ANNEX 1.1
No.
Recommendation
Related
recommendation/s
Timescale
Chapter 1
All of the recommendations in the Executive Summary appear in the relevant Chapters
Chapter 2
There are no recommendations in this Chapter
Chapter 3
3.1
R3.1
Further Consultation
Hold further meetings and consultations with stakeholders, especially LAs and HA (and equivalents), as
product of the recommendations
8.5
5.8. 6.3 and 6.4
Short-term
Chapter 4
4.1
Publications
R4.1
Undertake further investigations into requirements for (a) quarterly statistics and (b) their frequency of
publication
Short-term
R4.2
Develop revisions strategy
Medium-term
Chapter 5
5.1
Road lengths
R9.2
R5.1
Agree definitions of types of thoroughfare with key users
R5.2
SR2 to move from OSCAR to ITN road network,working closely with ITN users in the Dept and OS;
SR2 to maintain and update ITN in similar manner to OSCAR
6.1
Short-term
SR2 to seek, through OS, separate identification of C roads on ITN
6.1
Short-term
R5.3
5.2
Core Census (outside London)
Medium-term
R5.23 and R8.4
R5.4
Extend Core Census sites to locations previously considered problematic, particularly motorways and in
urban locations
Medium-term
R5.5
The Expansion Factor categories should be reviewed to ensure that they are appropriate.
Short-term
R5.6
Estimate selection probabilities of all Core Census sites and use inverses of these quantities in
calculations for Expansion Factor categories
Short-term
R5.7
When a Core Census site changes its road-class category, it can continue to be used in calculations for
the new category
Short-term
5.3
R5.8
5.4
London Core Census
DfT to review with TfL statisticians, to consider appropriate stratification of sites, renewing ATC
equipment, devising methods of obtaining classifiied counts, and deciding responsibility for processing
data
R10.2
Short-term
Manual Counts
R5.9
Review the coverage of the "neutral" counting days, the case for 16-hour counts and the
merits/demerits of counting for only 6 in every 15 minutes
Short-term
R5.10
Review different methods for calculating expansion and growth factors
Short-term
5.5
Minor Roads and Pedal Cycles
R5.11
Use panel of minor road sites to estimate year-on-year changes, combined with benchmark exercise as
soon as practicable and at 5/6 year intervals
Short/mediumterm
R5.12
Include some sites chosen to represent high pedal cycle flows, and some of those sites should involve
counts in summer months and at weekends
Short-term
R5.13
Major road pedal cycle traffic can continue to be estimated as at present, and cycle traffic on off-road
sections of the National Cycle Network should continue to be covered by SUSTRANS
Short-term
R5.14
Review continued publication of provisional quarterly estimates for pedal cycles and HGVs on minor
roads, once confidence intervals for these estimates have been calculated
Medium-term
5.6
R5.15
5.7
Seasonal Adjustment
The Quarters of Cyclical Dominance diagnostic from X11 ARIMA should be reported, along with advice
on the use of seasonally adjusted figures
Short-term
Disaggregated Outputs
R5.16
Vehicle kilometres by main (major roads) speed-limit bands can be calculated straightforwardly using
the AADF data
Ongoing
R5.17
Vehicle kilometres by main time-periods can be disaggregated using a combination of AADF and ATC
data
Ongoing
5.8
Local Authority and HA data
R5.18
Local Authority ATC data for major roads should be used to provide AADFs, along with LAs estimates
of AADFs based on shorter counts
3.1, 6.3 and 6.4
Medium-term
R5.19
There is a need for greater standardisation of traffic counting practice across and between LAs and HA
regions, and DfT should play its part in co-ordinating this
3.1, 6.3 and 6.4
Medium-term
5.9
Data Quality
R5.20
Check counts are needed for the ATCs and manual counts. These would indicate the need for
corrective action and, in the case of manual counts, would allow the variance due to random counting
error to be calculated
R8.1
Short-term
R5.21
Standard errors of AADFs should be calculated, as should standard errors of the total vehicle-km
estimates and disaggregations of these totals; it is further recommended that DfT commissions work to
develop these in view of the complexities involved
R8.1
Medium-term
5.10.
Vehicle Types
R8.3
R5.22
Some tidying up required to achieve consistency between the ATCs and manual counts in the
treatment of vans and lorries with draw-bar trailers
R5.23
Trials of revised loop signatures should be undertaken to improve ATC classification between buses
and rigid lorries
R5.24
Bus and motorcycle traffic estimates can be presented separately in the quarterly tables, if necessary
as moving averages
R5.25
Explore feasibility of linking ANPR data with traffic counts, with view to exploiting finer classifications
available on the DVLA website
Medium-term
5.2 and R8.4
Medium-term
Short-term
6.2 and 7.1
Medium-term
R3.1
Medium term
Chapter 6
6.1
R6.1
Mapping Data
DfT to work towards setting mechnisms in place to ensure ITN traffic-related features meet LA users'
requirements
6.2
Technologies
R6.2
DfT to consider needs of national and local monitoring and commission, collate and disseminate
research into the accuracy, reliability and costs (equipment and human resources) for existing and new
technologies
R5.25 and 7.1
Long Term
R6.3
Establish mechanism, with LAs, to systematically trial new technologies
R5.25 and 7.1
Long Term
6.3
R6.4
R6.5
6.4
LA data
DfT to work with LAs with view to using their AADFs for major road links in the road traffic estimates
DfT to explore how LA data on minor roads traffic can best be exploited
3.1 and 5.8
Short/Medium
term
Short/Medium
term
3.1 and 5.8
Highways Agency (and equivalent Scottish and Welsh) data
R6.6
Undertake trials of HA data to test their reliability for DfT monitoring
3.1 and 5.8
Short term
R6.7
If trials are successful, identify most cost-effective means of disaggregating HA data by vehicle-type, eg
special manual counts, "twinning" with nearby DfT sites, use of existing ANPR data (subject to
availability)
3.1 and 5.8
Medium Term
R6.8
DfT should aim to maximise its use of HA data, subject to the above trials [referred to in other
recommendations]. The aim would be to obtain data on AADFs on as many motorway and trunk road
links as possible. The more ambitious aim would be to use HA data to extend the coverage of
continuous counting throughout the year, so that the Core Census extended its reach to the more
'difficult' sites that it cannot currently handle.
3.1 and 5.8
Long Term
R6.9
DfT to exploit any traffic data that the Highways Agency collects for its own ad hoc operational
purposes
3.1 and 5.8
Medium Term
6.5
VOSA data
R6.10
6.6
R6.11
Monitor availability of MOT test and other data to compare mileages against road traffic estimates
Long Term
RIF
DfT to play key role in development of RIF, including with the aim of ensuring a common network
referencing system is implemented
Ongoing
Chapter 7
7.1
Assessment of new technologies
R7.1
Assess possible use of new technologies against defined set of criteria
R5.25 and 6.2
Long -term
R7.2
Identify when and how data from the new technologies can be incorporated into the road traffic
estimates
R5.25 and 6.2
Long -term
Chapter 8
8.1
Manual Traffic Counts
R8.1
Establish systematic programme of spot checks
R5.21
Medium-term
R8.2
Introduce indicators of manual counts' data quality
R5.21
Short-term
R8.3
Update guidance on vehicle classifications
5.10
Short-term
8.2
R8.4
8.3
Automatic Counters outside London
Implement improved vehicle classification algorithms used by ATCs outside of London
Quarterly Provisional Estimates
5.2 and R5.23
From 1
January 2007
(subject to
satisfactory
assessment of
the impact)
R8.5
8.4
R8.6
8.5
R8.7
8.6
R8.8
8.6
R8.9
8.7
R8.10
Implement more automated and systematic mechanism for checking and patching data when move to
new IT system
Medium-term
Automatic Data Quality
Introduce indicators of automatic counters' data quality
Maintenance of ATCs
DfT, HA, WAG, SE and local authorities to explore the potential for closer working on ATC issues
Short-term
3.1
Short-term
Medium-term
Heavy Goods Vehicle Estimates
Investigate discrepancies between HGV mileage estimates based on CSRGT and the Road Traffic
Estimates
Long-term
BuiltUp/NonBuilt Up Splits
In view of the poor accuracy of the BU/NBU split, it is suggested that SR2 ceases to supply these data
to users. However, before a final decision is made, further consultation should take place with LGF,
SR3 and other users.
Short-term
Road lengths
SR2 move to the latest version of the ONS Urban/Rural polygons
Chapter 9
9.1
R9.1
IT systems
Medium-term
R9.1
9.2
R9.2
9.3
R9.3
Explore the possibility of transferring the current provisional quarterly estimates' systems to a system
accesssible to both SR1 and SR2
R8.5 and R9.3
Medium-term
5.1
Medium-term
R9.1
Medium-term
GIS: road lengths
Investigate the necessity and practicality of a benchmarking exercise for road lengths in England
Risk Assessment of Current Procedures
Complete risk assessment of current procedures, focusing on documentation of procedures and
staffing continuity
Chapter 10
10.1
R10.1
10.2
R10.2
Manual counts and automatic counters
Undertake further trials and research into possible replacement of manual counts with automatic
counters
Long-term
London traffic counts
Agree way forward with TfL on replacement of London data collection system
R5.8
Short-term
ANNEX 2.1
Road Traffic Statistics Quality Review
Project Initiation Document
Project Name:
Road Traffic Statistics Quality Review
Date:
September 2004
Author:
John Garnsworthy, TSR2, DfT
Contents:
1. Introduction
2. Aims and scope of the review
3. Nature of the review
4. Roles and responsibilities
5. Quality Review Procedures
6. Communication and Consultation Plan
7. Project Timetable
8. Project Risk Register
9. Quality Assurance
10. Annex - Background note on road traffic estimates methodology
1. Introduction
1.1 The Government’s White Paper – Building trust in statistics – includes a
commitment to assure the quality of national statistics and refers to a programme of
reviews of outputs. The Road Traffic Statistics are part of the Quality Assurance
programme.
1.2 The Department for Transport produces annual and quarterly traffic estimates with
varying degrees of disaggregation by road and vehicle type. The Annex sets out key
aspects of the methodology that the Department uses to produce its estimates. In
essence, it hinges on combining Annual Average Daily Flow data (AADF) with road
lengths to produce an estimate of the distance travelled in Great Britain by different
vehicle types.
1.3 The Department collects a large volume of data to estimate AADFs for 22 road
categories. It undertakes, through contractors, around 10 thousand manual counts of
traffic each year across both the major and minor road networks. However, manual
counts only take place for 12 hours on a given day and the sites represent just a tiny
proportion of the road network. The Department also funds around 190 automated
traffic counters that collect traffic data by 20 different vehicle types 24 hours a day.
These are used to estimate expansion factors that can transform data from the manual
counts into a 24 hour AADF for the 22 road categories.
1.4 Road length data are derived from a number of sources. Estimates of road length
for major roads were agreed two years ago with key stakeholders and remain in use
with minor maintenance by the Department to take on board changes to the major road
network. Data for minor roads are estimated from Ordnance Survey data, which is
agreed annually with local authorities through the Department’s R199b return.
1.5 The Department produces its own annual and quarterly publications of traffic
estimates. The outputs are of significant interest to policy colleagues within the
Department and feed into the Local Government Finance Settlement for allocating
Government grant to local authorities. Business finds the figures useful in their own
development and planing decisions. These are but a few uses. The importance of
sound, reliable and valued traffic estimates extends far beyond the headline
publications that the Department produces.
1.6 The current methodology is designed primarily to provide robust and sound
statistics at a national and regional level. It does this through a "bottom up" approach
by aggregating data for individual roads and local authority areas. The Review will
address how best to handle the considerable and growing demand for statistics for
these building blocks. Allied to this is the need for detailed data to feed into measures
of congestion. Ideally these are required for individual road links and specific times of
the day. Local authorities have a key role to play, both as providers of data for such
estimates and as end-users to meet their Public Service Agreement requirements.
1.7 This paper sets out the proposed arrangements for this review and outlines its
scope.
2. Aims and Scope of Review
2.1 National Statistics, the umbrella under which Road Traffic Statistics fall, is a badge
of confidence. It must be able to signal to users that the statistics produced under the
label have been rigorously prepared to the highest standards, are free from bias and,
as such, can be relied upon and generate confidence from users.
2.2 The ultimate aim of this review, therefore, is to ensure that Road Traffic Estimates
are fit for purpose by ensuring that the Department produces traffic outputs which:





Meet the need for information at a variety of geographic levels;
Are sufficiently robust and reliable for the purposes that they are used;
Are consistent with other traffic information;
Are competitive in terms of cost and efficiency;
Use sound methodology, systems and techniques.
2.3 It is recognised that there might be insufficient time to address all relevant issues
during the course of the review itself. Recommendations for follow-up work will
therefore be made as appropriate.
3. Nature of the Review
3.1 The National Statistics (NS) Planning Board decided that all NS outputs or sources
earmarked for review should be allocated to one of the following categories of review:
Strategic Review – this would apply to a group of related outputs or cross-cutting
sources which involve major risks and areas of significant or changing user demand.
Fundamental Review – this would apply to a ‘key’ NS output which has generated
known concerns, has not been subject to a recent review or where the risk of not
conducting a review is deemed to be high.
Standard Review – this would apply to any output which is important and where the
purpose of the review is to ensure that the output is (still) of good quality and fit for
purpose.
Targeted Review – this would apply to any output which may have already been the
subject of a broader review but which has since generated a specific area of concern.
3.2 Road Traffic Statistics will be subject to a Standard Review carried out over a
period of around nine months. The review will cover traffic in Great Britain only and be
based on five strands of investigation and consideration:
Producing the right outputs – this is about better understanding who the customers
are for Road Traffic Statistics and what outputs they need from the Department to
effectively conduct their own business. It will require examination of whether existing
outputs meet this need, including the growing demands for statistics at more
disaggregated geographic levels. The review will also examine how the outputs are
being used, the understanding that customers have of them and what problems
customers perceive there to be with the outputs. The review will look at what other
outputs are being used by customers and whether or not the Department needs to
consider specialist products for diverse customers;
Quality inputs – the quality of the Department’s outputs will be significantly driven by
the quality of the data it collects. The review will seek to establish the quality and
accuracy of the data from both the manual and automated traffic counts and
consistency with data from other sources (e.g. Highways Agency, local authorities, and
Transport for London). Consideration will be given to the possible means for different
agencies to pool resources and utilise each others data for traffic estimation A key
element of this will be the need for close liaison with local authorities to maximise the
consistency of their data collection methodologies. The potential of new technology
(e.g. satellites, traffic control systems, aerial photography) to collect more timely, cost
effective and accurate data will also be explored. An important aspect of this strand will
be the storage and handling systems that the Department uses when it receives the
data to ensure they are robust, effective and efficient.
Sound methodology and reliability – this strand will be about the methods employed
to produce estimates and the quality of the outputs. It will examine international
approaches to producing traffic estimates, what completely new approaches may be
available and the robustness of the Department’s existing methodology (e.g. how it
estimates expansion and grossing factors, understanding urban and rural definitions).
The review will seek to establish the reliability of current outputs by examining revisions
to the estimates, reconciling the outputs with other known results, including CSRGT,
and calculating confidence intervals for estimates. An important output of the review will
be to establish a definition of quality for Road Traffic Statistics which will necessarily
consider balancing an easily understandable methodology against the high level of
robustness that customers rightly require. Storage, handling and analysis systems will
be a feature of this strand, very much linked to some of the work within strand two.
Communication – good, robust outputs only have worth if they are effectively
communicated to the audiences they seek to serve. This strand will establish how well
the Department communicates its outputs to its customers: does it sufficiently and
effectively explain its methodology to customers, are results readily available in a
helpful and informative format and could more use be made of new technology,
particularly the internet? The review will look at ownership and supply issues
surrounding data and results and will consider the issues of charging.
Future development – the estimates are currently made using manual and automatic
traffic counters only. This strand will focus on how the Department will position itself to
take advantage of new technological developments to further improve Road Traffic
Statistics. In addition, this strand will recommend how the Department can best keep
pace with customer views on its outputs and their changing needs.
Currently, the road traffic estimates are made using a well-established methodology,
defining what data are collected and how. However, there is no overriding rationale for
this to continue. A possible outcome of the review might therefore be for TSR to set
out its requirements for road traffic statistics and invite contractors to provide an
appropriate service to meet them. This could include data collection and perhaps
subsequent handling and analysis.
4. Roles and responsibilities
4.1 A Review Board will oversee the work of a Review Team that will be led by a
Review Manager. A Review Sponsor will oversee the review as a whole and has
ultimate responsibility for the review. Details of the roles and
responsibilities of the teams and their members are outlined below:
Alan Oliver
Tom Worsley
DfT, Transport Statistics Roads,
Chief Statistician
Integrated Transport and Economic
Analysis Division, DfT
Chair, Review Sponsor
Internal user
John Disney
Nottingham Trent University
External Adviser
Nick Fenton
Highways Agency
External User
Frank Dixon
Scottish Executive
External User
Sarah Demery
Welsh Assembly
External User
Ray Heywood
Leeds City Council, LA representative
External User
External Consultant:
methodologist
Martin Dale
John Garnsworthy
DfT, Head of TSR2
Review Manager
Louise Buckle
DfT, TSR2 Assistant Statistician
Secretariat
Review Board
4.2 The composition of the Review Board is:
4.3 The responsibilities of the Review Board are:
 to ensure the Review Team utilises all its resources such that the review
meets its objectives (including agreeing any changes to the timetable), and
resolves any difficulties that may arise; to agree scope of the review and the
PID;
 to assure the review's technical and business integrity, to ensure that
effective quality assurance is in place and check the review is being
conducted according to specified guidelines;
 to consider emerging findings throughout the review taking account of wider
issues related to the areas of the review;
 to consider the possible sensitivity of the report findings and ensure that
stakeholders are appropriately consulted throughout the review;
 to consider and agree the final report, containing recommendations for the
future of the Road Traffic Statistics, to be submitted to the National
Statistician;
 to commission an initial implementation plan to take forward the
recommendations, to be produced within 3 months of the final report being
released.
4.4 In addition, the External Adviser will:
 Provide advice on the consultation with users, including the consultation
documents
 Comment on the proposals arising from the methodological advice.
Review Sponsor (Alan Oliver)
4.5 The responsibilities of the Review Sponsor are:
 to identify and manage risks to the project in the context of corporate
management responsibilities;
 to help overcome barriers to the successful completion of the review which
are brought to his attention
 to oversee the work of the Review Team; and
 to chair the Review Board.
Review Team
4.6 The composition of the Review Team is:
DfT, Head of TSR2
John Garnsworthy
Review Manager
Martin Dale
External Consultant, Methodologist *
HA representative
David Pearce
User representative
LA representative
William Bryans, Surrey County
Council
User representative
ITEA representative
Malcolm Jay and Russell Harris
User representative
TSF representative
Chris Overson
User representative
TSPT representative
Stephen Reynolds
User representative
TSR4 representative
Mouna Kehil
User representative
Scottish Executive
representative
DfT, TSR1
Frank Dixon
User representative
Andy Lees
DfT, TSR2
Andrew Smith
Topic Manager: Automatic Traffic
Counts
Topic Manager: Manual Traffic Counts
DfT, TSR2
Kate Thomas
Topic Manager:- Dissemination and
GIS issues
DfT, TSR2
Louise Buckle
Statistical Methodology/Data Quality,
User Requirements and Secretariat
* Co-opted member of the Review Board
4.7 The responsibilities of the Review Team are:
 to draw up the PID and make any subsequent changes to it;
 to prepare additional documents and consult with users;
 to develop options and future recommendations for road traffic statistics;
 to report progress to the Review Board
4.8. It is envisaged that most of the work of the Review Team will be undertaken by small subgroups addressing specific, well-defined issues.
Review Manager (Head of TSR2)
4.9 The review manager is responsible to the Review Board for the following:










to assign tasks to individual team members and ensure that they deliver to
the agreed timetable;
to ensure the effective day-to-day management of the Review Team;
to produce and maintain a record of changes to the plans and timetable;
to ensure that the Project Risk Register is maintained;
to prepare the draft PID and reports for discussion by the Review Team;
to regularly inform the Review Team and Review Board of progress;
to consult as necessary with team members and other interested parties on
the subject matter of the review;
to report regularly to the Review Sponsor progress and any problems
encountered;
to produce necessary papers for the Review Board; and
to maintain a document and decision log.
Review Team Members
4.10 The responsibilities of the Team Members are:
 to take forward individual work strands and deliver on time;
 to contribute papers, as necessary, on issues of particular interest or
importance;
 to provide data, details on methodology and its rationale as required;
 to represent the views of colleagues where appropriate; and
 to advise the Review Manager of any problems which may affect their
participation.
5. Quality Review Procedures
5.1 The conduct of a quality review is a key element of the National Statistics approach
as set out in the Framework document (available from www.statistics.gov.uk). In each
case, the reviews should involve thorough consultation with data users and incorporate
a quality assurance procedure involving external experts.
5.2 The procedure for the review is:
i.
ii.
Review Board considers and agrees the scope of the review;
Review Manager/Team consults users and suppliers and identify the major
issues
to be reviewed;
iii.
Review manager/team consults users and suppliers and prepares a report of
emerging findings;
iv.
Review Board considers/agrees emerging findings;
v.
Review manager/team produces final report and recommendations;
vi.
vii.
Review Board considers/agrees final report
Ministers notified that review completed and appraised of key
recommendations
viii.
Final report submitted to National Statistician and published on NS website.
5.3 If the quality review shows areas of Road Traffic Statistics that need to be changed
and improved the Review Team will assess the various options in turn. In each case a
preliminary assessment of the costs and operational practicability and other information
will be provided. Where appropriate the Review Team will make recommendations to
the Review Board with regard to the desirability and priority that should be attached to
each improvement taking account of user preferences.
5.4 Any recommendations for change that emerge from this review would be subject to
the usual processes of consultation with users before implementation.
5.5 The review is being managed using project management techniques. A project
risk register has been developed (section 8) and will be updated and amended as
necessary during the course of the review. A timetable for the review is given at
section 7; this may be revised and updated as the review progresses.
6. Communication and Consultation Plan
6.1 The main aim of the review is to assess whether Road Traffic Statistics are fit for
purpose and of good quality. This is about ensuring that the right products are
produced well. Consultation with key customers and stakeholders is critical to the
success of the review. The review will not be able to ensure Road Traffic Statistics are
fit for purpose if it has not fully considered the views of those most in need of and
affected by the Department’s outputs.
6.2 The Review Team will ascertain views on (amongst other things):















The breadth of use of Road Traffic Statistics;
What outputs customers require;
Do existing outputs meet customer needs;
Are any specialist products required;
How are Road Traffic Statistics being used;
How well do customers understand the outputs;
What other data do customers use;
What problems are encountered with Road Traffic Statistics;
How reliable do customers find the Department’s outputs;
Does the Department’s perceived reliability compare favourable with other
data;
Do customers understand or want to understand the underlying
methodology;
How well does the Department communicate Road Traffic Statistics;
How do customers wish to receive outputs and data;
What should the Department make available to customers;
How much should the Department charge for specific services.
6.3 One particular area that the review will focus on is estimates of pedal cycling
which individual local authorities and other organisations, as well as the Department,
have considerable interest in. The estimates of cycling levels display a substantial
degree of volatility, rendering their interpretation difficult. It is possible that this, at least
partly, reflects deficiencies in the current methodology, especially lack of
representativeness of the sample, rather than genuine variations in cycling levels.
6.4 In order to address these issues, DfT is preparing a pilot study with a limited
number of local authorities for the installation of automatic pedal cycle counters within
their areas. It is proposed that LAs will maintain the equipment, with the Department
analysing the data and feeding back results to them. Local authorities will be closely
consulted on this and related issues. These will include what types of journey should
be covered by the estimates, eg on roads only (as currently collected) or off-road
journeys on tow-paths and designated cycleways in parks also?
6.5 It is therefore vital that input from a wide variety of users on the quality of the Road
Traffic Statistics is obtained. The Review Team will draw up a questionnaire that will
gauge users’ thoughts on the quality of the data as well as ascertaining the uses made
of the statistics and identifying areas of particular importance to them. The
questionnaire will be circulated to a wide range of users.
6.6 The Review Team will use the questionnaire as a benchmark and starting point.
This will be followed up through more detailed discussion with customers and
stakeholders to fully understand and develop specific issues that emerge from the
questionnaire and other review activity. The Department’s website will be actively
used to openly publish Review Board papers and seek feedback on review activity and
issues.
7. Project Timetable
7.1 The proposed timetable for the main stages of the review is outlined below:
Review Team to draft PID, including scope of review (August) – The scope of the
review presented in this PID will be discussed and agreed by the Review Team and
presented to the Review Board for approval.
First meeting of the Review Board (early September) – The main aim of this meeting will be
for the Review Board to approve the PID, especially the scope of the review, the timetable and
suggested procedures. The Review Board will note that after the PID has been approved it will
then be published on the National Statistics website.
Consultation Phase (mid-September to end December) – The initial plans for the
consultation process are presented at section 6. In late October, a progress report will
be produced and circulated to the Review Board for consideration. It will set out how
the review is developing, reassess the timetable and suggest any changes that are
necessary.
Diagnostic Phase ( early October to December) – This phase of the review will
consider how well the Road Traffic Statistics meet users needs, the quality of the input
and methodological issues. The views that have been gathered in the consultation
phase along with the knowledge of the Review Team will be used to produce a
thorough assessment of the Road Traffic Statistics. A second ‘progress report’ will be
circulated to the Review Board in November, This will consider similar issues to the
first ‘progress report’ and include findings from the user and supplier consultation and
methodological investigations.
Second meeting of the Review Board (late November) - This meeting will provide the
opportunity for the Review Board to consider how the project is progressing and initiate
any necessary changes as well as review the preliminary findings from the consultation.
Follow up work and future options (December to April) – During this stage the
Review Team will follow-up any work arising from the second Review Board meeting.
The Review Team will also begin considering whether there are any changes required
to the Road Traffic Statistics and the costs and practicalities involved in any of the
options suggested. It may be necessary to hold a third Review Board meeting to
consider the results of any follow-up work and future work programme.
Preparation of the Final Report (May) – The Review Team will prepare a final draft
report, with a recommended course of action. This will be circulated to the Review
Board in May.
Third meeting of Review Board (May) - The Review Board will consider the draft of
the final report. The main purpose of this meeting is to determine that the review has
been carried out satisfactorily and to approve the final report.
Final report and dissemination (early June) – The final report, after approval from the
Review Board, will be presented to the National Statistician. The report will then be
published on the National Statistics website.
Within 3 months of the publication of the final report, a plan for implementation of the
recommendations and/or progress made will be drawn up and made publicly available.
This will allow users and suppliers to assess the progress that is being made on each
recommendation. This plan will be published on the National Statistics website.
8. Project Risk Register
Risk
No:
1.
Risk
Likelihood
Impact
Staff changes and
vacancies in the
Review Team
Low
Medium
2.
Other work is
deemed higher
priority
High
Medium
3.
Unplanned
expansion of the
scope of the
review
New issues
emerging during
the consultation
process
Medium
Medium
Medium
Medium
5.
Inability of the
Review Board to
reach agreement
Low
High
6.
Unexpected
intervention at high
level that has
major implications
on Road Traffic
Statistics outputs,
e.g. unexpected
change to major
policy need.
Low
High
4.
Implication and Countermeasure
May lead to delay in the
completion of work. Solution:
Ensure that adequate
documentation of work is
maintained. Review Manager
to ensure that should staff
leave, adequate resource is
given from elsewhere.
Delay and/or reduction of
quality of the review. Solution:
Ensure the key staff are aware
of responsibilities and senior
managers give adequate
commitment to the review.
More topics to cover will lead
to an extension and/or
reduction of the quality of the
review.
A low-level risk as key users
have been consulted when
drawing up the list of issues to
be reviewed. The Review
Manager should monitor the
consultation process carefully
and report any potential
difficulties to the Review
Board.
It is vital that the members of
the Review Board are fully
committed to the review and its
outputs. Solution: the Review
Sponsor should be aware of
any potential points of
contention and will seek to find
common agreement between
members.
A low-level risk, as policy
colleagues are members of
both the Review Board and
Review Team. Solution:
Members of Review Board and
Review Team should inform
the Review Manager of any
likely events that may have a
major impact on the review
9. Quality Assurance
9.1 It is important that the quality of the review itself is monitored. This will be done in
the following ways:

Members of the Review Team will hold one another accountable with regard
to quality, assuring all documents that are sent out and that the Quality
Review procedures are adhered to. Regular team meetings will help
facilitate this.

Records of the Review Team's key activities will be available to the Review
Board on request.

The Review Team will include an outside consultant.

A representative from the ONS National Statistics Policy Division (NSPD) will
be available, if required, to advise on the process of the review.

Membership of the Review Board includes a member of the academic
community.

The Review Board will be responsible for quality assuring and agreeing the
PID, the progress reports and final report.

ONS will be advised about the progress of the review

The implications of the review, including time and costs of any proposed
changes to the Road Traffic Statistics, will be thoroughly investigated by the
Review Team before being put forward to the Review Board.
Annex
How the National Road Traffic Estimates are madeAmendment History
Version
number
1.1
1.2
Date of
Amendment
March 2004
April 2004
1.3
1.4
May 2004
August 2004
Section amended
Summary of change
All sections
Minor Roads
paragraph 17
Paragraph 22
Paragraph 18
Completely revised
Minor change to improve accuracy in
methodology description.
Amended link to CLIP web site
Amendment made to description of minor
road methodology
How the national traffic estimates are made
Introduction
1. In 2002, motor vehicles travelled about 485 billion kilometres while pedal cycles travelled
4.4 billion kilometres along the public roads of Great Britain. How do we know this? We
know this from the traffic counts conducted about many different types of roads and
information on road lengths. However it is not easy to convert count data to total traffic data
2. The road network consists of about 50,000 kilometres of motorways and class "A" roads,
with a further 340,000 kilometres or so of minor roads. The road system as a whole is thus
much too extensive to allow the collection of comprehensive traffic data for every part.
Moreover, the density of traffic carried (and the mix of traffic by vehicle type) varies
enormously from place to place and from hour to hour. Flows are less than 100 vehicles a
day on many minor roads but exceed 150,000 a day on some motorway links. Even within a
road class, one site may easily carry ten times as much traffic as another. Flow also varies by
time of day, by day of week and by month of year. Furthermore, there is variation within the
variation - car traffic levels, for example, may change relatively little over the seven days of
the week, but goods traffic is usually at far lower levels on Saturdays and Sundays than on
other days. These characteristics mean that estimating national traffic volumes requires a
fairly complicated sampling design, the collection of substantial volumes of data and
complex computational procedures. This note describes the new methodology used for the
traffic estimates from 1993 onwards.
3. Estimation of traffic levels uses information from both manual and automatic counts. They
are each described briefly below.
The manual counts (previously referred to as the rotating census and the
biennial counts)
4. These counts operate somewhat differently for major and minor roads. The major roads are
split into five road classes: motorways, trunk roads and principal roads with the latter two
divided into urban and rural roads. Urban1 roads are defined as those within the boundaries
of the Urban Area polygons for settlements of 10,000 population or more, based on the 2001
Population Census. On the outskirts of urban areas, bypasses are normally treated as rural
even if part of the road may lie within the urban area polygon. Conversely, roads between
urban areas with short lengths outside the polygons are normally treated as urban. Minor
roads are divided into 6 classes: B class, C class and U (unclassified) roads, each subdivided into urban and rural.
5. For major roads (motorways and A-roads), the traffic on every link - normally a section of
road between consecutive junctions with other major roads - must be regularly assessed.
This is done by counting the traffic at a statistically random point on most links at regular
intervals – traditionally, once every three years in England and Wales and once every six
years in Scotland2. It is recognised that with the exception of motorways, traffic levels will
1
Before 8 May 2003, roads were, instead, classified as built-up and non built-up. Built-up roads were
those with a speed limit of 40mph or less. Non built-up roads were those with a higher speed limit.
2 Since 1999, the frequency of counting has been broadly based on the variance of the traffic of the link
(variance of average flow * link length). Intervals are every one, two, four and eight years in England and
Wales. In Scotland they are two, four, eight and sixteen years. However, Scotland is proposing to
increase its budget for traffic counting and so may come into line with level of counting in England and
Wales.
vary along the length of a link. However, the procedure of counting at a statistically random
point on each link can be expected to lead to good estimates at national level although
estimates on some individual links may be less reliable.
6. In total, about 5,300 major road sites are scheduled to be counted in 2004. In addition to
traffic count data, information is collected about the characteristics of each link, such as its
length and the road class and road width at the place of the count. At each chosen point,
trained enumerators count vehicles of each of eleven types (pedal cycles, two-wheeled motor
vehicles, cars and taxis, buses and coaches, light vans, and six separate categories of goods
vehicle) for the 12 hours from 7am to 7pm. These counts are all scheduled to take place on
weekdays, but not on or near to public holidays or school holidays. To minimise the effects
of possible seasonal factors, counting is confined to the so-called "neutral weeks". These are
namely most weeks in March, April, May, June, September and October.
7. Some major road links are unsafe to count or are too short to be worth counting in the
normal way. In these cases, traffic estimates are derived from the judicious use of flow data
on adjacent links. These are called derived links. Further, because all links are now defined
as ending at a local authority boundary, some links are treated as dependent links. In these
cases, it is assumed that the flow is the same along all of the link. So, a count in one local
authority can be used as a proxy for the flow on the dependent link. In 2003, there were
15,500 normal links, 1,200 derived links and 1,000 dependent links. Complete coverage of
the minor road network is not attempted as it is too extensive. It is not practicable to define
the minor road network in terms of individual links; even if all the links could be identified,
their number would be far too great to allow traffic data for each link to be collected. Minor
road traffic estimates are therefore made by grouping minor roads into one of the six road
classes3. An attempt is then made to measure the average flow on each of these road types
by carrying out a number of counts along them. A random sample of approximately 4,500
sites across GB is visited each year. These same sites are counted each year. Most of these
counts are carried out in neutral weeks. However about 200 counts per year, the "summerwinter counts", are carried out in non-neutral weeks and on weekends. These 200 sites are
visited each year and are mainly used to provide extra information about two-wheeled traffic
throughout the year. This is because pedal cycles and motor cycles are not always accurately
identified by automatic counters (see paragraph 9).
8. The manual counts have the advantage over automatic counts of complete coverage of major
road sites and moderately good coverage of minor roads. However, the data (hourly by
vehicle type), are very sparse since traffic is counted for only 12 hours on each visit. Thus
these counts give no information about traffic at night, at weekends, over public holiday
periods, and little about the seven non-neutral months. In calculating national traffic
estimates, therefore, use must be made of data from automatic counters.
The automatic counts (previously referred to as the core census)
9. The automatic counters fill the gaps left by the manual counts. There are some 160 sites in
GB outside of London where traffic is monitored continuously using automatic sensors,
which classify the traffic into vehicle type. The numbers of vehicles of each type detected
are combined into hourly totals and stored on-site until it is downloaded during the night to a
computer in the DfT headquarters building. The automatic counting equipment recognises
3
These include: urban and rural 'B' roads, 'C' roads and unclassified roads
22 different types of vehicle; these are then combined to provide estimates for the eleven
vehicle types4 used by DfT.
10. The automatic counters do not give 100% accuracy. For example, the equipment cannot
classify vehicles into their different types when the traffic is moving very slowly (5mph or
less). They also have a tendency to malfunction, though the new sensors recently introduced,
are more reliable than the previous ones. The equipment cannot distinguish between cars and
car-based vans and can also have difficulty distinguishing between some types of buses and
coaches and goods vehicles having similar axle spacing and chassis height. The equipment is
also prone to failing to identify two-wheeled vehicle traffic, both bicycles and motorcycles5.
Nevertheless, they do have the big advantage over manual counts that they operate
continuously and so can give a complete picture of traffic at the points where they are sited.
11. The automatic counters in London are slightly different to those outside London. There are
56 automatic counters in London and they are "volumetric" classifiers that only distinguish
between short (up to 5.2 metres) and long (greater than 5.2 metres) vehicles. They need 24hour manual counts every three months to provide estimates of the breakdown of traffic by
vehicle type in each hour of the day. These counters suffer from similar problems as those
outside London, but are more reliable than the automatic counters used outside London.
Annual Average Daily Flows (AADFs)
12. The data for all manual counts done in neutral months6 are combined with information from
automatic counters on similar roads to provide an estimate of the AADF at that site. This is
normally done by multiplying the raw count data by factors derived from the automatic
counts in that same year. There are a large number of such expansion factors since there are
separate factors for each vehicle type, day of counting and expansion factor group7. Because
these counts are done in neutral weeks, the expansion factors used do not usually vary too
much from year to year, except when bad weather has restricted traffic during the winter
months. For cars, the factors are usually between 1.15 and 1.25 (except on motorways and in
London where the factors are higher) while for goods vehicles the factors vary between 0.75
and 0.9 - lower because of the greater drop in traffic at weekends.
13. The automatic counters provide a reasonable guide to changes in traffic over time. This
information is used in two ways. Firstly it is used to provide provisional quarterly estimates
of traffic, which are published on the sixth Thursday following the end of a quarter.
Secondly they are used to provide growth factors between consecutive years. These growth
factors are used for links not counted, or not counted satisfactorily in the latest year. In these
cases, the AADF for the previous year is multiplied by the appropriate growth factor to give
a reasonable estimate of the AADF for the latest year.
4
These include: Pedal cycle, two-wheeled motor vehicle, car, light goods van, bus, rigid 2 axle lorry, rigid
3 axle lorry, rigid 4 or more axle lorry, 3 axle or 4 axle articulated, 5 axle articulated and 6 axle or more
articulated.
5 This deficiency underpins the need for the extra sites for manual counting on minor roads in the
summer and winter months
6 Counts done in the summer and winter months are not grossed-up: the sites are normally counted at
the same time of the year each year and so are compared directly with each other.
7 There are now 22 expansion factors groups. These are based on type of area (from holiday area to
Central London), road category and, in some cases, traffic flow level.
Use of AADFs in Calculation of Annual Traffic Estimates
14. Different procedures are used for major and minor roads in converting AADF data to traffic
estimates. The difference arises because the link concept cannot be applied to minor roads.
Major roads
15. A major road link of length 2km with an AADF of 50,000 has a traffic figure of 100,000
vehicle-kilometres (2*50,000). This equates to 36.5 million vehicle kilometres a year.
Because every major road link is counted, in principle, total traffic on major roads can be
obtained by summing the traffic figures for every link.
16. As mentioned in paragraph 6, some links are not counted. In these cases, the traffic flows are
derived from adjacent links using suitable formulae (derived links) or using the flow of the
adjacent link as a proxy (dependent links).
Minor roads
17. In the base year (currently 1999), for each minor road class in each local authority an AADF
is estimated based on a sample of traffic counts, including those projected forward from
counts done in earlier years. These AADFS are then multiplied by the total road length for
the relevant minor road category to give an estimate of traffic for that road category.
18. Traffic for the latest year is then obtained by calculating changes in traffic flows derived
from the automatic minor road counts, after taking into account any changes in road length.
It is assumed that new minor roads have the same average flows as that of other minor roads.
This is plausible, since some of the newest will be quiet roads on housing estates whilst
others will be busy roads recently declassified from major road status.
Quarterly and annual estimates of traffic
19. As mentioned in paragraph 12, the automatic counters are used to provide provisional
quarterly estimates throughout the year. A first estimate for the year is published in early
February, and this is largely based on automatic count data. A final estimate for the year is
normally published in early May and this is produced by putting together the estimates for
major and minor road traffic as detailed above.
Availability of data
20. Basic quarterly data are included in the quarterly statistical bulletin Traffic in Great Britain,
which can be obtained by contacting Road Traffic Statistics The data are also included on
the DfT web site. They can be downloaded from the correct PDF file which can be located
underhttp://www.dft.gov.uk/stellent/groups/dft_control/documents/contentservertemplate/dft
_index.hcst?n=7405&l=4
21. More detailed information are available from the annual publication. This again is available
free, but also available from the DfT web site. This can be found at the following site:
http://www.dft.gov.uk/stellent/groups/dft_control/documents/contentservertemplate/dft_inde
x.hcst?n=8170&l=4
22. Advice on how to conduct local transport surveys can be found at the following web site:
http://www.clip.gov.uk/subgroups.asp?lsection=6&ccat=57
23. Some users need more detailed data for particular areas or particular roads. These can be
supplied, though there is a charge payable by external customers.
TSR 2
August 2004
Annex 4.1
User Consultation Report
Introduction
1. A key element of the Quality Review was consultation with users. This was undertaken
through two complementary mechanisms:


First, a seminar was held with users, under the auspices of TSUG, on 4 October 2004. At
the seminar, key speakers explained the background and purposes of the Review. More
importantly, it provided users with the opportunity to present their views on, and
requirements, for the statistics. To aid the discussion, attendees were invited to participate
in one of four groups, covering different topics. The first part of this Annex describes the
key points to emerge in discussion.
Second, a questionnaire was circulated to users seeking their views, in a more structured
manner on key issues relating to the statistics and users' requirements. The questionnaire
was circulated at the seminar and attendees encouraged to complete it. The findings from
the questionnaire are reported on pages 6 to 17.
USER SEMINAR
2. The results of the seminar were used to inform the design of the user questionnaire and the
drawing up of the work packages as part of the Project Plan. Limited weight was placed on
the results as it was felt that the attendees might not be representative of all users. There
were 42 participants including Review Board members and branch members.
DfT and
Agencies
HMCE
Review Board
members
Consultants
LAs
Academics
Journalists
Others
18
5
4
3
2
2
1
7
35
Note of User Seminar
3. John Disney of Nottingham Trent University chaired the seminar.
4. The first part of the seminar involved three introductory speakers from the Quality Review
Board: Alan Oliver and John Garnsworthy of the DfT and Ray Heywood from Leeds City
Council. This was followed by a brief discussion. The attendees then convened into their
pre-chosen discussion groups to cover one of four topics in detail. After these discussions
the groups then reconvened in a plenary session to present their findings and highlight key
issues raised.
5. Group A - Supplying Data for Local Needs
Chair: Ray Heywood
Scribe: Kate Thomas
Many local authorities undertake their own traffic counts with some also acting as DfT
manual count contractors. In some cases there might be unnecessary duplication of effort
and different approaches can lead to inconsistent results. In addition to local authorities'
data, a number of other current and potential data sources have been suggested to meet the
growing demands for traffic statistics.
Key points from the discussion:
 LAs generally use all of the published traffic and road length data
 The LAs undertake their own counts to ‘fill in the gaps’ left by the National Traffic
Census. The main areas of concern are with the minor roads data.
 Most LAs have long term systematic traffic monitoring programmes and often monitor
cycle traffic.
 There are some inconsistencies between DfT and LA data, especially in relation to
expansion factors and minor road flows.
 There is some consistency of manual counting practice across LAs and the DfT could
make more use of LA data.
 There was, however, a need to promulgate consistent definitions and count practice and
to give guidance on sampling methodology.
 On opportunities afforded by new technologies, group A identified GPS/IT IS, mobile
phones and image recognition. They were very much interested in journey times. There
were, however, implications for data protection and for storing large volumes of data.
6. Group B - Quality Counts
Chair: John Disney
Scribe: Chris Overson
A key element of the review is the quality of the outputs. It is essential that the statistics
produced are fit for purpose and as consistent as possible with estimates from other sources
Key points from the discussion:
 There was some scepticism about the need for quarterly data. The main need was for
good annual data.
 There was a need for both absolute figures (vehicle-kilometres and AADFs) and
percentage changes. Even those primarily concerned with traffic growth rates needed to
know the base line.
 Single ‘point’ estimates were not sufficient. Each estimate should have a 95%
confidence limit.
 Where there was more than one source of information about a variable, guidance should
be given on which data were appropriate for particular uses.



The need for revisions was recognised, in particular quarterly data, but users should be
given clear warnings when data were revised stating what had changed. Revisions that
are trivial for some users may be important for others.
There was a need to exploit Highways Agency and LA data. Incentives might be needed
to persuade others to pool their data.
There could be benefit in examining practice on presentation of outputs in other
European countries and in the EU
7. Group C - Delivering the Outputs
Chair: Nick Fenton
Scribe: Andrew Smith
Good, robust outputs only have worth if they are communicated effectively. DfT publishes
traditional paper-based quarterly and annual road traffic bulletins. These are also
disseminated on its website with other key statistics, including estimates for individual local
authorities. Whilst these dissemination mechanisms meet many users' needs, some 1,500
individual ad hoc requests per annum are dealt with too.
Key points from the discussion:
 DfT should show how the quarterly and annual data were linked together.
 More generally, they should show how the outputs were derived and how the various
inputs fitted in. This could be done by a flow diagram
 There were some doubts about the value of the quarterly bulletin
 All bulletins should show a full list of sources. Also a commentary on the trends would
be useful.
 Need for greater consistency in the breakdowns between the various publications.
 Explain more about what is available (and how this links to other data), what the stats are
normally used for and why certain stats are not available.
 Although road length data was not viewed as essential by many of the syndicate
members, it was regarded as essential by those who did require it.
 The DfT website can appear rather user-unfriendly. Tables should be downloadable to
Excel and other text to PDF. AADFs should be on a website.
 Frequency of publications is about right. However, could the Annual Estimates be
published earlier?
 The syndicate suggested looking at the neighbourhood statistics web site, part of the
ONS web site, for ideas regarding formats of downloadable material and other possible
items of good practice.
 The syndicate also thought that the web site should encourage user feedback and include
a section on Frequently Asked Questions (FAQs). This latter item might also be
replicated on the (DfT hosted) TRS2 web site
8. Group D - The Devil’s in the Detail
Chair: Tom Worsley
Scribe: Frank Dixon
Demands for traffic statistics vary enormously from broad aggregate statistics to detailed
requests relating to individual road links, and breakdowns by vehicle type and time periods.
Clearly, the robustness of the data suffers at the more disaggregated levels, particularly
affecting the less common vehicle types.
Key points from the discussion:
 On classification of HGVs, the ATCs deliver what is needed for most purposes.
 However, there was a need for different strands of information, for example gross
vehicle weight and links to the VOSA vehicle definitions.











If ANPR cameras were also being deployed there could be links to DVLA categories and
speeds in specifics lanes related to flows. Whether these strands could be addressed
solely by the National Traffic Census was uncertain, but it was important to recognise
these potential linkages.
Car traffic was generally satisfactory. However, if linkage could be made to the DVLA
database, then make- and model-specific traffic estimates could be made and be used as
measures of exposure in calculating traffic accident rates. A research study could explore
this. Linkage to the DVLA could also establish the sizes of cars used at different times
during the week
More details on bus sizes would be useful as a guide to bus loadings.
On geographical detail, there was some concern that the ‘rural’ definition was too coarse.
There were probably too few ATC counts in rural areas
Similarly there were concerns about the adequacy of the ATC coverage for urban areas,
all of which had distinct characteristics.
ATCs were probably a cost-effective way of obtaining local traffic information but they
could not provide the full picture.
Traffic data should be available by 15-minute periods. Manual counts should be used for
this.
There were mixed views on quarterly data: some wanted them and used them
extensively; others only used the annual figures.
There was concern that caveats attached to figures were often removed or downgraded
when the data were assembled for publication by organisations outside the DfT.
Confidence limits should be shown so that users know how to treat the results. Tighter
limits could be obtained if the data from other sources were used to boost the sample.
LAs need guidance on standard counting practice.
9. In the plenary discussion following the groups’ reports the following points were made:
i. Rather than displaying 95% confidence limits, it might be best to show standard errors,
together with guidance on how to construct confidence limits.
ii. Given that LA and HA traffic counting might often be ad hoc, some means needed to be
found to integrate these counts within the national system without biasing the results.
This might be straightforward for the major road AADFs but more complex for the
minor road system.
iii. It was assumed that the traffic counts could not per se handle vehicle occupancy.
However this would be useful and the possibility of roadside occupancy counts could be
considered.
iv. There was a general consensus for greater integration and standardisation of traffic data
across the network.
v. Likewise there was a consensus that the opportunities offered by new technologies
needed to be explored as part of the review. The recommendations could point the way
to useful research rather than early implementation.
10. Finally the chairman again reminded participants that it was essential that they completed
the questionnaire as soon as possible and returned it before the end of November to the DfT.
USER QUESTIONNAIRE
10. User Profiles
Profiles of Users: Questionnaire vs Enquiries
Database vs Bulletin mailing lists
45.0
40.0
35.0
30.0
25.0
20.0
15.0
10.0
5.0
0.0
Questionnaire
responses
Enquiries Database
Other (includes
Police for Quality
Review)
Consultants
Commercial
Academic/Students
Local Authorities
Other Gov
Departments
Other gov regional
bodies
Annual and Quarterly
Bulletin mailing lists
DfT (Central &
Agencies)
Percentage of Total
When the profile of the users who responded to the questionnaire was compared to
that of the users who request data directly from the branch, it was reassuring to note
that the profiles are similar except in the case of Commercial users where there is a
far higher percentage of commercial responses than enquiries. The profile of users
who are on the Annual and Quarterly Bulletin mailing lists was then added to this
comparison and it can be seen that the general profile remains the same except for the
very high percentage of DfT recipients and the very low percentage of Consultants.
Both these differences seem reasonable as it would be expected that many recipients
from within DfT would receive the bulletins as a matter of protocol. It is also makes
sense that consultants would request data as and when it was required rather than
receive the bulletins regularly.
Categories of Users
Analysis of User Questionnaire
Introduction
11. As part of the user consultation, 240 questionnaires were sent out by e-mail to individuals
and organisations believed to be users. Further copies were sent to user groups such as the
Transport Statistics User Group which distributed them on to their members. A further 50
copies were sent out by post including to members of the Annual Bulletin mailing list. In
addition a number of questionnaires were circulated to Local Authorities in Scotland. Thus
the total number of questionnaires sent out was believed to be around 400.
12. The questionnaire was also made available on the DfT website, in both Excel and PDF
format, for users to complete.
13. There were two e-mail reminders sent out to those on the consultation list.
Purpose
14. The aim of the questionnaire was to identify key requirements from Users.
Recommendations were made on requirements that were deemed to be within the scope of
the Review. These recommendations were passed onto the appropriate work package team
leaders. Those requirements that were out of scope were passed to appropriate colleagues
within the department where possible
15. A comprehensive analysis of the results of the questionnaire can be found in Annex 4.1.A
and the key findings are highlighted below. The questionnaire itself can be found at Annex
4.1.B.
Findings:
CURRENT AND FUTURE REQUIREMENTS
16. Breakdown of responses - 85 responses were received. Of these, the majority of the
responses were from Local Authorities (31) who comprised 36% of the total. The responses
were analysed by the following groups (strata) of respondents:
DfT - central,
DfT - agencies,
Other Government Departments,
Consultancies,
Miscellaneous #1 - Organisations categorised as 'other' in the user
questionnaire
Miscellaneous #2 - Other government regional bodies (e.g. Regional
Development Agencies), Academic, and
Commercial
were grouped together due to the
small number of responses
from these users.
17. Question 1
Use of Road Traffic vs. Road Length Statistics - The vast majority (92%) of respondents
use the Road Traffic Statistics (RTS) while 61% use the Road Lengths Statistics (RLS).
Similar proportions were evident in each stratum analysed.
18. Question 2
Frequency of use of statistics - 46% of respondents use the statistics at least once a month,
while 66% use them at least every quarter, emphasising the statistics' importance to a wide
range of users.
19. Question 3
Disaggregated Statistics by Road Type and Urban/ Rural split –
The majority of respondents for both Road Traffic Statistics and Road Length Statistics will
have the same requirements in the future as they do currently.

(Question 3.a) For the level of disaggregation of RTS by road type, there is a fairly even
demand across all of the requirements - for both current and future uses. Demand is
greater for Urban/Rural than Built-up/Non built-up splits. Small changes only were
recorded between current and future uses except for Individual road Links with future
demand for the latter 350% of the current level. This is driven by the LAs and
Consultants, with 5 and 3 responses of the 10 extra responses respectively. Overall the
numbers do not suggest any significant increase in demand for our statistics in the future.

(Question 3.b) For the RLS by road types it seems as if the demand for statistics is
currently centred on Principal A roads, Trunk A roads and motorways, followed by B, C
and unclassified roads, and from there tailing off to individual road links which in itself
is not insignificant. The main change from Current to Future use occurred for
Urban/Rural splits, with 8 and 9 more respondents requiring Urban and Rural splits
respectively. 7 more respondents will require individual road links in future than now.
Overall the numbers suggest an increase in demand for RLS across all the levels of
disaggregation.

For both the RTS and the RLS there was an unexpectedly high demand for the statistics
to be split by Built Up/ Non Built Up definitions. DfT is moving towards splitting the
estimates by the Rural/ Urban Polygon definitions. The demand for the Built Up/ Non
Built Up statistics by user is detailed in the table below. It centres on LAs and
Consultancies. It was not clear from the free text answers why this demand exists and so
a number of respondents were contacted. These subsequent enquiries indicated that the
responses were not, in fact, a true reflection of the actual (lower) demand.
RTS
Local Authority
Consultancy
Other Government
Other
Police
DfT Central
Regional
Government
DfT Agencies
Commercial
Academic
RLS
19
8
5
5
4
3
14
7
3
3
1
2
2
2
2
0
50
2
0
0
0
32
20. There is a high demand for both the RTS and the RLS to be split by all the road types that
were listed and as such none can be discarded.
Proposed Way Forward: Local Authority Demand for Built Up/ Non Built Up statistics to
be investigated.
Other results to be passed to team leader of WP7- Demand for Disaggregated Statistics.
Note: The demand for the BU/NBU statistics was investigated for the RLS and the
following results obtained:
Out of the 32 respondents:
15
Were unavailable to contact
6
Answered the question by mistake or said that their answer is no longer
correct
4
Said that that data supplied in the BU/NBU split would be easier to
analyse
3
Said that such data would be useful for future comparisons
2
Mentioned that they used the data for the Revenue Support Grant
2
Said that their users requested BU/NBU data from them
From those that were contactable: a third of the respondents had made a mistake in
answering the question and did not in fact require the BU/NBU split and from the answers
of the remainder the demand for these statistics is far less that initially perceived. None of
the contacted respondents said that the BU/NBU split is essential to their work.
It may be that the need for the BU/NBU split is being perpetuated by the fact that SR2 still
supplies data in this format to Local Government Finance (LGF). Should LGF decide to
move to the Urban/ Rural split then the small demand that currently exists would in all
likelihood disappear.
21. (Question 4)
Disaggregated Statistics by time periods - A significant proportion (32%) of respondents,
including ten LAs and five DfT Central, require traffic estimates for shorter time periods.
Of these, a substantial number suggested they would find both hourly and daily data useful.
64% of respondents do not require shorter time periods.
Proposed Way Forward: Results to be passed to team leader of WP7- Demand for
Disaggregated Statistics.
22. (Question 5)
Demand for Seasonally Adjusted Data - The overwhelming majority (92%) of
respondents do not specifically use the seasonally adjusted data. Of those 4 respondents
who do, only one was specific as to the use of the data stating that seasonality has road
safety implications due to varying light and weather.
Proposed Way Forward: In view of the overwhelming lack of interest in seasonally adjusted
data it may be worth considering whether or not it is worth calculating the adjusted figures.
Results to be passed to team leader of WP12.c - Dissemination
23. (Question 6)
Vehicle type disaggregation - There is a high demand for all the vehicle types listed to be
supplied separately, with the lowest demand being for HGV rigid and articulated (28
responses, 33% of all respondents). All the respondents who wanted the HGV total to be
split required both articulated and rigid totals. Of particular interest is the high demand for
the Bus and Motorcycle figures (currently combined in the Quarterly Bulletin) to be
published separately (62% and 60% respectively). Proposed Way Forward: Bus and
motorcycle figures to be published separately. Results to be passed to team leader WP9.b Quarterly Estimates Issues
24. (Question 7)
HGV disaggregation - Of those who require the HGV estimates separately, 18 of these
require these estimates to be further split by number of axles. This is 21% of all the
respondents and as such is a significant requirement.
Proposed Way Forward: Published HGV estimates to be further split by number of axles.
Results to be passed to team leader WP9 - Methodological Issues
25. (Question 8)
Uses of RT and RL statistics - There were 81 responses to this question. The uses can be
classified into seven broad categories:
 Monitoring/understanding traffic levels/trends including for LTP purposes - 37
responses (of which 20 were Local Authorities)
 Calculate/ inform casualty statistics - 14 responses
 Specific policy/ enforcement purposes - 11 responses
 Traffic/ transport modelling – 8 responses
 Manage/ maintain the Road network - 6 responses
 Calculate/ inform other (non-casualty statistics) - 4 responses
 Commercial use - 2 responses
40
100%
90%
35
80%
30
Frequency
70%
25
60%
20
50%
40%
15
30%
10
20%
5
10%
0
s
ng
tat
ori
sS
nit
o
a
C
M
0%
licy
Po
f.
Tra
t
en
s
ta t
m
sS
ge
a
a
mm
n
n-C
Co
Ma
o
N
Cumulative frequency as a percentage
Pareto Analysis of the uses of RT and RL statistics
ial
e rc
As demonstrated by the Pareto Analysis above it is clear to see that over 80 % of the
suggestions fall into the first four categories.
Proposed Way Forward: Efforts to improve the statistics should be informed by these uses.
This result to be passed to all work package team leaders.
26. (Question 9)
Additional statistics - 31 respondents identified 44 additional requirements, including
some respondents who identified more than one requirement. The comprehensive list of
requirements with a decision on whether they are in or out of scope can be found in Annex
4.1.C. Many of the requirements depend on being able to match ANPR data to DVLA
records e.g. Vehicle make.
Proposed Way Forward: Should DfT be able to match ANPR data to DVLA records it will
be necessary to reassess the requirements subject to this constraint. At this stage they have
been judged to be out of scope. The other requirements judged to be in scope will be passed
on to the relevant WP team leaders.
27. (Question 10)
Alternative sources of statistics - 49% of respondents indicated they are aware of other
sources. A wide variety of sources were highlighted although most were presumably to
supplement DfT data. Sources of similar or potentially useful data for DfT identified are the
Camera Partnership, information collected by Bus Operators, the Scottish Roads Traffic
Database and Traffic Master
Proposed Way
Forward: Results to be passed to team leader WP3 - Investigating existing data sources:
Highways Agency and other non-LA sources.
ACCESSING THE STATISTICS
28. (Question 11.a)
Use of publications - Most (86%) respondents obtain their data from the Annual Transport
Statistics Bulletin with 59% using the Quarterly Bulletin. 36% obtain statistics through adhoc queries while 19% said that they used other publications. The other most frequently
mentioned publications are Transport Statistics Great Britain and Road Casualties Great
Britain, both produced by DfT. There were no publications recorded that had not been heard
of and needed investigating.
Proposed Way Forward: Sufficient demand to justify continued publication of both
Quarterly and Annual Bulletins.
29. (Question 11.b)
Frequency of publication of Annual and Quarterly Bulletins - 96% of all those that
responded (72) felt that the frequency of publication was satisfactory. The remaining 4% of
respondents asked for more prompt publication.
Proposed Way Forward: Sufficient evidence to justify keeping frequency of publication the
same.
30. (Question 12.a)
Obtaining the statistics - There appears to be a fairly even spread between the different
media and methods of obtaining the statistics, with a
majority (76%) of respondents obtaining their statistics from the DfT website. It is also
worth noting that 58% of respondents also obtain the statistics from paper publications. 19
respondents said that they obtained road traffic/length statistics from a source other than
those listed explicitly in the question. However, it appears that this question was
misinterpreted by most respondents and no significant other sources were identified.
31. 12.b Publication format - A clear majority of users (over 80%) feel that the publications
are both clear and easy to read and easily comprehensible. There were however a number of
comments on how improvements might be made, these have been listed in the text box for
Question 12.b.
Proposed Way Forward:
List of comments to be passed to team leader WP12.c – Dissemination
32. (Question 12 c and d)
Website accessibility - 88% of respondents are aware that the statistics are available on
both the DfT and ONS websites. What is apparent is that there are some difficulties in
accessing the publications on the website. 10 respondents described the problems that they
encountered. 8 of the responses suggest that it is difficult to navigate through the website,
that the search facility is poor and that this can make it difficult to locate the data. One of
the respondents requested that the tables from the bulletins are provided in an Excel/Word
format and one mentioned connection difficulties. All the above suggest that there are
improvements to be made in this area.
Proposed Way Forward: The tables are already provided in Excel format and it is not felt
necessary to additionally provide them in Word format. The other concerns to be forwarded
to the web publishing team SLAM5.
33. (Question 12.e)
Telephone enquiries - Of those who use this mechanism it appears that the process is
satisfactory.
34. (Question 12.f)
Email/letter enquiries - There were 4 respondents out of 30 who felt that the time taken to
process the request was not satisfactory. However 26 respondents were satisfied with the
time taken and on balance it appears that overall the process is satisfactory in terms of time
taken, price and statistics supplied.
35. Contact point - 98% of respondents who have attempted to contact the enquiry point
succeeded in doing so, with only one respondent unable to get through. This is highly
satisfactory.
36. Overall it appears that the main concern for accessing data is through the website, whereas
all the other means of access are satisfactory.
OTHER ISSUES
37. (Question14.a)
Note describing how the National Traffic estimates are made - Only 54% of respondents
were aware of the note's existence and availability. .Of these, only half had tried to obtain
the note and nearly all were successful. Overall this result seems satisfactory as not all
respondents would need this information and if they did require the information they would
be able to make an enquiry which should be resolved satisfactorily.
38. (Question15.a)
Quality of RTS
 15% of respondents who answered this question (80 responses) "Strongly Agree"
that the RTS have a level of accuracy appropriate for their purposes, while 50%
"Agree". 23% “Neither Agree nor Disagree”, 10% “Disagree” and just 3% “Strongly
Disagree”. It is interesting to note that it is the “Other government regional bodies",
Local Authorities and Consultants who have the most neutral responses. This may
be because users have no way of knowing how accurate the figures are. It is
envisaged that the provision of quality measures may provide a solution to this in the
future.
 From the text answers that were received the most common reason for doubting the
SR2 statistics appears to be issues with the methodology including the sample size
and revisions. There were some who raised issues regarding the minor roads figures,
some the accuracy of the statistics at LA level and lastly those who questioned the
accuracy of pedal cycle stats and the lack of quality measures.
Proposed Way Forward: The concerns about methodology, sample size, minor roads and
Local Authority level statistics to be passed on to team leader WP9 – Methodological Issues
(Annual Estimates Issues and Sampling). Concerns over pedal cycle statistics to be passed
on to team leader WP14 – Cycling.
The demand for quality measures to be passed on to team leader WP10 - Quality Measures/
Quantifying Uncertainty
39. (Question 15.b)
Quality of RLS
 66 respondents answered this question, 9% of those "Strongly Agree", while 33%
"Agree". 47% are neutral while 11% "Disagree". No one "Strongly Disagreed".
Again it is the "Other government regional bodies", Local Authorities and
Consultants who have the most neutral responses.
 With respect to the LA responses, of those who responded 14 agreed (of which 2
"Strongly Agreed") and 10 were neutral and 1 "Disagreed".
 From the five text answers received the main reasons for doubting the quality of the
RLS related to the methodology used to calculate the RLS and the frequent revisions
to the RLS.
Proposed Way Forward: It is felt that the response to this question indicates that efforts
need to be made to improve the confidence of users in the RLS.
This recommendation along with these results and those from the text answers to be passed
on to the team leader WP11 - GIS and Road Network.
40. (Question 16)
Opinions on revisions - A total of 40 responses were given. Most of the respondents regard
revisions as inevitable but 7 explicitly refer to the problems and confusion such revisions
cause. 10 respondents support the revisions to varying degrees. A further 10 respondents
suggest that an accompanying explanation and notification of any revision should be made,
8 suggest consistency can become an issue and 4 suggest users should be consulted about
revisions. There were also suggestions for measures of quality, the legacy time series to be
made available and revisions to be backdated.
12
100%
90%
10
80%
70%
Frequency
8
60%
6
50%
40%
4
30%
20%
2
10%
ke
r
Q
ui
c
ea
su
re
s
Q
ua
l
ity
M
St
at
us
t
Co
ns
ul
Pr
ob
le
m
s
te
nc
y
Co
ns
is
Su
pp
o
Ex
pl
a
rt
0%
na
t io
n
0
Cumulative frequency as a percentage
Pareto Analysis of opinions on revisions
As can be seen from the Pareto Analysis above more than 80% of the comments were made
from just 4 categories. If we remove the category “Support” then it can clearly be seen, in
the chart below, that just less than 80% of the comments were made in the three categories:
Explanation, Consistency and Problems.
12
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
8
6
4
2
ke
r
Q
ui
c
ea
su
re
s
ity
M
St
at
us
Q
ua
l
t
Co
ns
ul
Pr
ob
le
m
s
te
nc
y
Co
ns
is
na
t io
n
0
Ex
pl
a
Freqency
10
Frequency as a cumulative
percentage
Pareto Analysis of opinions on revisions without support
category
Proposed Way Forward: Efforts are concentrated on addressing the comments in the three
categories: Explanation, Consistency and Problems. The original comments made in each
category to be passed to the team leaders WP9 - Methodological issues and WP12 –
Dissemination.
41. (Question 17)
Other issues - There was an overwhelming request for quality measures for the statistics,
specifically cited in (21 out of 28) 75% of responses. A further two raised the timeliness of
the output, one suggested more detailed explanations of the methodology, one requested
information on the sources used, one suggested that there are issues with minor roads and
one requested forecasting traffic flows figures.
Quality Measures – addressed by WP10 - Quality Measures/ Quantifying Uncertainty
Timeliness – As this was requested by only one user, and the majority of other users were
satisfied with the timeliness of the publications, it has been decided not to pursue this
requirements further due to limited resources.
Detailed explanations of the methodology – This is already available and mentioned in all
publications
Information on the sources used – This information is freely available on request from the
branch but a note will be added to the Annual and Quarterly publications.
Issues with Minor roads - addressed by WP 9 Methodological issues, Annual estimates
issues.
Forecasting Traffic flows - This is out of scope but it is worth noting that ITEA produce
National Road Traffic Forecasts.
Proposed Way Forwards:
 Pareto analysis confirms that requirement for quality measures to be passed to team
leader WP10 - Quality Measures/ Quantifying Uncertainty. Other requirements
either already addressed or judged to be out of scope due to lack of demand.
Pareto Analysis of other requirements
100%
90%
20
80%
Frequency
70%
15
60%
50%
10
40%
30%
5
20%
10%
Fo
re
ca
st
M
in
or
R
oa
ds
og
y
et
ho
do
l
M
So
ur
ce
s
m
Ti
Q
ua
l
el
in
es
s
0%
ity
0
Frequency as a cumulative
percantage
25
Conclusions
42. It is believed that most of the major User requirements had been anticipated and considered
in the work already undertaken. However additional user requirements identified from the
analysis have been passed to the relevant Work Package leaders to take into consideration.
LB
April 2005
Annex 4.1.A
Department
for
Transport
Road Traffic and Road Length Statistics Quality Review
National Statistics User Consultation: Questionnaire
The tables below show the breakdown of the questionnaire responses. The first column shows the totals of all 85 responses. There then follows eight categories as
listed below. Three of the categories on the original questionnaire have been combined in this analysis as there individual results were felt to be too few to provide
significant data.
A number of questions show the number of respondents who did nt answer the question or for who it was not applicable.
What type of organisation do you work for (or what type of work are you involved in)?
1. Department for Transport (DfT) - central
9
2. Department for Transport (DfT) - agencies
5
3. Other Government Departments
12
4. Local Authority
31
5. Consultancy
10
6. Police
5
Misc. # 1: Organisations categorised as 'other' in the user
7. questionnaire
5
Other government regional bodies (e.g. Regional
Development Agencies)
3
Academic
3
Commercial
2
Total
YOUR CURRENT AND FUTURE REQUIREMENTS
*One response gave two differing time periods presumably one for use of RTS and one for use of RLS
Misc. # 1
Misc. # 2
Consult.
Police
DfT Ag.
Oth Gov.
Misc. # 1
Misc. # 2
3
2
3
0
1
0
0
Consult.
Police
18
21
17
12
9
3
6
LA
2 Generally how often do you use the statistics? About:
Every week
Every month
Every quarter
Every six months
Every year
Less than once a year
No answer given
3 7 10 3 3 2 2
1 0 2 1 0 0 2
1 4 19 6 2 3 5
DfT Cen.
3
0
6
All
33
6
46
DfT Ag.
Oth Gov.
All
DfT Cen.
Number of Responses
1 Which of the following statistics do you use?
Road traffic statistics Only
Road length statistics Only
Both
85
LA
8. Misc. #2 :
0
3
0
0
1
1
0
7
7
5
5
5
0
2
1
3
3
1
0
1
1
1
2
1
0
0
1
0
2
2
3
2
2
0
0
3
1
1
0
0
0
0
1
1
1
4
0
0
3
3a If you use the road traffic statistics, please indicate which of the following levels of disaggregation of our road traffic
statistics you currently use and which you may require in the future for Road Traffic Estimates?
25
36
37
34
41
44
46
4
1
3
3
5
4
4
1
0
2
1
3
3
0
0
4
7
6
2
3
4
13
18
10
9
11
14
18
2
6
7
6
9
9
9
1
2
0
0
2
2
2
Misc. # 1
Misc. # 2
0
0
1
0
1
1
0
0
0
0
0
0
0
1
5
4
4
3
5
5
4
4
2
5
5
4
0
0
2
0
0
0
0
0
0
0
0
0
0
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
2
0
0
1
1
1
1
1
2
2
1
1
3
1
0
1
1
1
1
1
1
1
2
1
0
2
2
1
2
2
2
2
2
2
2
1
1
1
1
2
3
2
2
2
3
4
6
6
7
7
7
14
10
9
5
10
8
9
1
1
2
0
0
0
0
0
0
0
1
1
1
1
0
1
1
0
2
1
1
5
3
3
1
2
1
2
3
2
0
0
1
1
1
2
0
0
0
1
1
1
0
0
1
1
1
1
1
Misc. # 2
Misc. # 1
Police
Consult.
LA
Oth Gov.
DfT Ag.
DfT Cen.
All
Misc. # 2
Future Only
Misc. # 1
Police
Consult.
LA
Oth Gov.
DfT Ag.
DfT Cen.
All
Both Current and Future
Consult.
Police
1
2
1
3
3
2
2
2
2
2
2
2
1
LA
4
14
14
9
11
13
12
11
11
9
13
13
9
Individual Road Links
Local Authority Level
English region(s) or Wales or Scotland
Great Britain as a whole
Motorways
Trunk A roads
Principal A roads
B Roads
C Roads
Unclassified Roads
Urban areas
Rural areas
Built up/ non-built-up (based on speed limit)
DfT Ag.
Oth Gov.
All
DfT Cen.
Current Only
3
3
2
2
2
2
2
0
0
0
2
9
14
2
2
1
1
0
0
0
6
10
6
0
0
0
0
0
0
4
4
2
0
1
1
0
0
1
1
0
0
0
0
0
0
0
0
0
0
1
Misc. # 2
1
1
1
0
0
1
2
3
1
0
2
2
2
1
1
1
1
1
1
Misc. # 2
Misc. # 1
Police
0
0
0
0
0
0
0
0
0
0
0
0
0
2
2
2
1
1
0
Misc. # 1
0
0
1
0
0
0
1
0
0
0
0
0
0
Police
0
1
0
0
1
2
2
2
1
0
2
2
2
Consult.
LA
Oth Gov.
0
0
0
0
0
0
0
0
0
0
1
1
0
14
13
12
14
15
15
6
5
5
5
2
4
Consult.
0
0
1
0
1
1
0
0
0
0
0
0
0
2
1
1
3
3
2
LA
1
1
0
1
2
2
2
2
2
2
1
1
0
2
2
2
2
2
2
DfT Ag.
6
6
5
7
7
7
Gov.
Misc. # 2
Misc. # 1
Police
Consult.
LA
Gov.
DfT Ag.
All
DfT Cen.
3
1
All
3
5
4
2
6
7
7
5
4
3
5
5
3
Individual Road Links
Local Authority Level
English region(s) or Wales or Scotland
Great Britain as a whole
Motorways
Trunk A roads
Principal A roads
B Roads
C Roads
Unclassified Roads
Urban areas
Rural areas
Built up/ non-built-up (based on speed limit)
15
26
16
14
13
12
10
7
DfT Ag.
3b If you use the road length statistics, please tell us which of the following levels of
dissaggregation you currently use and which you may require in the future?
2
2
2
3
3
1
DfT Cen.
0
0
0
0
1
1
DfT Cen.
2
2
2
3
3
2
All
36
32
30
35
31
26
3
0
1
1
1
3
3
2
5
6
7
6
7
7
1
0
0
1
1
1
1
0
0
1
1
1
1
1
1
0
0
1
2
2
0
2
2
2
5
4
7
9
9
7
7
7
13
14
5
1
0
0
0
0
1
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
2
2
1
1
1
2
3
0
3
2
1
2
2
2
2
3
6
6
3
Currently, most of the traffic estimates are made for annual (and quarterly) intervals.
Do you require estimates in respect of shorter time periods, such as hourly, daily,
weekly or monthly?
Yes
No
No answer/Not Aplicable
27
54
4
0
0
1
1
1
1
1
0
1
1
1
0
0
1
1
1
1
1
1
1
1
0
5
4
0
0
0
1
1
1
1
1
1
1
1
1
2
3
0
0
2
3
2
2
0
0
0
2
2
0
1
9
1
10
19
2
3
6
0
1
4
0
1
4
0
Misc. # 2
2
3
6
7
7
6
5
5
5
2
4
Misc. # 1
1
2
2
2
2
1
1
1
2
2
1
Police
0
0
1
1
1
1
1
1
1
1
1
Consult.
5
4
7
7
6
6
6
5
4
4
6
LA
7
6
9
9
14
14
13
12
7
6
8
Oth Gov.
1
1
3
2
2
2
2
2
1
0
2
DfT Ag.
0
0
0
0
0
0
0
0
0
0
0
All
4
2
3
3
3
3
2
2
2
3
3
2
DfT Cen.
18
19
31
31
35
32
30
28
23
18
24
4
5
1
If "Yes", please specify what time periods you require statistics for and why.
Conclusions from free text answers
Misc. # 2
Misc. # 1
Police
Consult.
LA
Oth Gov.
Do you specifically use the seasonally adjusted series which we currently provide?
DfT Ag.
5
DfT Cen.
All
A substantial number of respondents suggested they would find both hourly
and daily data useful for reasons including seasonal variability analysis,
casualty related statistics, congestion monitoring and commercial uses. There
were also responses requesting weekly figures, a high number requesting
monthly figures and other responses for 15-minute and peak periods.
Yes
No
No answer
4
78
3
0
9
0
0
4
1
1
10
0
3
27
1
0
10
0
0
5
0
0
4
1
0
9
0
If "Yes", please explain why seasonally adjusted rather than unadjusted figures are used.
Conclusions from free text answers
Misc. # 2
Misc. # 1
Police
Consult.
LA
Oth Gov.
DfT Ag.
DfT Cen.
4
3
2
1
1
2
3
1
4
4
3
3
1
1
2
2
3
3
4
4
5
2
2
4
4
7
5
59
4
4
7
23
6
2
4
9
Misc. # 2
9
9
10
4
4
5
5
9
5
Misc. # 1
26
23
27
15
15
24
25
24
23
Police
7
5
7
2
2
6
6
8
6
Consult.
4
4
4
1
1
4
4
4
4
LA
Do you require the HGV estimates (as you have indicated above) to be
further split by number of axles?
7
6
6
2
2
6
2
5
3
Oth Gov.
7
65
57
64
28
28
53
51
61
53
DfT Ag.
Cars
Light Vans
Heavy Goods Vehicles (HGVs) - total
HGVs - Rigid only
HGVs - Articulated only
Buses
Two wheeled motor vehicles
All Motor Vehicles
Pedal Cycles
All
Traffic
DfT Cen.
Please indicate which of the following vehicle types (or categories) you require
separate traffic estimates for:
All
6
All
While four people indicated they did use the seasonal adjusted series only
three reasons were given. Of these only one was specific as to the use of the
numbers stating that seasonality has safety implications due to varying light
and weather
Yes
No
No answer/Not Apllicable
8
18
57
10
0
7
2
0
5
0
1
8
2
11
18
2
4
0
0
0
3
2
1
2
2
1
14
0
Please describe what you use or will use road traffic and road length statistics for, including the
requirements indicated in response to questions 3 to 7.
Conclusions from free text answers
Yes
No
No answer/Not Apllicable
If "Yes", please describe briefly what you require and why
Conclusions from free text answers
Not all respondents answered this questions of those who did 47% required
some additional stats, of these some were clearly out of scope. A list of the
balance of the requirements follows:
Motorcycle engine capacity
Make & Model
Vehicle/Axle Weights
31
35
19
2
5
2
1
2
2
5
4
2
10
13
8
4
0
1
1
3
1
2
2
1
Misc. # 2
Misc. # 1
Police
Consult.
LA
Oth Gov.
DfT Ag.
Are there any other statistics/analyses/data that you would use if we were
to provide them?
DfT Cen.
9
All
There were 81 responses to this question. The uses were divided into 5
categories:
Traffic, transport modelling, monitoring/understanding traffic levels/trends
including for LTP purposes 37
Specific policy/enforcement purposes 11
Calculate/ inform casualty statistics 14
Calculate/ inform other (non-casualty statistics) 5
Management/ maintenance of the Road network 6
Commercial use 2
6
6
2
Yes
No
No answer/Not Apllicable
42
37
6
2
6
1
3
2
0
5
5
1
14
15
2
6
4
0
4
1
0
1
3
1
Misc. # 2
Misc. # 1
Police
Consult.
LA
Oth Gov.
DfT Ag.
Do you use or are you aware of other sources of statistics that meet your current
needs?
DfT Cen.
10
All
Smaller MGV
Company Cars
Fuel type
London Borough Disaggregation 2
Pedal Cycle usage
Goods Vehicles broken into MGV and over 7.5T
Disagg. By Unitary Authority
Data on B, C and U roads
LA Disaggregation
Separate figures for Scotland
Emissions data
Regional flows
International Freight (country of owner)
Public transport data
7
1
1
If "Yes", what are the sources?
11.a
Do you obtain your information from:
Misc. # 2
Misc. # 1
Police
Consult.
LA
Oth Gov.
DfT Ag.
All
ACCESSING THE
STATISTICS
DfT Cen.
Conclusions from free text answers
A wide variety of other sources were highlighted though most were presumably
to supplement our data. Some sources of similar or pontentially useful data wer
the Camera Partnership, information collected by Bus Operators, the Scottish
Roads Traffic Database and Traffic Master
i. Quarterly Transport Statistics Bulletin: Traffic in Great
Britain
ii. Annual Transport Statistics Bulletin: Road traffic Statistics
iii. Other Publications
iv. Ad hoc queries
v. Other
50
73
25
31
16
5
7
5
7
1
3
3
1
1
0
7
8
3
4
3
16
29
9
9
11
6
9
5
3
0
4
5
0
2
1
3
5
0
2
0
6
7
2
3
0
Conclusions from free text answers
A number of other publications were identified which could be catergorised into
three groups; SR Publications, DfT Publications and Other. By far the most
common were DfT publications which included TSGB and RCGB which were
mentioned on a number of occasions. There were then more SR Publications
identified than other.
If you answered "i" or
"ii", please go to
Question 11b.
Need less frequent publication
Frequency is satisfactory
Need more frequent
publication
*Not all people responded to part b even when it was apllicable to their previous answer
Please explain your answer
Conclusions from free text answers
Misc. # 2
Misc. # 1
Police
Consult.
LA
Oth Gov.
DfT Ag.
If you answered "i" or "ii", does the frequency of publication meet your current
needs and future requirements?
DfT Cen.
11.b
All
If you did not answer "i"
or "ii", please go to
Question 12a.
1
69
0
6
0
4
0
8
0
27
0
9
0
5
0
4
1
6
2
0
0
0
1
0
0
1
0
i. Paper Publication
ii. DfT Website
iii. ONS Website
iv. Telephone
enquiry
v. Email enquiry
vi. Personal contact with someone within DfT
vii.
Letter
viii.
Other
Misc. # 2
Misc. # 1
Police
Consult.
LA
Oth Gov.
How do you obtain the statistics?
DfT Ag.
12.a
DfT Cen.
All
There were few responses that were not currently satisfied with the frequency
of our publications. The most notable comment albeit from just three
respondents, did not regard the frequency. These resonses requested a more
promt publication. Other comments regarded the usefulness of the Quarterly
Bulletin with two respondents suggesting it wasn't necesary and one response
indicating that it was vital.
49
65
14
5
6
1
2
2
0
6
6
1
23
26
7
4
9
2
3
5
0
3
5
1
3
6
2
14
29
25
2
5
6
0
0
0
4
6
4
6
13
10
2
2
0
0
1
1
0
1
0
0
1
4
2
0
0
1
1
0
0
0
0
19
3
0
1
5
2
1
0
7
Conclusions from free text answers
No other significant sources were identified in this question
Consult.
Polic
e
Misc. #
1
Misc. #
2
LA
If you have used any of the publications (whether paper copy or Internet
version found on the DfT website)
DfT Ag.
Oth
Gov.
12.b
All
DfT
Cen.
The following questions are dependent on wether or not users use a particular method and therefore we do not
expect the number of responses to total 85. We have excluded the no/answer row from our tables.
Is the publication clear and easy to read
?
Yes
No
74
2
8
0
4
0
9
0
28
0
9
1
4
1
5
0
7
0
70
3
8
0
4
0
8
0
26
0
8
2
4
1
5
0
7
0
Is it easily comprehensible?
Yes
No
How could it be improved?
Conclusions from free text answers
11
0
26
3
10
0
4
1
5
0
Misc. # 2
Misc. # 1
Police
Consult.
LA
Oth Gov.
DfT Ag.
3
0
7
3
Consult.
Polic
e
Misc. #
1
Misc. #
2
If you obtained statistics from the DfT website were you able to obtain them
easily?
9
0
LA
12.d
75
7
DfT Ag.
Oth
Gov.
Yes
No
DfT Cen.
Were you aware that the statistics published by DfT are available on the
DfT and ONS websites?
All
DfT
Cen.
12.c
All
The responses fell into three distinct catergories regarding the format, contents
and web related issues. Seven comments on the content of the publications
proved the most common. There were then six issues regarding the online
publication and a further three who had suggestions on the format. A list of the
problems and suggestions will follow.
Annual Bulletin
Yes
No
57
6
7
1
2
0
5
2
20
1
7
1
4
1
5
0
7
0
Yes
No
42
6
4
2
1
0
4
1
15
1
5
0
3
2
4
0
6
0
25
9
0
3
0
0
3
0
13
3
0
1
3
2
2
0
4
0
Quarterly Bulletin
Local Authority Traffic Estimates
Yes
No
If "No" for any of the above, please describe any problems you encountered.
Conclusions from free text answers
Misc. # 2
Misc. # 1
Police
Consult.
LA
Oth Gov.
DfT Ag.
If you obtained statistics by telephone enquiry:
DfT Cen.
12.e
All
Eight of the ten responses suggest the navigation of the web site including the
search facility is poor and this can make it difficult to locate the data. Two other
comments include one who requested the provision of Excel files, which has
been mentioned in previous questions and another who had previously had
connection difficulties.
Were you satisfied with how long it took to process your request?
Yes
No
21
1
3
0
0
0
4
0
10
1
3
0
0
0
1
0
0
0
8
0
0
0
7
1
0
0
0
Did you feel the price you paid was reasonable?
Yes
No
0
0
0
0
0
0
0
0
0
19
0
1
0
0
0
4
0
11
0
3
0
0
0
0
0
0
0
Misc. # 2
Misc. # 1
Police
Consult.
LA
Oth Gov.
If you obtained statistics by email or letter enquiry:
DfT Ag.
12.f
DfT Cen.
Yes
No
All
Did the statistics provided meet your needs?
Were you satisfied with how long it took to process your request?
Yes
No
26
4
5
1
0
0
6
1
11
1
1
0
0
0
2
0
1
1
12
0
0
0
0
0
2
0
8
0
0
0
0
0
1
0
1
0
24
0
3
0
0
0
6
0
12
0
1
0
0
0
0
0
2
0
Did you feel the price you paid was reasonable?
Yes
No
Yes
No
Not applicable
13.b
If "No" please explain briefly what happened.
Conclusions from free text answers
40
1
44
6
0
3
2
0
3
6
0
6
17
0
14
4
0
6
0
0
5
2
1
2
Misc. # 2
Misc. # 1
Police
Consult.
LA
Oth Gov.
Have you successfully reached the contact point within the Department for
queries about the Road traffic Statistics?
DfT Ag.
13.a
DfT Cen.
Yes
No
All
Did the statistics provided meet your needs?
3
0
5
Only one of the respondents had not successfully reached the contact point
within the Department citing problems with accessing information concerning
what data was available.
Yes
No
No answer
14.b
Misc. # 2
Misc. # 1
Police
Consult.
LA
Oth Gov.
46
38
1
6
3
0
1
4
0
4
7
0
19
11
1
6
4
0
4
1
0
3
2
0
3
6
0
24
20
40
1
4
3
0
1
4
1
2
8
13
7
11
6
0
4
0
4
1
2
1
2
1
1
7
21
4
60
1
0
8
0
0
5
1
0
10
12
3
16
5
1
4
0
0
5
2
0
3
0
0
9
If, "Yes", have you tried to obtain this note?
Yes
No
No answer
14.c
DfT Ag.
Are you aware that a note describing how the national road traffic
estimates are made is available on the DfT website or on request from the
DfT?
DfT Cen.
14.a
All
OTHER ISSUES
If "Yes", have you been successful in obtaining either a hard copy or a
print-out from the DfT website?
Yes
No
No answer
If you have not been successful, please state the reason why
Conclusions from free text answers
Of the free text answers given only one was relevant. This was simply to say
Misc. # 2
Misc. # 1
Police
Consult.
LA
Oth Gov.
DfT Ag.
To what extent do you agree or disagree with the following statement: "The road traffic
statistics have a level of accuracy appropriate for my purposes?"
All
15.a
DfT Cen.
that it would be easier if there was a complete document available as opposed
to several individual documents.
Strongly Agree
Agree
Neither Agree nor Disagree
Disagree
Strongly Disagree
12
40
18
8
2
2
5
1
1
0
0
3
0
0
2
1
5
4
1
0
3
17
5
3
0
1
3
6
0
0
2
2
1
0
0
0
3
1
1
0
3
2
0
2
0
No Answer
5
0
0
1
3
0
0
0
1
If you 'Disagree' or 'Strongly Disagree' please state why.
Conclusions from free text answers
Police
Misc. #
1
Misc. #
2
LA
Consult.
To what extent do you agree or disagree with the following statement: "The road
DfT Ag.
Oth
Gov.
15b.
All
DfT
Cen.
The most common reason for doubt in our statistics appears to be issues with
the methodology including the sample size and revisions with five responses.
There were four who suggest issues reagrding the minor roads figures, three
who question the accuracy of the statistics at LA level and two each who
question the accuracy of pedal cycles and the lack of quality measures.
length statistics have a high level of accuracy"
Strongly Agree
Agree
Neither Agree nor Disagree
Disagree
Strongly Disagree
6
22
31
7
0
1
2
3
0
0
0
2
2
0
0
0
0
5
3
0
2
12
10
1
0
0
0
6
2
0
0
1
2
0
0
1
2
2
0
0
2
3
1
1
0
No Answer
19
3
1
4
6
2
2
0
1
If you 'Disagree' or 'Strongly Disagree' please state why.
Conclusions from free text answers
The only responses citing lack of confidence in the road length statistics were
due to the methodology including sampling size and revisions. Out of these five
responses the revisions were the most common reason.
16
From time to time, the statistics are subject to revision as new data and/or improved methodology is
introduced. Please let us have your comments on this.
Conclusions from free text answers
A total of forty free text responses were given for this question.
Most of the respondents regard revisions as inevitable but seven explicitly
mention the problems and confusion such revisions induce.
Nine respondents support the revisions to varying degrees.
A further ten respondents suggest an accopnaying explanation and notification
of any revision is important, eight suggest consistency can become an issue
and three suggest users should be consulted about revisions. There were also
suggestions for a measure of quality, status of the figures and quicker
publication of revisions.
17
Please describe any other issues that you wish to raise, e.g. would you be interested in measures of
quality, such as standard errors of the estimates?
Conclusions from free text answers
There was an overwhelming request for quality measures of our statistics.
Twenty-one of a total of twenty-eight responses specifically cite this. A further
two suggested more detailed explanations of the methodology and sources
used, another two the timeliness of the output, one suggesting issues with
minor roads and one requesting forecasted figures
Note from Q12
=
Note: while 14 people answer cite telephone enquiries as a
source from which they obtain the statistics 22 choose to
comment in part 'd'. Therefore there is a discrepancy
between the two questions.
Note from Q14b =
Note: while 24 people answer 'yes' in part 'b', 25 respond in
part 'c'. On closer inspection it became apparent that there
were in fact 27 people who answered one or the other.
Annex 4.1.B
Department for Transport
Road Traffic and Road Length Statistics Quality Review
National Statistics User Consultation: Questionnaire
YOUR DETAILS
Name
Organisation
Job title
Address
Telephone
Fax
E-mail
What type of organisation do you work for (or what type of work are you
involved in)?
Department for Transport (DfT) - central
Department for Transport (DfT) - agencies
Other government regional bodies e.g. Regional Development Agencies
Other Government Departments
Local Authority
Academic
(mark only one option)
Commercial
Consultancy
Other (please specify)
YOUR CURRENT AND FUTURE REQUIREMENTS
1
Which of the following statistics do you use?
(mark as appropriate)
Road traffic statistics
Road length statistics
2
Generally how often do you use the statistics? About:
(mark as appropriate)
Every week
Every month
Every quarter
Every six months
Every year
Less than once a year
3a
If you use the road traffic statistics, please indicate which of the following levels of disaggregation of our road traffic
statistics you currently use and which you may require in the future for Road Traffic Estimates?
(mark one or more boxes in each column)
Curre
nt
Use
Future Use
Individual Road Links
Local Authority Level
English region(s) or Wales or Scotland
Great Britain as a whole
Motorways
Trunk A roads
Principal A roads
B Roads
C Roads
Unclassified Roads
Urban areas
Rural areas
3b
Built up/ non-built-up (based on speed limit)
If you use the road length statistics, please tell us which of the following levels of dissaggregation you currently use and
which you may require in the future?
(mark one or more boxes in each column)
Curre
nt
Use
Future Use
Individual Road Lengths
Local Authority Road Lengths
Road lengths for English region(s) or Wales or Scotland
Road lengths for Great Britain as a whole
Motorways
Trunk A roads
Principal A roads
B Roads
C Roads
Unclassified Roads
Urban areas
Rural areas
Built up/ non-built-up (based on speed limit)
4
Currently, most of the traffic estimates are made for annual (and quarterly)
intervals. Do you require estimates in respect of shorter time periods, such
as hourly, daily, weekly or monthly?
Yes
If "Yes", please specify what time periods you require statistics for and why.
No
Not
applic
able
5
Do you specifically use the seasonally adjusted series which we
currently provide?
Yes
No
If "Yes", please explain why seasonally adjusted rather than unadjusted figures are used.
6
Please indicate which of the following vehicle types (or categories) you
require separate traffic estimates for:
Cars
Light Vans
Heavy Goods Vehicles (HGVs) - total
HGVs - Rigid only
HGVs - Articulated only
Buses
Two wheeled motor vehicles
(mark all that are
appropriate)
All Motor Vehicles
Pedal Cycles
All Traffic
Yes
No
7
Do you require the HGV estimates (as you have indicated above) to be
further split by number of axles?
8
Please describe what you use or will use road traffic and road length statistics for, including the requirements
indicated in response to questions 3 to 7.
9
Are there any other statistics/analyses/data that you would use if we
were to provide them?
If "Yes", please describe briefly what you require and why
Yes
No
Not
applic
able
10
Do you use or are you aware of other sources of statistics that meet your
current needs?
Yes
No
If "Yes", what are the sources?
ACCESSING
THE
STATISTICS
11.a
Do you obtain your information from:
i. Quarterly Transport Statistics Bulletin: Traffic in Great Britain
ii. Annual Transport Statistics Bulletin: Road traffic Statistics
iii. Other Publications (please specify)
iv. Ad hoc queries
v. Other (please specify)
(mark all that are
appropriate)
If you answered
"i" or "ii", please
go to Question
11b.
If you did not
answer "i" or "ii",
please go to
Question 12a.
11.b
If you answered "i" or "ii", does the frequency of publication meet your
current needs and future requirements?
(mark as appropriate)
Need less frequent publication
Frequency is satisfactory
Need more frequent publication
Please explain your answer
12.a
How do you obtain the statistics?
i. Paper Publication
ii. DfT Website
(mark all that are
appropriate)
iii. ONS Website
iv. Telephone enquiry
v. Email enquiry
vi. Personal contact with someone within DfT
vii. Letter
viii. Other (please state)
12.b
If you have used any of the publications (whether paper copy or
Internet version found on the DfT website)
Yes
Is the publication clear and easy to read ?
Is it easily comprehensible?
How could it be improved? (Please state which publication you are referring to)
No
12.c
Were you aware that the statistics published by DfT are available on
the DfT and ONS websites?
12.d
If you obtained statistics from the DfT website were you able to obtain
them easily?
Yes
No
Yes
No
Annual Bulletin
Quarterly Bulletin
Local Authority Traffic Estimates
If "No" for any of the above, please describe any problems you encountered.
12.e
If you obtained statistics by telephone enquiry:
Yes
No
Yes
No
Were you satisfied with how long it took to process your
request?
Did you feel the price you paid was reasonable?
Did the statistics provided meet your needs?
12.f
If you obtained statistics by email or letter enquiry:
Were you satisfied with how long it took to process your
request?
Did you feel the price you paid was reasonable?
Did the statistics provided meet your needs?
13.a
Have you successfully reached the contact point within the Department
for queries about the Road traffic Statistics?
13.b
If "No" please explain briefly what happened.
Yes
No
Yes
No
OTHER ISSUES
14.a
Are you aware that a note describing how the national road traffic
estimates are made is available on the DfT website or on request from
the DfT?
14.b
If, "Yes", have you tried to obtain this note?
14.c
If "Yes", have you been successful in obtaining either a hard copy or a
print-out from the DfT website?
If you have not been successful, please state the reason why
Not
applic
able
15.a
To what extent do you agree or disagree with the following statement: "The road
traffic statistics have a level of accuracy appropriate for my purposes?"
(Mark only one option)
Strongly Agree
Agree
Neither Agree nor Disagree
Disagree
Strongly Disagree
If you 'Disagree' or 'Strongly Disagree' please state why.
15b.
To what extent do you agree or disagree with the following statement: "The
road length statistics have a high level of accuracy"
Strongly Agree
Agree
(Mark only one option)
Neither Agree nor Disagree
Disagree
Strongly Disagree
If you 'Disagree' or 'Strongly Disagree' please state why.
16
From time to time, the statistics are subject to revision as new data and/or improved methodology is introduced. Please let
us have your comments on this.
17
Please describe any other issues that you wish to raise, e.g. would you be interested in measures of quality, such as
standard errors of the estimates?
Yes
No
Yes
No
18
If we were to hold a seminar in you region would you like to attend?
19
We may wish to contact you to clarify your responses. Would you be
happy for us to contact you again?
Please return completed questionnaires as soon as possible, by 30 November 2004 at the latest, via email, with "User
Questionnaire Completed" in the subject heading.
Alternatively, completed paper copies of the questionnaire should be returned to Martin Trevor, Department for Transport, Zone
2/14, Great Minster House, 76 Marsham Street, London, SW1P 4DR.
Contact Louise Buckle for enquiries at:
E-mail:
roadtraff.qualityreview@dft.gsi.gov.uk
Telephone: 020 7944 6389
Annex 4.1.C
REQUIREMENTS IN/OUT OF SCOPE
Organisation
Requirement
Significant from
quesionnaire
Individual road links split
Significant from
quesionnaire
traffic estimates for shorter time periods (hourly &
daily)
Significant from
quesionnaire
separate bus & motorcylce figures
Significant from
quesionnaire
HGV estimates split by number of axles
Significant from
quesionnaire
easier access to publications on the website
Significant from
quesionnaire
higher level of accuracy
Significant from
quesionnaire
explanation and notification of revisions made
Significant from
quesionnaire
measures of quality
Significant from
quesionnaire
forecasting traffic flows
Seminar
give consistent definitions and count practice
Seminar
give guidance on sampling methodology
Seminar
journey times
Seminar
absolute figures and percentage change
Seminar
95% confidence limit or standard errors for
estimates
Seminar
exploit Highways Agency and LA data
Seminar
show how the quarterly and annual data are
linked together.
Seminar
show how the outputs were derived and how the
various inputs fitted in
Seminar
bulletins should show a full list of sources
Seminar
greater consistency in the breakdowns between
the various publications
Seminar
explanation of what the stats are normally used
for and why certain stats are not available
Seminar
Annual Estimates published earlier
Seminar
More details on bus sizes
Seminar
more ATC counts in rural areas
Seminar
Traffic data available by 15-minute periods
In/Out of
Scope
Contraints
Seminar
greater integration and standardisation of traffic
data across the network
Academic
Age distribution of vehicle fleet
Academic
Intermodal Traffic
Academic
International Freight (country of owner)
Academic
Regional flows
Commercial
Pedestrian flow data
Commercial
Pedestrian flow data
Commercial
Public transport data
Commercial
Public transport data
Consultancy
LA disaggregation
O
Subject to ANPR links to
DVLA Database - to be reevaluated if this should
become possible
O
I
O
O
I
Consultancy
Make & Model
Consultancy
Smaller MGV
Consultancy
Speed Limit information
Consultancy
Speed Limit information
Consultancy
Vehicle/Axle Weights
DfT Agencies
Detailed breakdown of Road Traffic Accidents
with regard to nation of drivers licence
DfT Central
Daylight/Darkness
DfT Central
Emissions data
Local Authority
% over 30mph on urban roads
Local Authority
Adjustment factors by month
Local Authority
Company Car statistics
Local Authority
Congestion levels
Local Authority
Data on B, C and U roads
O
Subject to ANPR links to
DVLA Database - to be reevaluated if this should
become possible
I
O
Subject to ANPR links to
DVLA Database - to be reevaluated if this should
become possible
Subject to ANPR links to
DVLA Database - to be reevaluated if this should
become possible
O
Subject to ANPR links to
DVLA Database - to be reevaluated if this should
become possible
O
Subject to ANPR links to
DVLA Database - to be reevaluated if this should
become possible
O
O
I
Local Authority
Goods Vehicles broken into MGV and over 7.5T
Local Authority
Graphical representation of link based network
Local Authority
LA Disaggregation
Local Authority
Peak hour disaggregation
Local Authority
Public transport data
Other
Government
Other
Government
Consistent Road Length statistics
Other
Government
Fuel type
Other
Government
Other
Government
Regional
Government
Regional
Government
TfL
Hard copy of statistics
TfL
Comprehensive data on cycling
TfL
London Borough Disaggregation
TfL
London Borough Disaggregation
TfL
More effort to monitor walking
TfL
Pedal Cycle usage
TfL
Pedestrian Exposure
O
Subject to ANPR links to
DVLA Database - to be reevaluated if this should
become possible
I
I
I
I
Disagg. By Unitary Authority
I
O
Subject to ANPR links to
DVLA Database - to be reevaluated if this should
become possible
I
Stats for non-principle roads at LA level
I
Pedestrian flow data
O
Speed Limit information
Alternative adjsustment factors for cycles
I
I
I
O
I
Organisation
Requirement
TfL
Vehicle Occupancy included in traffic surveys
SUSTRANS
Comprehensive data on cycling and walking (not
currently provided)
SUSTRANS
Cycle/Pedestrian figures on non-traffic routes
SUSTRANS
Development of mechanisms for monitoring the
levels of walking in both road (adjacent) and
traffic free route context.
Greater transparency in the methodological
details of cycling levels
O
In/Out of
Scope
O
Cycling I,
Walking O
O
SUSTRANS
O
I
Contraints
SUSTRANS
SUSTRANS
Improved coverage of cycling monitoring on the
road network
Monitoring of cycling on traffic-free routes
Police
Motorcycle engine capacity (for road accidents)
Police
Seatbelt use
NCSB (National
Cycling Strategy
Board)
NCSB (National
Cycling Strategy
Board)
NCSB (National
Cycling Strategy
Board)
NCSB (National
Cycling Strategy
Board)
ITEA
Accurate national picture of cycle traffic on all
routes
ITEA
Traffic estimate on every link of the roads
network, including minor roads
Road length disagg. Into carriageway standards
I
O
O
O
I
Cycle traffic data broken down by type of route
I
Increased number of sites from which data is
collected
I
More reliable data collected from individual count
sites is not always reliable
Inclusion of Bus Lanes in road length statistics
I?
Government
Office
ATOC
I
I?
Congestion measures
O
Subject to ANPR links to
DVLA Database - to be reevaluated if this should
become possible
SUMMARY OF USER REQUIREMENTS
Annex 4.1.C
Requirement
Quality of Statistics
Data collected from individual count sites is not always reliable
Increase number of sites from which data is collected
Consistent Road Length statistics
Greater transparency in the methodological details of estimating
cycling levels
Development of mechanisms for monitoring the levels of walking in
both road (adjacent) and traffic free route context.
Improved coverage of cycling monitoring on the road network
Access to statistics
A web link, rather than hard copy, of the Annual Traffic Statistics
Bulletin would suffice
Disaggregated statistics
by geographic area
Regional flows
LA Disaggregation
Organisation
Disagg. By Unitary Authority
Separate figures for Scotland
Stats for non-principal roads at LA level
London Borough Disaggregation
by vehicle type
Smaller MGV
Company Car statistics
Pedal Cycle usage
Intermodal Traffic
Accurate national picture of cycle traffic on all routes
Monitoring of cycling on traffic-free routes
by time interval
Seasonal Adjustment factors by month
Peak hour disaggregation
Hours of Daylight/Darkness
by type of route
Cycle traffic data broken down by type of route
Cycle/Pedestrian figures on non-traffic routes
Relating to vehicle itself
Age distribution of vehicle fleet
International Freight (country of owner)
Make & Model
Vehicle/Axle Weights
Fuel type
Emissions data
Motorcycle engine capacity (for road accidents)
Other traffic (and-occupancy) related requirements
Vehicle Occupancy included in traffic surveys
Seatbelt use
Congestion levels
Congestion measures
Detailed breakdown of Road Traffic Accidents with regard to nation of
drivers licence
National Cycling Strategy Board
National Cycling Strategy Board
Other Government Departments
SUSTRANS
IN or OUT
of Scope
IN
IN
IN
IN
SUSTRANS
SUSTRANS
OUT
IN
Other Government Departments
IN
Academic
Consultancy
LA
IN
IN
Other Government Departments
Other Government Departments
Other Government Departments
Transport for London
IN
IN
IN
IN
Consultancy
LA
Transport for London
Academic
National Cycling Strategy Board
SUSTRANS
IN
OUT
IN
OUT
IN
OUT
LA
LA
DfT Central
IN
IN
IN
National Cycling Strategy Board
Other
IN
OUT
Academic
Academic
Consultancy
Consultancy
Other Government Departments
DfT Central
Police
OUT
OUT
OUT
OUT
OUT
OUT
OUT
Transport for London
Police
LA
Other
OUT
OUT
OUT
OUT
DfT Agencies
OUT
Comprehensive data on cycling
Transport but not Traffic
Public transport data
Methodological Requirement
Alternative adjustment factors for cycles
Traffic estimate on every link of the roads network, including minor
roads
Road Network
Inclusion of Bus Lanes in road length statistics
Road length disagg. Into carriageway standards
Speed Limit information
Transport for London
IN
Commercial
LA
OUT
Transport for London
ITEA
IN
IN
ITEA
Gov Office
Consultancy
Regional Government
IN
IN
IN
% over 30mph on urban roads
Data on B, C and U roads
Graphical representation of link based network
Pedestrians
Pedestrian flow data
LA
LA
LA
IN
IN
IN
Regional Government
Commercial
OUT
Pedestrian Exposure
Transport for London
SUSTRANS
OUT
Cycling IN,
Walking OUT
OUT
Comprehensive data on cycling and walking (not currently provided)
More effort to monitor walking
Transport for London
Annex 5.1
ANNUAL TRAFFIC ESTIMATES METHODOLOGY ISSUES
1.
Introduction
1.1
Under this task we consider the way that the annual vehicle-kilometre estimates
are derived. We concentrate upon the estimates for major roads, because minor
road procedures are covered separately.
1.2
Estimates of major-road vehicle-kilometres are based on the product of
estimated Annual Average Daily Flow [AADF] and the road length on individual
links. Therefore the reliability of the vehicle-kilometre estimates depends
directly on the reliability of the AADFs, since the road-lengths are now obtained
accurately from digital maps. Issues associated with road lengths, such as
newly built roads not being included in the more recent versions of the OSCAR
dataset, will be addressed in the near future as SR2 introduce OS's ITN dataset
and derive their road length information from it. ITN is continually updated by
OS with information relating to the road network, including sites where new
roads are opened.
2.
Estimating AADFs for sites counted in the current year
2.1
The 12-hour manual classified counts taken each year are converted to AADFs
by multiplying by expansion factors [EFs] and the EFs are derived from Core
Census ATC data for the appropriate expansion factor categories [EFCATs].
These EFs are derived for each vehicle-type and each EFCAT, although for
certain vehicle-types some of the 22 EFCATs are combined.
2.2
Various methods can be used to calculate EFs, some of which are summarised
below. To derive an EF for [say] Wednesday 13 May:

[Flow-weighting]. We can simply divide the total of all the AADFs for the
ATC sites in the relevant EFCAT by the total of the ATC counts between
0700 and 1900 on 13 May;

[Site-weighting]. We can calculate the EF for 13 May for each ATC site
and then take the mean of these factors or the median. The latter is
currently the procedure used by SR2;

[Regression-based]. We can calculate an EF for each ATC site and then
regress this factor against the 12-hour count.
2.3
Each of these approaches has its own rationale. Flow-weighting relates the
average of all the AADFs to the average of all the 12-hour counts, and as such
allows an unbiased estimate of overall average flow. But it does not provide an
unbiased estimate of the EF at a random site on the network, because it places
most weight upon the high flow sites’ data.
2.4
Site-weighting, on the other hand, does provide an unbiased estimate for the EF
at a random site, if the mean of the site factors is taken. The rationale for taking
the median is that this discounts the influence of ‘unusual’ sites with very high or
low factors, although the median factor may be slightly biased.
2.5
The regression approach is able to cope with the situation where the size of the
EF is found to be related to the size of the flow at the site. If this is the case,
then a significant regression coefficient will be found and the bias in the AADFs
for the various sites will be reduced. Examination of ATC data for some of the
EFCATs indicates that this approach could provide more robust EFs.
2.6
The decision on the appropriate calculation method is not straightforward. In
principle it could vary depending on the date of the manual count, or between
vehicle-types or between EFCATs; however, it is hard to envisage
circumstances where we would allow such inconsistency. Given the need to
provide unbiased AADF estimates for major roads as a key output from the RTS
system, this tends to favour the site-based or regression approaches, but if the
flows at the various sites are of a similar order the three approaches are likely to
produce similar EFs.
2.7
It will be worthwhile calculating the standard errors of the EFs for each vehicletype. This can be achieved directly for most methods of calculation, but if
medians are derived, Jack-knife or Bootstrap procedures may be needed. Any
method tending to produce large standard errors would be avoided.
2.8
In Annex 5.3: Sampling Methodology Issues, we emphasise the importance of
weighting the ATC site data by the inverse of the probability of the site being
selected. This applies whenever the ATC data are used to derive EFs and
whichever method of calculation is used; similarly these weights would apply
when standard errors were obtained.
3
Standard errors of AADFs
3.1
To derive the standard errors of the AADFs, we note that
AADF = EF * 12-hour count. Here both the EF and the 12-hour count are
subject to random error. The standard error of the EF can be derived directly
from the ATC data for the relevant EFCAT sites, as described above.
3.2
The 12-hour counts are also subject to error because of mis-counting and misclassification. To measure these effects it would be necessary to undertake
trials at some typical sites, where a normal counting team’s data were duplicated
by a dedicated team whose error-rate was assumed to be negligible. The
normal team should not be aware of the dedicated team’s presence, so that the
latter would need to locate themselves at a different spot on the link. If an ATC
site were chosen, there would be the added benefit of a further comparison, as
well as a check on the reliability of the ATC itself. To obtain a completely
reliable set of counts, a video-film could be obtained. The classified counts for
each hour or half-hour from the normal team can then be compared with those
of the dedicated team [or the ATC or the video-film if available] and the mean of
the sum of the squares of the differences computed. This gives us an estimate
of the variance of the counting error for each hour or half-hour, from which can
be derived the variance and standard error of the 12-hour count.
3.3
To derive the variance of the AADF we use the formula for the variance of the
product of two random variables, which is based upon the means and variances
of the individual variables, X and Y:
Var [ X * Y ] = YBAR * Var[X] + XBAR * Var[Y] + Var[X] * Var[Y]. This holds
provided that X and Y are independent, which applies in this case. Hence we
can obtain the standard error of the AADF on a link, from which we derive the
standard error of the overall mean flow and of the overall vehicle-kilometres.
Confidence limits for the AADFs and vehicle-kilometres can then be derived.
4.
Estimating AADFs for sites counted in earlier years
4.1
Where a site is not counted in the latest year, it is possible to estimate the AADF
by updating the AADF calculated in an earlier year, using the growth-factor
derived from the Core Census ATC sites.
4.2
These year-on-year growth factors [gfacs] are obtained from the AADFs for the
ATC sites in each EFCAT. As with EFs there are alternative methods for
deriving gfacs. In each method we should take account of the weights [the
inverse of the selection probabilities] attached to the different sites. The
decision on the appropriate method of deriving gfacs would again take account
of the standard errors.
4.3
When the AADF for the latest year is estimated from the AADF for an earlier
year, there is clearly an additional source of random error as a result of using
the gfacs. We have
[AAFDF latest year] = [AADF counting year] * gfac.
We have the means and variances of the AADFs in the counting year, and of
gfac, so again we can apply the formula for the product of two random variables,
because the two quantities are independent.
5.
Assessing the performance of the AADF estimates
5.1
For a site that is counted in the latest year we can derive its AADF directly using
EFs. But we can also look at the AADF that would have been ‘predicted’ from
an AADF obtained for an earlier year. If we take a set of these sites and plot the
latest AADFs against the predictions in a scatter diagram, we will be able to
assess whether the predictions are systematically high or low, and obtain an
estimate of the variance of the ‘prediction errors’. These comparisons will tell us
how well our growth-factors are performing.
5.2
We could also investigate whether alternative methods for calculating these
factors would have proved more robust, which might lead in turn to a revision in
the methodology for future years.
6.
Need to revise AADFs and vehicle-kilometres for earlier years
6.1
We have stressed the need to compare the performance of the growth factors in
predicting AADFs from those in earlier years. When bias is found to be
statistically significant, then some revisions will need to be made to the traffic
estimates for the years covered by these growth factors. The revised vehiclekilometre figures for the intervening years can be obtained by scaling the growth
factors for individual years up or down so that the overall growth is correct.
6.2
It is not always necessary to issue revised AADFs for individual links when the
overall vehicle-kilometres are revised. If the 95% confidence ranges for the
AADFs are large [say of the order +/- 40% of the mean] and the implied change
in traffic levels relatively small [say, 3%], the change in the overall limits
becomes trivial.
7.
Bias-correction factors
7.1
Bias-correction factors are used currently:

Because the Core ATC sites do not cover the full range of high-flow/lowspeed sites on motorways and trunk roads and therefore have biased
expansion factors. The factors are in fact too low, because the ‘missing’
sites tend to have higher than average proportions of traffic in the periods
before 0700 and after 1900 and at weekends;

Because comparisons with the CSRGT suggest that the RTS counts of
goods vehicles may be biased and involve some misclassification.
7.2
In Annex 5.3: Sampling Methodology Issues, we have recommended extending
the ‘reach’ of ATC sites to all parts of the network. If satisfactory arrangements
can be made for the difficult sites, then the need for correcting biased expansion
factors should diminish. Also if we follow the recommendation to weight the data
by the inverse of site-selection probability, this should also tend to reduce bias in
the expansion and growth factors.
7.3
Under discussion of the findings of the CSRGT review (Annex 8.4), we
recommend some trial counts of goods vehicles, in order to assess the extent to
which the manual counts are biased and hence provide improved guidance and
training for the enumerators. If successful, this would allow us to dispense with
the corresponding correction factors.
7.4
It may be optimistic to anticipate eliminating the need for bias correction-factors,
and we should continue to draw comparisons between AADFs and other
sources, such as the CSRGT and AADFs produced by LAs and the Highways
Agency.
8.
Summary of Recommendations
8.1
For the annual estimates our recommendations are as follows:

The different methods of calculating expansion factors should be
reviewed, and their standard errors calculated;

In calculating expansion factors, growth factors and their standard errors,
the ATC site data should be weighted by the inverses of the site-selection
probabilities;

Standard errors of 12-hour manual counts should be estimated using
counting trials at a few sites;

Standard errors of AADFs should be derived from the standard errors of
the 12-hour counts and the expansion factors;

Standard errors should be obtained for AADFs derived from growth
factors applied to AADFs for earlier years;

The standard errors of AADFs should be used to derive standard errors
for vehicle-kilometres. These standard errors should be shown alongside
published statistics;

Comparisons should be made between AADFs based on current year
counts and AADFs ‘predicted’ by growth factors from earlier years.
Significant differences could lead to revisions in methodology and to
revised vehicle-kilometre estimates for earlier years;

When revisions to overall traffic levels are required, this does not
necessarily require revisions to individual AADFs;

Whilst we would hope to eliminate the need for bias-correction factors, by
various improvements in sampling and other measures, there is a
continued need for cross-comparisons with other sources, such as
CSRGT, LA data, and Highways Agency data.
Annex 5.2
QUARTERLY TRAFFIC ESTIMATE METHODOLOGY ISSUES
1.
Introduction
2.1
Quarterly traffic statistics are published on a provisional basis six weeks after the
end of the quarter in question. The figures comprise

A breakdown of vehicle-kilometres by the main vehicle-types: cars, vans,
lorries [which are somewhat misleadingly described as ‘goods vehicles’],
pedal cycles and other vehicles [these being the total of buses and
motorcycles]. These are also shown in seasonally adjusted form as
indices based on 1993=100;

A breakdown of motor vehicle traffic by the major road-classes and a total
for minor roads, both split between urban and rural areas;

For cars and lorries only, a similar breakdown by road-class.
2.2
These figures are necessarily provisional, being based solely on trends derived
from the Core ATC sites, applied to an earlier set of finalised annual figures.
Figures for the full calendar year depend also upon the estimated AADFs, and
this further information can cause revisions to the quarterly figures some 5-6
months after the end of the year.
2.3
Provisional quarterly figures are needed as an adjunct to the quarterly road
casualty and congestion figures: casualty-rates are expressed in terms of
casualties per 100 million vehicle-kilometres and changes in speed and travelrate are negatively correlated with changes in traffic level. There is also interest
in the evolving trend of traffic growth or decline, and the latest trends can be
deduced from the rolling 12-month total or from the seasonally adjusted
quarterly data.
2.4
In reviewing the provisional quarterly statistics, we consider the following
questions:

Are the figures sufficiently reliable? Are the potential revisions so large
that we risk providing misleading pointers to traffic trends in the
intervening period before the figures are revised?

Do we show too much detail, given the likelihood of revision? [An
example might be the quarterly pedal cycle figures]. Or do we obscure
some of the key messages by excessive aggregation? [An example is
the aggregation of bus and motorcycle figures];

Can the provisional figures be improved as data from the 12-hour manual
counts come available during the year?

Is the seasonal adjustment process sound?
3.
Reliability of the provisional figures
3.1
Provisional estimates for Q1 of 2005 will depend upon growth factors between
Q1 2003 [this being the latest years with final figures] and Q1 2005. Using
procedures similar to those recommended for the annual growth-rates, we could
in principle estimate the standard errors of these growth-rates and therefore be
able to assign 95% confidence limits to the year-on-year changes in the
quarterly figures. If there were sufficient demand from users, it would be helpful
if Table 1 of the Bulletin showed these confidence limits, alongside the
percentage changes.
3.2
We cannot expect the quarterly changes necessarily to be significant at the 95%
confidence level, especially when these are based purely on the ATC data. By
implication, if we present provisional changes as being significant, we would
expect the final figures to show changes in the same direction. So for example,
if we say that provisional car traffic has grown by 2.9% and the 95% confidence
limits for the change are +/- 2 percentage points, then we would expect the final
figures also to show an increase in car traffic [though the final increase might not
be significant]. It would be instructive to see a table comparing provisional
figures with final figures for the years 2000 onwards.
3.3
Provided that the confidence limits were published, we need not be too
concerned about showing summary figures for the main vehicle-classes,
including buses and motorcycles. Combining these two vehicle-types together is
conceptually unhelpful, even though the combined total may be more robust
than the component parts.
3.4
On the other hand we have some concerns about publishing provisional pedal
cycle figures. Even the annual pedal cycle figures have proved volatile and we
cannot rule out significant revisions to the quarterly results. Until a reliable panel
has been established [as we have recommended elsewhere], it seems likely that
the standard errors of the quarterly changes will be large. It may be best to
publish quarterly cycle traffic figures only when they have been finalised, ie for
past years.
3.5
There must be some doubts about the figures for lorries on minor roads in Table
4 of the Bulletin, given the low flows on minor roads. It would not be surprising
to find that the standard errors of the percentage changes were very large, or
that quite large revisions might occur when the figures were finalised.
3.6
It seems unlikely to be worthwhile attempting to use the manual count data to
refine the provisional quarterly estimates, in advance of the AADFs being
calculated. This is partly because timing considerations: the counts require
validation, so that reliable data from the spring surveys could not be used until
the autumn, by which time the provisional Q2 estimates would have been
prepared. But more fundamentally, the full implications of the manual data
cannot be absorbed until AADFs have been estimated, and final annual
estimates of vehicle-kilometres prepared.
4.
Seasonal Adjustment
4.1
The quarterly figures are input to the X11 seasonal adjustment program, which
examines the series for potential turning points and irregularities, and then
provides seasonal adjustment factors that take account of normal seasonal
movements. Table 1 of the Bulletin shows the seasonally adjusted series.
4.2
X11 allows for the effects of Easter to be modelled. Given that Easter can occur
in Q1 or Q2, depending on the year, it will be important to include these effects
in the X11 model. One way of doing this would be to have a dummy variable
which takes the value 0 if the quarter does not cover Easter, 1 if it covers only
Good Friday or Easter Monday, and 2 if it includes both days. However, this will
have a comparatively small effect on the seasonal adjustment.
4.3
One of the diagnostics from X11 is the QCD [Quarters of Cyclical Dominance].
This expresses the intrinsic reliability of the quarterly figures, ie whether real
trends from one quarter to another dominate irregular movements. A QCD=1
implies that a single quarter’s seasonally adjusted figure is robust, but if the
QCD=2, then it would be safer to average together two successive quarterly
figures.
4.4
The final quarterly figures are derived by scaling the seasonally adjusted
provisional figures, so that they agree with the final total.
5.
Summary of Recommendations
5.1
Our recommendations can be summarised as follows:

Table 1 in the Quarterly Bulletin could show standard errors or confidence
limits for the percentage changes if there were sufficient User demand;

The Bulletin should show bus and motorcycle figures separately, unless it
was shown that the figures were likely to undergo major revisions;

On the other hand, it seems hard to justify publishing provisional quarterly
figures for pedal cycles, and for lorries on minor roads, unless the
confidence limits are shown to be tight;

It is not worthwhile attempting to use the available manual count data to
refine the quarterly vehicle-kilometre estimates;

The seasonal adjustment process should take account of the possibility of
Easter holidays occurring in Q1 or Q2;

The Quarters of Cyclical Dominance diagnostic from X11 should be
reported, along with advice on the use of seasonally adjusted figures.
Annex 5.3
SAMPLING METHODOLOGY ISSUES
1
Introduction
1.1
In this section we cover a number of issues relating to sampling within the Core
Census and within the one-day manual count programme. These relate to the
location of sites within and between the various road-class categories [usually
referred to as EFCATs], and the times of year when manual counts occur.
1.2
Inevitably there is overlap with other issues. For example
1.3

Site location in the Core Census is related to suitability of sites for ATC
equipment and with the need for bias correction in the quarterly and
annual estimates;

The overall size of the Core Census affects the precision [ie the size of
the standard errors] of the quarterly and annual estimates;

The frequency of counting of major road counts in the one-day manual
count programme affects the precision of the quarterly and annual
estimates;

How we choose to sample minor roads affects the annual estimates for
all traffic and for pedal cycle traffic on these roads. Minor road and cycle
traffic estimation deserve special attention and are covered elsewhere.
We will tackle the following issues under the Sampling heading:

Core Census: historical basis for sampling; the need for sampling
probabilities to be established and used in calculations; site location bias
and need to extend the ‘reach’ to low-speed sites; sampling within
London; size of the Core census and allocation between road-class
categories;

One-day Manual Count Programme: frequency of counts on major road
sites; allocation of sites between months; possibility of sub-sampling
within 15 minute periods; alternatives to link-based counts.
2
Core Census sampling issues
2.1
The Core Census is based mainly upon ATCs which classify traffic at 190 sites
in GB outside London, plus a further set of 56 non-classifying ATCs within
London. The main outputs from the Census are

Factors that relate a 12-hour flows [usually on a weekday in a neutral
month] to the annual average weekday flow. These are known as
expansion factors [EF] and are used to convert 12-hour manual counts
taken at various sites to AADFs. Similar factors could also be derived
relating other counting periods to other annual averages;

Factors that relate the average daily flow in one period [for example a
calendar year or a quarter] to the average flow in some later period.
These are known as growth factors [gfac].
The importance of sampling probabilities
2.2
These factors are calculated for ‘expansion factor categories’ [EFCATS] which
are defined in terms of road-classes, urban and rural areas as defined for the
2001 Population Census, and type of region [in particular in relation to London
and holiday-areas]. These categories have been drawn up within SR2 in order
to achieve reasonably homogeneous factors within EFCATs. They have been
developed with experience over the years and, not surprisingly, are different
from those used in drawing up the original sampling-frame.
2.3
Hence the ATC sites found within a particular EFCAT will originally have
belonged to a number of different categories as originally defined; they will also
be associated with different likelihoods of selection, because the sampling rates
varied between the original categories. All this suggests that we should take
account of the site selection probabilities in calculating our factors; if we ignore
these probabilities [as is currently the practice], then we risk giving too much
weight to data from sites with a high probability of selection and insufficient
weight to sites with a low probability. It is therefore important to examine the
ways in which sites were selected and added to the Core Census and to assess
the corresponding probabilities.
2.4
The ATC sites were selected at different times by different methods:

An original batch of about 120 were selected in 1980 randomly
throughout GB within 11 road-classes, where the 5 non-motorway roadclasses were split between built-up and non built-up sections [built-up
roads being those with 40 mph or lower speed limits]. The chance of a
link being selected was proportional to its length and generally the
sampling rate was higher for built-up roads than for non built-up roads.
However, a handful of low-speed sites had to be abandoned and
replaced by sites more suitable to ATCs and therefore a slight element of
non-randomness was introduced;

In 1986 a set of about 50 sites were randomly selected within inner and
outer London and by road-class.
These were volumetric ATCs
supplemented by manual counts to provide a split by vehicle-type;

In the early 1990s a second batch of about 13 sites was added [outside
London], apparently at random;

In the mid-90s a third batch of 70 sites was introduced, about half of
which were in Scotland. The basis of selection was a perceived need for
more sites on certain road classes on urban or rural roads [urban/rural
being defined in terms of agglomerations of Census output areas] in
certain regions. The sites were not selected completely at random; the
chosen links were found by the engineers looking for suitable ATC sites,
and typically were in areas with a power supply nearby and uncongested
traffic.
2.5
It is clear that the chances of selecting particular ATC sites vary considerably.
Those in the first batch, and the London sites, can be assigned probabilities
fairly accurately, these being related to the number of sites on the relevant roadclass, the length of the link on which they are situated, and the total length of
road of the given class. [The probability is approximately equal to the number of
sites selected, times the length of the link, divided by the total road-class length.
The lengths of the 11 road-classes can be found by examining past road-length
figures].
2.6
Similarly it should be possible to assign probabilities to the second batch of nonLondon sites, provided that the basis for their selection can be determined; a
slight complication is that, strictly speaking, these probabilities should be
multiplied by the probability that they were not selected in the first batch.
2.7
With the third batch of non-London sites, assigning probabilities is much more
difficult, because the method of selection is perhaps less apparent. One could
take the view that these sites were picked out deterministically, so that each site
represents only itself and that the probability of selection was therefore 1; but
this is an excessively restrictive approach, given that the engineers could have
chosen different sites. An alternative approach would be to examine the context
in which each set of sites was chosen. Suppose, for example, that there was a
need for 3 more urban motorway sites in the West Midlands. Then the
probabilities for these sites will be approximately 3 times the link length divided
by the length of urban motorway in the West Midlands not already covered by
existing sites, provided that this total length is suitable for ATCs; if parts of the
network are unsuitable for ATCs, then the total length has to be reduced
accordingly. In general the probabilities for the different sites in the third batch
will be larger than those in the first two batches, because they have already
been passed over in the first 2 batches [which has to be reflected in the
probabilities] and because the new sites were normally directed towards specific
localities.
2.8
Suppose we now want to calculate an expansion factor based for a particular
EFCAT. If we do this with a flow-weighted calculation, the formula is
EF =

AADFi /

Ci
where AADFi is the AADF for site i and Ci is the count on the day for which the
factor is required; the summation is over all sites i within the EFCAT. But taking
account of the site selection probabilities, the formula becomes
EF =

[AADFi / pi ] /

[ Ci / pi ]
where pi is the probability that site i was selected originally. Similar
considerations would apply if the expansion factor was first calculated for each
site, and then the mean or median taken; the calculation of the mean or median
would need to weighted by [1/ pi].
2.9
The need for weighting by the inverse of site-selection probabilities becomes
apparent when we define Expansion Factor Categories [EFCATs] in ways that
cut across the categories used as a basis for selecting the ATC sites. This is
because, for example, a category defined in terms of urban roads will contain
sites from the original 120 that were selected by built-up and non built-up areas,
using different sampling rates for these two types of road. If the original
sampling-rates are not taken into account, then we are likely to have ‘too many’
built-up sites in our expansion factor categories and our EFs will be excessively
influenced by these sorts of site.
2.10
This is also relevant when we have a site on a link which changes its category,
for example from class A to class B. If we know its selection probability, then
the site data can now be used alongside data from other class B sites, provided
that they are weighted by the inverse of this probability.
Site location bias: need to extend the ‘reach’ of Core Census sites
2.11
The Core Census was established to provide continuous traffic monitoring,
classified by vehicle-type, at a set of representative sites on all types of road.
The need for continuity arose because EFs need to relate counts taken on any
day of the year to an annual average [AADF] for the whole year and gfacs need
to relate AADFs for different years. Therefore when a Core census site is
elected, the presumption is that it will be covered by a classifying ATC.
2.12
The latest generation of ATCs have been found to function well at the majority of
sites, but there were two factors that gave rise to problems:
2.13

If there was no mains power supply, then the ATCs would have to rely on
batteries, which required frequent site visits;

If the traffic was very slow-moving [eg below 5 km/hour] or in stationary
queues, then the classification algorithms were unable to distinguish
individual vehicles properly.
Given these difficulties, there has been a tendency to avoid sites with these
characteristics when new sites are selected. Whereas one would want to select
sites on links that have been selected randomly in order to obtain a ‘probability
sample’ [ie one which gives an equal probability to all potential sites within a
stratum], in practice compromises have been made. Three examples are

On M1 southern sections and on M25 the problems of getting ATC sites
installed [both because of electrical supply problems and slow-moving traffic]
have prevented any Core Census sites covering these sections of the
motorway network. As a consequence the EFs and gfacs for motorways
have become biased, given that the data do not take account of these busy
sections. Bias correction factors have had to be introduced to allow for this.
[This has been feasible because non-classifying volumetric counts on these
sections are available from the Highways Agency]. However, the use of
correction factors is a comparatively crude adjustment process;

When further Core Census sites were being selected for principal roads
during the mid-1990s, the site selection was in part governed by the need to
avoid sites with slow-moving traffic. The possible bias induced by these
considerations is not known; it could only be quantified by conducting some
other type of counts at the ‘unsuitable’ sites and comparing the factors for
these sites with those based upon ‘suitable’ sites;

Within London it was considered probable that a large proportion of the Core
Census sites would prove difficult, because slow-moving conditions
frequently arise. Therefore in 1986 it was decided to use simple volumetric
ATC equipment, which is less susceptible to the slow traffic problem
[although under extreme conditions this also can fail to distinguish all the
vehicles passing the site]. To obtain the vehicle-classification, the ATCs
divide traffic into short and long vehicles, and manual classified counts are
mounted 4 times a year, to provide the further breakdown by the standard
vehicle-classes. The London sampling issues are considered separately
below.
2.14
This is an unsatisfactory situation, because it effectively ‘writes off’ important
parts of the network, notably those parts most subject to congestion and those
parts remote from electricity supply. It also makes the assignment of siteselection probabilities difficult because some sections of the network are not
available for sampling and the precise length of such sections is not known.
And, as we have seen above, it gives rise to bias, some of which cannot be
quantified or identified.
2.15
How can we address these problems? The electricity supply problem can now
be overcome by the use of independent power sources, ie solar panels and
windmills which charge batteries at the sites. It appears that virtually any
location in GB would be suitable, so that we can now consider re-selecting sites
that were chosen randomly but then rejected because of the power-supply
problem. This should allow a number of potential motorway sites to be
introduced.
2.16
The low-speed sites are more difficult to deal with, because the technical
difficulties remain. In the meantime three alternative approaches should be
considered:

1. Volumetric [non-classifying] ATCs could be installed and the breakdown
by vehicle-class handled in the same way as in London. This is a
comparatively expensive option because of the use of manual classified
counts;

2. Volumetric [non-classifying] ATCs could be installed and the breakdown
by vehicle-class handled by data from ANPR cameras. Whilst ANPR
systems are currently expensive and do not identify every vehicle that
passes, they could provide sufficient data to provide the split by vehicle-type.
Furthermore ANPR cameras have already been installed at a large number
of motorway and trunk road sites, and within central London for congestion
charging. It is worth investigating the possibilities of using these data,
provided that the appropriate assurances can be given on commercial
confidentiality and privacy;

3. On motorways and trunk roads the Highways Agency ATCs could be
used. This would be a similar option to 1. above but the split by vehicle-type
might be assisted by the classification by vehicle-length at some HA sites.
[This is discussed further in Annex 6.4].
2.17
The respective merits of these approaches [and any other technologies that look
promising] could be investigated by undertaking trials at known low-speed sites
on motorways and in urban areas. The economic factors, such as the cost of
ANPR systems, are likely to change from year to year, so that, for example,
manual counting might be the most cost-effective approach in the next 5 years
but ANPR might be preferred from 2010 onwards. As regards data quality,
ANPR is likely to provide the best vehicle-type split, because it is based on
DVLA records and is available for any 24-hour period; manual counts are
limited to just a few days.
Sampling within London
2.18
As noted above, ATC sites within London are based upon random sampling.
The sites were selected within road-classes in Inner and Outer London, and
hence can be assigned probabilities. Therefore from a sample selection
viewpoint, the existing sites are probably representative.
2.19
The ATC equipment at the London sites is known to be ‘falling apart’ and the
central processing system for handling the London data is also in need of
replacement. So the options for the London Core Census will have to be
considered. The three main options would be

To use the standard Core Census classifying ATCs. This may be the right
approach for the sites that do not suffer from low-speed or queuing
problems, perhaps up to a third of the sites;

To use volumetric ATCs supplemented by MCCs, as at present;

To use volumetric ATCs supplemented by ANPR camera systems.
2.20
Whichever approach is adopted, it should be possible to integrate the London
Core Census more closely with the non-London Core Census. The London
sites would remain as separate strata within the GB system. There is no
obvious need to revise the current sample of sites, since these form a
‘probability sample’. However, the current number and distribution of sites
between strata could be reviewed as part of a more general review of the GB
sampling and stratification.
2.21
The important consideration within London is to take into account the
information requirements of Transport for London. TfL organise three cordon
surveys, covering traffic which crosses the external boundaries to central, inner
and outer London; these long-standing surveys are normally conducted at 2/3
year intervals. These surveys are complemented by counts across strategic
screenlines at the River Thames and at other lines designed to detect orbital
traffic. Regular counting with ATCs takes place on the strategic network for
which TfL is responsible.
2.22
Currently TfL are carrying out extensive monitoring related to congestion
charging. This involves both ATCs and manual counts at sets of random points
within the charging area and on its boundary road. The congestion charge
monitoring is expected to cover the proposed Western Extension to the charging
area. However, it is unlikely that these counts will continue indefinitely; the
primary aim has been to monitor the effects of the scheme during its first three
years of operation.
2.23
2.24
In 2001 TfL commissioned a review of its transport flow monitoring activities.
This review recommended that TfL discuss with DfT the best means of
monitoring road traffic volumes within London as a whole. Pending the work on
congestion charging, these discussions have been put on one side, but the
current RTS review provides a good opportunity for them to get underway. The
aim of any joint monitoring programme would be

To continue to monitor trends in strategic movements in different parts of
London [in effect maintaining at least some of the long-term time-series of
cordon and screenline surveys];

To develop further the system of random ATC counts [currently
undertaken by the London Core Census and by one-day counts at other
sites] so that it can produce reliable trends on overall travel by major
roads [vehicle-kilometres]. The TfL strategic network counts would also
be part of this process.
The second aim specifically mentions major roads, because it is unlikely that
sufficient minor road sites could be covered in the London Core Census to
produce reliable trends for minor roads. However, this might be achievable
using some other survey programme and this is discussed separately under the
topic of Minor Roads [see Annex 5.4]. For major roads the initial task is to
assess the standard errors of the AADFs and hence the overall vehiclekilometres, based upon the current DfT programme and the TfL strategic
network counts. It is then for TfL and DfT to consider whether the resulting
confidence limits are adequate, bearing in mind that traffic estimates for London
will be used – in conjunction with speed surveys - to measure changes in
congestion.
Overall size of the Core Census and its distribution between strata
2.25
The overall ‘size’ of the Core Census is governed by the number of sites. Table
X shows the current number of active sites.
Table X: Number of Automatic Count Points in each Expansion Factor Category for CARS (as
at November 2004)
EFCat
Total
2-way
1-way
01
2
2
0 Motorways in holiday areas
02
3
3
0 Motorways in other rural areas with an estimated AADF of of up to 59,999
03
3
3
0 Motorways in other rural areas with an estimated AADF of 60,000 or more
04
10
9
1 Motorways in part rural and part urban areas and conurbations
05
4
4
06
10
10
0 Motorways in mostly urban areas and Greater London
Rural A roads in holiday and very rural areas with an estimated AADF of up
0 to 4,999
07
5
5
08
7
7
Rural A roads in holiday and very rural areas with an estimated AADF of
0 between 5,000 and 7,999
Rural A roads in holiday and very rural areas with an estimated AADF of
0 8,000 or more
09
9
8
1 Rural A roads in all other areas with an estimated AADF of up to 13,999
10
10
10
0 Rural A roads in all other areas with an estimated AADF of 14,000 or more
11
1
1
12
4
4
13
5
5
0 Urban A roads in holiday areas
Urban A roads in all other areas except Greater London with an estimated
0 AADF of up to 19,999
Urban A roads in all other areas except Greater London with an estimated
0 AADF of 20,000 or more
14
15
15
0 Urban A roads in Outer London
15
14
13
1 Urban A roads in Inner London
16
2
2
0 Urban A roads in Central London
50
5
5
0 Minor rural roads in holiday areas with an estimated AADF of up to 399
51
5
5
0 Minor rural roads in holiday areas with an estimated AADF of 400 or more
52
7
7
53
8
8
0 Minor rural roads in all other areas with an estimated AADF of up to 2,499
Minor rural roads in all other areas with an estimated AADF of 2,500 or
0 more
54
9
9
0 Minor urban roads in all areas except Greater London
55
19
157
18
153
1 Minor urban roads in Greater London
4
2.26
The total number of sites and their distribution will affect the accuracy and
precision of the two sets of factors derived from the Core Census; it will also
affect the capital costs and the annual running costs.
2.27
To assess the precision of the EFs we can look at the variation between the
EFs for different sites within each EFCAT, where these factors are averaged
over the ‘neutral’ months, or alternatively we can simply use factors for a typical
neutral month such as May. Let the standard deviation of these factors within
EFCAT s be denoted by SDs and let VKs be the total vehicle-kilometres within
EFCAT s. Then the optimum distribution of sites between the EFCATs [ie the
distribution which minimises the overall confidence limits for the expansion
factors] will be proportional to [SDs * VKs]. [It may not be possible to include all
22 EFCATs in this review of the expansion factors, because several have only 3
or fewer Core sites with which to estimate variances].
2.28
But even using this ‘mechanistic’ approach there may be conflicting indications;
for example, it may indicate different balances of sites for cars and lorries, or
for different neutral months. Similar calculations can be made to obtain the
optimum distribution of sites for estimating growth factors. Whilst there may be
a number of conflicting criteria, and other considerations such as site-costs
come into play, nevertheless it will be worthwhile reviewing the distribution of
sites between the different road strata. Indeed the review might lead to the
strata themselves being revised, with certain strata which contributed
excessively to the overall variance being split, and with other strata that had
similar factors being combined. Analysis packages such as CHAID would
probably be helpful.
3
Sampling for the One-day Manual Count Programme
3.1
This programme used to be called the Rotating Census, because the objective
was to cover around 3000 sites in ‘neutral’ months each year, and after a fixed
cycle of five or six years, to ‘rotate’ through all the potential sites in GB. In this
way the Rotating Census would provide a basis for estimating the average
levels of flow in different regions and on different classes of road. This the Core
Census, with its limited number of sites, could never do; on the other hand the
Rotating Census could not estimate year-on-year trends reliably, because by
definition the set of sites surveyed was always changing.
3.2
In the light of experience a number of modifications were made to the Rotating
Census during the late 1980s and subsequently. First it was always recognised
that it would be impossible to cover the complete set of possible minor road
sites, so a finite set of just over 4000 was used to represent minor roads,
equivalent to 800 per year over a five year cycle. It became increasingly
obvious that minor road flows varied enormously even within a single road-class,
and that some of the very low-flow sites were not worth covering, because they
wasted resources [eg sending a surveyor to count fewer than 100 vehicles over
a 12-hour day] and because these sites actually contributed very little either to
the overall average flow or to changes in average flow from year to year.
Consequently a number of changes were made to the make-up of potential
counting sites on minor roads, the most recent change being at the 1999
benchmark review. Following the review by Vic Barnett in the early 1990s, it
was decided that cycle and motorcycle counts on minor roads needed to be
undertaken in winter and summer months and at weekends, because these
vehicles had profiles distinctly different from other vehicles. This review and the
Scottish review in 2001 [??] also recommended an increase in the number of
minor road sites.
3.3
The other important change to the Rotating Census was for major roads.
Instead of a fixed rotation cycle, this was made to vary depending upon the level
of flow at each site. Thus the high-flow sites are now covered annually, and
others at intervals of two, three or four years. The lowest flow sites are covered
only once in eight years.
Frequency of counts on major roads
3.4
If we wanted to review the counting frequency for road links belonging to the
different strata, this could be addressed in a similar way to the method we have
suggested for the Core Census sites. The ‘optimum’ strategy would minimise
the size of the confidence limits for estimates of overall vehicle-kilometres. This
would be achieved if the number of sites counted each year in an EFCAT s was
proportional to the product of the vehicle-kilometres [VKs] and the standard
deviation of average flow [SDs]. Both of these quantities can be calculated using
the data from the latest year.
3.5
Such a review would also allow us to consider the balance between major and
minor roads sites. The procedures for minor roads are considered separately,
because a number of alternatives need to be considered. [See Annex 5.4].
Allocation of counts between months
3.6
The one-day counts on major roads are spread between the ‘neutral’ weeks and
months in such a way that nearby sites tend not to be counted in the same
month. There is no attempt to count a site on the same day of the week or the
same month on each year it is included. This is because the primary objective
of the one-day counts is to cover a wide range of locations and dates. This
tends to minimise bias in the estimates of average flow over the main cycle of
surveys.
3.7
From time to time it would be worthwhile reviewing the range of weeks on which
the one-day counts are taken. This would be undertaken by examining the
corresponding EFs derived from the Core Census.
Possibility of sampling within 15-minute count periods
3.8
At present the one-day counts are taken for 12 hours continuously. An
alternative procedure, which has been used in London for over 30 years, would
be to count for 6 minutes in every 15 minutes and multiply this count by 2.5. It is
important that in taking these 6 minute counts, the counting starts at a random
time, rather than at an arbitrary point such as the beginning of a green trafficsignal cycle.
3.9
This of course makes the counts less precise: the confidence limits for the
hourly or daily totals would be multiplied by the square-root of 2.5, ie by 1.58.
The increase in the standard error of the AADF would be less than this, because
the random errors from the counting process would be compounded with those
from the expansion factors. The benefit from 6-minute counting is that the
surveyors can carry out another 6-minute count during each 15 minute period.
Typically, for example, it is possible to cover two directions at busy sites with
fewer surveyors than would be needed for continuous counts. Another
possibility is to use the same surveyors to conduct counts at two arms of a
junction in a 15 minute period.
3.10
The implications of this type of sampling should be considered by DfT, with a
view to increasing the overall coverage of the one-day count programme.
4.
Summary of Recommendations
4.1
The following recommendations are made under the heading of sampling:

The sampling probabilities of all the Core Census ATC sites need to be
determined. These can be derived from considering the lengths of the links
on which sites were chosen, and the total length of road that was available
for sampling [allowing for the fact that some links were ruled out as
unsuitable] , at the time the samples were chosen;

Calculations of expansion factors and growth factors using Core Census
data should weight the data from each site by the inverse of its selection
probability. Failure to carry out this weighting may cause bias in our
estimates;

To overcome locational bias in the Core Census, we need to extend its
‘reach’ so that it can cover sites where ATC operation has proved difficult.
These ‘difficult’ sites include M25, southern sections of M1, and congested
links in conurbations.

Electricity supply problems can be met by using solar power or windmills,
which would allow a number of motorway sites to be covered;

Low-speed site problems will have to be addressed. The most viable
approaches currently are: [a] a combination of volumetric ATC counts,
supplemented by classifying manual counts, [b] a combination of volumetric
counts and existing ANPR data, used to disaggregate by vehicle-type, or [c]
on motorways and trunk roads, by similar methods using Highways Agency
data. None of these options is straightforward, and site trials will be needed;

London Core ATC monitoring should be reviewed with Transport for London
statisticians, to devise a programme that meets both national and London
monitoring needs. There is no apparent need to revise the ATC monitoring
sites, but given that many of the sites are low-speed, the counting options at
these locations will be as above;

The overall size of the Core census, and the balance of sites between
expansion-factor categories [EFCATs], should be reviewed, based upon
achieving satisfactory standard errors of the expansion factors;

Within the one-day manual count programme, the counting frequency [ie the
number of counts per year] within each EFCAT should be reviewed, based
upon minimising the standard errors of the AADFs;

The allocation of counts between weeks and months is satisfactory, but we
could review the range of weeks and months over which counts might be
taken, based upon the requirement for satisfactory expansion factors;

The arguments for and against counting for only 6 minutes in each 15 minute
period should be quantified in terms of their effects on standard errors and
costs.
Annex 5.4
MINOR ROAD TRAFFIC ESTIMATES METHODOLOGY
4
Introduction
4.1
Minor roads cover all public roads other than motorways and class A roads.
They account for 87% of total road length but only about 36% of total vehiclekilometres. In terms of traffic counting they account for 29% of the Core Census
ATC sites and 20% of manual counts. Minor road traffic is therefore sampled
much less intensively than major road traffic: allowing for the grossing-up
process, 3.1 billion vehicle-kilometres per year are represented on average by
each minor road ATC site, compared with 2.3 billion for each major road ATC
site. The figures are shown in Table 5.4(i).
4.2
The main reason for the less intensive sampling is that there is much less
interest in obtaining AADFs for individual sections on the minor road network.
Links on the major road system provide for inter-urban travel and act as major
distributors within towns and conurbations; as such they have high flows and
attract considerable attention when problems of capacity, safety or environmental
concerns arise. Minor roads, by contrast, have lower flows and the problems of
individual sites attract only local interest. Therefore the need for AADFs on
individual sections is normally purely statistical: as a means of estimating the
overall volume of minor road traffic within a local area or region.
4.3
Nevertheless with over one-third of the total traffic, any problems arising from the
estimation of minor road traffic will have a significant effect on the estimates of
total traffic. It is clear that minor road traffic is difficult to monitor satisfactorily,
mainly because traffic flows vary considerably over its 340,000 kilometres and
any sample of sites will represent only a tiny fraction of the whole.
4.4
One class of vehicle for which minor roads are particularly important is pedal
cycles. If we improve minor road traffic estimates, then ipso facto we improve
pedal cycle traffic estimates. But pedal cycle traffic is different from general
minor road traffic and there are other dimensions to be considered. These are
handled separately in Annex 5.5.
5
Alternative approaches to monitoring minor roads
5.1
In Table 5.4(ii) we review some of the alternative approaches that have been,
and might be, adopted for minor roads. In each case we summarise the
methodology and the positive and negative aspects. Every approach involves
some sort of compromise.
The ‘standard’ approach
5.2
The ‘standard’ approach to estimating minor road traffic was used when the Core
and Rotating Censuses were established in the early 1980s. The approach was
to treat minor roads the same way as for major roads. A random sample of sites
– of the order 3000 initially but growing over the years to over 4000 – was
selected on all types of minor roads. Each year one-sixth of the sample was
counted for one day and AADFs estimated using expansion factors derived from
the Core Census sites for the same class of road or site. For sites counted in
earlier years AADFs were updated to the latest year using growth factors from
the Core Census. The resultant set of AADFs was then weighted together to
provide the estimated vehicle-kilometres.
5.3
The difference between the procedures for minor and major roads is that for
major roads it is possible to rotate through every major road link over a cycle of
five or six years, whereas with minor roads we can handle only a small sample of
sites. This is arguably the weakest feature of the ‘standard’ approach. The
sample counted will vary from year to year and in a purely random sample we
have no control over the types of sites we select, which might range from busy
shopping roads to cul-de-sacs. Often the sites will have very low flows which are
likely to exhibit large percentage changes. Also of course these low flow sites
offer poor value for money in terms of traffic counted and can give rise to
comment in the media.
5.4
The other problem is that we have very few ATC sites on minor roads from which
to compute expansion and growth factors. All of this implies that we run the risk
of obtaining volatile estimates from one year to another, based on the traffic
changes at just a few sites.
5.5
Prior to the Core and Rotating censuses being established in the early 1980s,
there was a need for major benchmarks to be established every 5 to 7 years.
These covered a wide range of sites, to ensure that the set of monitoring sites
represented average flows correctly within each road-class. With the revised
systems there was less need for benchmarks on major roads, because a
complete sample of links was covered over a cycle of about 6 years. But for
minor roads benchmarks have to be undertaken periodically.
Using Core Census trends
5.6
Over the years there have been attempts to improve the robustness of minor
road estimates. In 1999 a benchmark study was undertaken, examining in detail
the make-up of the minor road system in different types of locality and using a
relationship between flows on principal roads and flows on different types of
minor road to establish more reliable figures at the local authority level. As a
result the overall level of minor road traffic in 1999 was obtained robustly and
that the main need from that point onwards was to monitor the year-on-year
changes accurately, based upon that baseline. [Some minor revisions were
made to the earlier minor road figures for the 1990s, for consistency].
5.7
Two different ways of monitoring year-on-year changes have been considered.
The first is to use the Core Census ATC sites to estimate the trend for the minor
road categories, which are

Minor rural roads in holiday areas with an estimated AADF of up to 399 [5
ATC sites]

Minor rural roads in holiday areas with an estimated AADF of 400 or
more [5 ATC sites]

Minor rural roads in all other areas with an estimated AADF of up to
2,499 [7 ATC sites]
5.8

Minor rural roads in all other areas with an estimated AADF of 2,500 or
more [8 ATC sites]

Minor urban roads in all areas except Greater London [9 ATC sites]

Minor urban roads in Greater London [19 ATC sites].
This procedure has been used from 1999 to 2003. It is comparatively
straightforward to work out growth-factors from the Core Census site data and
apply these to the baseline vehicle-kilometres, and there is no need for AADFs to
be calculated. However, the trends within some of these categories are based
upon very few sites and there is a risk that oddities could arise for certain types
of vehicle which tend to have low flows on minor roads.
Using a panel of higher-flow sites
5.9
When the 1999 benchmark was carried out, a special panel of 2300 sites was
selected with the intention of using these to track the year-on-year changes from
the 1999 baseline. The sites were chosen to avoid the very low-flow sites, on
the basis that the main changes in flow would be exhibited more clearly at the
higher-flow locations. The flows at the panel sites would be compared with the
baseline to provide updating factors for each year.
5.10 Various methods have been suggested for calculating the trends using the panel:

One can use either AADFs or the raw counts. The latter are taken on the
same weekday and the same week in each year, so should provide a
like-for-like comparison. Using AADFs takes into account the actual
seasonal patterns at the ATC Core Sites;

One can use the change in mean flow or median flow within each
category to estimate the change from the baseline. It is usual to use the
mean in applications where the results for each stratum are weighted up;
but given that the panel is not a random sample of minor roads, there is
also a case for using the median, which is less affected by extreme
observations;

One can work out the percentage change at each site and take the mean
or median of those percentages changes as the overall change for each
road category. This method gives equal weight to each site in a stratum,
regardless of its flow.
5.11 A wide range of different figures can be obtained from these methods, which is
one reason why this panel was not used to derive minor road trends between
1999 and 2003. [However, the change from 2003 to 2004 was based upon
these sites, using AADFs]. There is also some concern that the concentration on
higher-flow sites could bias the estimates. Nevertheless the panel approach may
offer the best method of maintaining stability in the estimates of year-on-year
trends, while covering a reasonably wide range of locations, and we suggest a
modified panel below.
Treating class B sites like major roads
5.12 Before we resume the theme of a panel survey, we can consider a completely
different approach for class B roads. These roads have higher flows than other
minor roads and account for a slightly shorter length than class A roads. Could
we consider a rotating sample of B road links, similar to that used for A roads?
5.13 If we were to divide the B road network into links between junctions with other B
roads or with A roads or motorways, then we might find we had about 9-10,000
links in total. If this network were covered over a 6 year cycle, that would imply
about 1500 B road links to be counted each year. This compares with only about
900 manual counts in the current panel. Clearly an expansion of this order
would not be feasible. An alternative would be to cover only 1 in 5 [say] of the
links in our total sample [selected by a systematic sample], and to count 1 in 6 of
these each year, ie about 300-330 per year. By arranging a systematic sample,
we would obtain a wide range of sites each year, although the sample would not
be very large in each region. We would have expansion factors based upon the
8 rural and 17 urban ATC sites on B roads, and therefore be able to produce
reasonably reliable AADFs.
5.14 If we were to treat B roads this way we would in effect be inventing two additional
expansion factor categories: rural B roads and urban B roads. This of course
would have implications for the treatment of other minor roads. The decision
whether or not to treat B roads as an additional category would depend inter alia
upon the standard errors of minor road AADFs produced by the alternative
methods.
A stratified panel of sites based upon a new benchmark
5.15 The potential use of the panel of sites was discussed above. We expressed
some concerns that the selected sites were not representative of sites in general,
being based mainly on higher-flow sites.
5.16 The last benchmark analysis of minor roads was for 1999 and it is time to
consider an update. To do this we could repeat the 1999 methodology which
looked in detail at certain prototypical local authority areas. But perhaps this
process needs a more systematic approach, such as that described below.
5.17 We need to find out more about the taxonomy of minor roads, especially those in
the unclassified stratum: for example the relative proportions of cul-de-sacs,
other short links on housing estates, and the longer and busier types of
unclassified road. [Alternatively we could define the sub-types simply to be high,
medium and low flow-levels]. This would be impracticable for the whole network,
but would be feasible if we took random samples of kilometre squares in each
region, the samples being taken from conurbations, other urban areas and rural
areas. Within those squares we would itemise and classify each link, probably
with a different classification system for urban and rural areas. We would also
count traffic on these links; for the very low-flow sites, the counts could be very
short. This gives us a profile of average flows for each type of link. The sampled
squares for each region can be weighted by road length and grossed up to the
total regional road-length to provide an estimated profile for the sub-types of link
we have found. This would also give us the basis for selecting a panel of sites.
5.18 An alternative approach would be to examine in depth the road system defined
by an ‘envelope’ of major roads. Under this approach we could also examine
whether flow levels on minor roads varied systematically with distance [or some
other measure of separation] from the major roads forming the boundary.
Another approach could be to associate flows with local land-use data, provided
that the latter data could be related to the network.
5.19 The panel would not be a simple random sample within each road-class, but be
stratified by sub-type. According to stratified sampling theory, the optimum
weighting within each stratum would be determined by the product of the ‘size’
and the standard deviation of flow [SD]. The ‘size’ of the stratum is given by the
vehicle-kilometres, or the average flow multiplied by the road-length. Earlier
work on the 1999 benchmark suggested that the variance of minor road flow was
of the order {flow}0.75, so that the SD would be of the order {flow}0.4 . Therefore
the sample size within sub-type j would be of the order
{Flow} *{length} * {Flow}0.4 = {Flow}1.4 * (length}.
5.20 This would imply that the panel was weighted towards strata with higher flows;
road-length would also be a factor but flow-level would predominate. Some lowflow links would be covered, but these would not need to be counted for a full 12
hours.
5.21 The objective would be to estimate each year the growth in average flow within
each stratum. This growth would be applied to the benchmark vehiclekilometres. The counts would be taken on the same day-of-the-week and the
same week each year, to minimise any seasonality effects.
5.22 SR2 have used the panel data to estimate the growth between 2003 and 2004,
using AADFs to establish the year-on-year change. The aim is to obtain as
accurate estimate as possible for minor roads in total, so that the projections of
vehicle-kilometres based upon the 1999 benchmark are as close as possible to
those eventually obtained from the next benchmark; in that way we would
minimise any revisions to the back-series.
5.23 It is not clear from the information currently available, which method will achieve
this most effectively. One way of addressing this would be to select the method
with the smallest standard errors. When the panel survey data are used then
we need to be assured that growth-rates were unbiased, and this may be more
likely if we work with AADFs rather than raw data, because the AADFs take
account of actual seasonal variation. However, once a new benchmark had
been obtained we would be able to establish by hind-sight which methodology
would have proved most accurate in predicting the changes between the two
benchmark years.
Alternatives to a Benchmark Survey
5.24
Benchmarking has traditionally involved a large-scale survey, designed to exhibit the
full range of flows on the minor road system. However, a full-scale survey might
involve 10,000 to 20,000 sites and would be very expensive.
5.25
In the section above we have suggested various approaches to sampling that would
help to build up a profile of the minor road network. The challenge is to devise
procedures that allow this profile to be established robustly, but also offer good
value-for-money.
6
Summary of Recommendations
6.1
The following recommendations are proposed for the monitoring of minor road
traffic:

A further benchmark exercise for minor roads is recommended as soon as
is practicable. The extent of any data-collection during this benchmark will
need to be determined, with a view to obtaining good value-for-money.
Existing information will need to be exploited to the maximum;

Like its counterpart for 1999, the benchmark will need to apply in-depth
studies in sample areas within urban and rural areas in each region, to
estimate the main sub-types within each road-class and the corresponding
road-lengths and average flows;

Based upon the benchmark strata, a panel of minor road sites should be
established, where the number of sites in each stratum within each region
will be influenced mainly by the average flow, but also by the relative length
of road;

The panel sites should be counted for one 12-hour day each year, using the
same day-of-the-week and week within the year. Low flow sites would be
counted for shorter periods;

Until the new benchmark is processed, SR2 should continue to use the
existing panel of higher-flow sites. The method which produces the lowest
overall standard errors of year-on-year trend is to be preferred;

When the benchmark is complete, it will become clear which method of
estimating year-on-year trends was most successful in ‘predicting’ the
outcome. The lessons from that comparison should then inform the
methodology to be used over the following years. The expectation is that
the panel method will prove the most robust, but there are concerns about
possible bias, and there are different methods of estimation to be compared;

As a separate consideration, SR2 should review the possibility of treating B
roads as a further class of major roads, where a network of links is
established and a rotating sample of links is surveyed over a cycle of [say] 6
years. This should improve the estimates of class B traffic, but would incur
extra expense; possibly there is an acceptable compromise position that
offers good value-for-money;

Pedal cycle monitoring is considered separately in Annex 5.5.
TABLE 5.4(i) Major and Minor Traffic and Road-length
Road Class
Motorway
Vehicle kms
(millions)
92899.6
Length
(kms)
3476.4
Average flow
Number of
(veh/day)
automatic counters
73214
60
Number of
manual counts
1026
Vehicle-km covered
per ATC
per MCC
1548
91
Rural A roads
Urban A roads
139428.6
81634.2
35516.8
11115.7
10755
20121
46
29
7707
8857
3031
2815
18
9
Rural B roads
Urban B roads
20666.0
15760.3
24546.5
5641.5
2307
7654
8
17
542
326
2583
927
38
48
Rural C roads
Urban C roads
22166.3
18372.6
73878.4
11098.0
822
4536
9
5
665
403
2463
3675
33
46
Rural unclassified
Urban unclassified
21587.4
77811.9
112231.5
114816.1
527
1857
7
10
607
1761
3084
7781
36
44
Sub-totals
Major roads
Minor roads
313962.4
176364.5
50108.9
342212.0
17166
1412
135
56
17590
4304
2326
3149
18
41
All roads
490326.9
392320.9
3424
191
21894
2567
22
Table 5.4(ii) Approaches to estimating minor road traffic
Data sources
Positive aspects
Negative aspects
1. Standard approach [as Should produce unbiased Many low flow sites:
originally proposed]. Use estimates
over
longer counting these is poor
AADFs from a set of term, provided that the value-for-money and gives
random sites on minor total number of sites is volatile results. Volatility
roads, a proportion of large.
will also arise because the
which is counted each
sites will vary from year to
year. AADFs are derived
year and because the
using
Core
census
number of minor sites in
expansion and growth
the Core Census is small.
factors.
2. Use Core Census only. May produce reasonably Based upon so few Core
Use minor roads Core stable trends for the main census sites, the trends
sites to give year-on-year vehicle-classes.
Method are
unlikely
to
be
trends. These are then currently used in SR2.
representative. A firmer
applied to a benchmark
grounding is needed.
total of minor-road vehicleAlso depends upon a
km.
reliable
benchmark
of
minor roads.
3. Use a panel of higher- Should produce stable May be biased because
flow sites.
This panel estimates, because panel panel
represents
only
provides various methods is fixed. Can also arrange high-flow sites. Different
to estimate year-on-year for each site to be counted methodologies
produce
change: using mean or at fixed day/week to allow different results. Does not
median % change, and raw
counts
to
be handle seasonal effects
based upon either raw compared.
Avoids the properly, even if AADFs
data or AADFs. [Such a volatility of low-flow sites.
are
used,
because
panel was constructed
expansion
factors
are
around 2001, and annual
based upon very few Core
counts taken; but results
sites.
have not yet been used in
Also depends upon the
minor road estimates]
1999 benchmark of minor
roads.
4. Treat B roads like
major roads. The B road
network would be divided
into links, and a proportion
counted each year over a
cycle of [at least 6] years.
AADFs would then be
calculated as for A roads.
Should produce stable and
unbiased estimates, based
upon a systematic sample.
For many B roads, AADFs
would be of interest in their
own right. If it was too
expensive to count all
class B links over a cycle,
we could decide to cover a
random fraction [say half]
without drastically affecting
accuracy.
5. Use a stratified panel As with 3. above, should
of sites, taking account produce stable trends,
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 105
Would probably increase
costs unless we were
prepared to accept a very
long cycle [eg 15 years]
over which to ‘rotate’ the
sites.
Increasing the
number of annual one-day
counts would also imply
increasing the number of
Core sites, otherwise the
AADFs
would
be
unreliable.
Similar to 3. above, but
should be less likely to be
Data sources
of different ‘sub-types’ of
link.
The panel would
represent the different subtypes within the urban/rural
B/C/UC road-classes. The
number of sites in each
stratum
would
be
governed by its size and
the standard deviation of
flow.
Positive aspects
using various methods of
estimation. Takes some
account of low-flow sites.
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 106
Negative aspects
biased.
Also depends upon a
reliable
benchmark
of
minor roads, to be taken in
2006/7.
Annex 5.5
CYCLE TRAFFIC STATISTICS METHOLOGY
7
Introduction
7.1
Under this heading we consider ways in which cycle monitoring might be
improved. It is generally accepted that cycle counting and cycle traffic
estimation sits uneasily alongside general traffic monitoring. Cycle traffic
represents only a small proportion of total traffic, so that it has not been
thought worthwhile to adopt distinctive monitoring and estimation procedures
to cover it. As a result there is a perception that cycle trends from the RTS
system are volatile, producing erratic changes from year to year, and that
these are possibly biased.
7.2
Below we consider ways in which cycle traffic statistics could be made more
robust, while at the same time developing the monitoring programmes
undertaken by SUSTRANS and by local authorities. Inevitably this means
treating cycle traffic as a phenomenon in its own right, rather than as an
adjunct of general traffic monitoring.
7.3
If the costs of a programme dedicated to pedal cycle monitoring are deemed
to be too high, then some compromises will be necessary: in effect cycle traffic
will have to continue to be counted alongside general traffic.
8
User Requirements and Interests
8.1
The main policy interest of central government is to increase the amount of
cycling, in absolute terms and as a percentage of all journeys. Cycle traffic
estimates [cycle-kilometres] are also required to measure exposure to
accidents, enabling the numbers of cycle casualties and cycle involvements in
accidents to be expressed as casualty- or involvement-rates per cyclekilometre.
8.2
LAs are asked to promote cycling and improve facilities for cycling in their
LTPs, and LEAs are to encourage more cycling to school, mainly at secondary
level.
8.3
SUSTRANS is a charity established to encourage cycling, inter alia through
the promotion of cycles routes. It is supported financially by government.
8.4
All of the above interests imply the need to monitor the amount of pedal cycle
usage from year to year. Estimates of cycle usage are needed at the
aggregate level [such as all GB and for regions of GB], at the local level
[trends within LAs] and at individual facilities [cycle lanes and routes].
8.5
Cycling takes place mainly on public roads but also on dedicated cycle routes,
in parks, by canals and rivers, and on footways at the side of roads. The latter
is illegal except for small children, but is observed to be quite common. There
is interest in distinguishing where cycling takes place, as a guide to investment
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 107
and provision of cycling facilities. We assume that investigating the extent of
illegal cycling on footways is outside the scope of this review.
9
Current Monitoring Systems
9.1
The National Travel Survey should in theory provide the best means of
covering total cycling activity. But the sample size of active cyclists amongst
overall respondents is small and it appears that not all trips are recorded. For
2002 the NTS found an average of 33 miles per person, which equates to 2.9
bn veh-km overall. This is actually less than the estimate from the Road
Traffic Statistics surveys, which reported 4.4 bn veh-km on public roads alone.
It appears that only direct observation in places where cycling takes place can
provide full coverage.
9.2
The RTS surveys operate on the basis of sampling sites on different types of
public road. They do not cover cycle paths on non-public roads, or other offroad locations. Nor do the traffic counts on public roads include cycles using
the footway. In 2003 it is estimated that 83% of cycling on public roads occurs
on minor roads [below class A] and the vast majority of usage is in urban
areas. Within the major and minor road classes it is found that cycling is
concentrated on a few sites with relatively high flows; at the majority of sites
cycle flows are very low and volatile.
9.3
The RTS surveys, however, are designed to monitor general traffic trends,
rather than cycling per se. The main preponderance of monitoring sites, both
on the Core/ATC census and the annual one-day counts, is towards major
roads, which carry two-thirds of total traffic. As a consequence the estimated
year-on-year trends in cycle traffic have proved to be volatile and subject to
revision. If the RTS surveys were to be designed specifically to cover cycling,
the balance of survey effort would be entirely different. They would be
focussed principally on minor roads and within the minor road classes they
would attempt to distinguish sites where cycling tends to be concentrated. In
this sense some of the problems of cycle monitoring are common to the
monitoring of minor road traffic in general [see task 9c]. But even within the
minor road classes [B, C and unclassified roads] the places where cycle traffic
is high do not normally coincide with places where general traffic is high.
9.4
SUSTRANS undertakes surveys at sites on the National Cycle Network.
Many of these sites are on cycleways off the public road system. The
monitoring sites are concentrated mainly at places where cycle usage is
believed to be highest. In this way they can detect the main changes in cycle
usage, although it is more difficult to estimate average flows over the whole
network when most monitoring is at the high-flow sites. SUSTRANS has
limited resources to undertake the processing and validation of the survey
data, and this may inhibit the expansion of their monitoring system.
9.5
LAs are encouraged to undertake their own cycle monitoring. DfT suggests –
in its advice notes - that it is best to concentrate LA monitoring effort on roads,
corridors and screenlines where higher flows might be expected. This is
understandable, given the poor value-for-money involved in counting small
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 108
numbers of cycles and the likelihood that the greatest increases – at least in
absolute terms – are likely to occur at the higher flow sites.
9.6
Currently most cycle counts are undertaken manually. Whilst automatic
counters are available, their performance in mixed traffic conditions is patchy;
cycles may be missed if they do not cross the detectors at the expected place
near the edge of the road, and ‘ghost’ signals can arise from larger vehicles
nearby. Trials of cycle detector equipment are underway in two LAs. Also a
general-purpose counter which may be able to detect cycle usage at certain
sites is being tested.
10
Objectives
10.1 Given this background, the main objectives of a review of cycle monitoring can
be summarised as follows:

To improve estimates of the year-on-year trends on public roads

To improve estimates of the absolute level of cycle usage [cyclekilometres] on public roads

To estimate the proportion of cycling that takes place away from public
roads

To estimate trends on the SUSTRANS/National cycle Network

To estimate trends within LAs.
10.2 All of the above objectives have to be considered in the light of the costs
involved and the weight of policy interest.
11
Options for Improvement
11.1 A number of approaches can be adopted to improving estimates of cycle
usage, based upon surveys within the DfT RTS system and building on the
SUSTRANS and LA initiatives. The main approaches are summarised in
Table A, which shows how each approach might be used to estimate trends
and absolute levels of cycling; some of the arguments supporting each
approach [positive aspects] and casting doubt on it [negative aspects] are
shown in the table.
11.2 The approaches are as follows:

1. The ‘standard approach’. We could adopt the standard approach
to estimating AADFs, ie using a set of one-day counts of all vehicletypes which are converted to AADFs by expansion factors derived from
the Core/ATC sites. Under this approach we would have a set of
about 4500 thousand minor road sites, a fraction of which were
covered by rotation each year; sites counted in earlier years would be
updated using Core/ATC trends. Whilst this might in the longer term
produce unbiased estimates, it has already been found to produce
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 109
volatile figures for minor road traffic in general, mainly because the
sites change from year to year. It is even less reliable for cycle traffic
because the balance of sites does not reflect the balance of cycle
flows.
Nevertheless it appears to estimate effectively the small amount of
cycle traffic that uses major roads.

2. Using Core Census ATC sites. We could use the Core Census
ATC data to measure the year-on-year trends, based upon the current
set of major and minor road sites. Provided that these sites are
representative, then the trend estimates will be unbiased. Provided
that we also have a reliable benchmark for a past year, this also gives
us a measure of absolute cycle-kilometres. However, the ATC sites
are generally weighted towards major roads and are probably
unrepresentative of cycle traffic even on minor roads.
Therefore this approach seems unsuitable, although it was, up until
recently, being used to estimate minor road traffic in general, and
hence the bulk of cycle traffic.

3. Using a Minor Road panel. Under this option we would maintain a
fixed panel of one-day counts for minor road sites, rather than a
varying set, and use these to estimate year-on-year trends for traffic in
general. This is the approach we have recommended for estimation of
general traffic on minor roads. This approach would tend to reduce the
volatility problem, but the panel of minor road sites will tend to be
representative of general traffic, rather than of cycle flows. The
correlation between general traffic flows and cycle flows is not very
high.

4. Fixed Panel of Cycle Monitoring Sites on Minor Roads. Here we
would devise a special panel of minor road sites, based upon historical
benchmarks and other data, where the panel was slanted towards
sites known to have higher-than-average cycle flows. [This panel
would be entirely separate from any panel used for general traffic
monitoring, as in option 3 above, although some of the sites might
coincide]. These sites would be counted on one weekday per year in a
neutral month. By comparing counts year on year [assuming that
these were taken at similar times] trends in cycle usage could be
obtained using a larger cycle-flow count than could be obtained from
the more general sites used in options 1 and 3.
These would produce more stable trend and cycle-km estimates,
although care would need to be taken to ensure that the sample was
indeed representative.

4a. Cycle Panel supplemented by Weekend and Summer/Winter
Counts. Under this variation of 4. we extend the concept of a special
panel of cycle sites, to include weekend counts and counts in non-
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 110
neutral months in the winter and summer at a proportion of the sites.
These are needed to take account of the fact that a large proportion of
cycle usage takes place at weekends [more, for example, than the
proportion of car traffic] and that unusual summer or winter weather
can distort the annual profile; these counts were recommended by an
earlier review.
However, options 4 and 4a are expensive, involving probably over
1000 extra counting sites. And option 4a multiplies the number of
counts by a further factor.

Under options 4 and 4a, we have the option of comparing the raw
count data from year to year, or converting the counts to AADFs using
the Core Census expansion factors. Using AADFs would exploit the
seasonal patterns already recorded by the Core/ATC sites, and hence
reduce the need for comprehensive coverage of weekends and
summer and winter months [although probably some such counts
would remain]. If raw data comparisons were to be used, it would be
necessary to do the counts on the same weekdays and weeks, to
avoid some of the complications of seasonal and day-to-day variation.
There would also be benefits in standardising the counting days if
AADFs were used.

5. SUSTRANS monitoring.
The numbers of sites covered by
SUSTRANS monitoring could be expanded, to provide better
coverage. Ideally the set of sites would include some lower-flow points
[so that all types of the network are represented] and each site could
be assigned a grossing-up factor to represent the proportion of the
network it covered. There may be a need for central processing of the
data by some agency, eg DfT or a consultancy, in which case further
costs would be incurred.

6. LA monitoring. As the number of sites covered by LA surveys
expands, there will be a need to exploit this large dataset to determine
trends at the regional and sub-regional level. Initially the imperative
would be to draw out the main findings of the LA data, bearing in mind
that the sites concerned might not be representative of all cycling in the
areas concerned.
11.3 For major roads the standard approach [option 1] is sufficient. For minor
roads the most promising approach would appear to be the special panel
surveys of options 4 or 4a. The panel approach depends upon a benchmark
in some earlier year; the most recent benchmark was in 1999 and we have
recommended a further benchmark in 2006. At the same time further
encouragement could be given to expand the SUSTRANS and LA monitoring
programmes. However, all these approaches require considerable extra
expenditure [mainly on manual enumerator staff] and might be deemed to be
poor value-for-money compared with the more general traffic monitoring; for
example, the cost per vehicle counted would be much higher.
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 111
11.4 In the section below we suggest an approach based upon a fixed panel of
cycle-monitoring sites, such as option 4a. We consider how this panel might
be made statistically representative and then consider whether it might be
possible in the longer term to integrate this panel with the SUSTRANS and LA
panels.
12
Establishing a Panel of Cycle Monitoring Sites on Minor Roads
12.1 The objective would be to establish a ‘Cycle Panel’ of up to 2000 sites on
minor roads. Most of these would be on minor urban roads, where the
objective would be to classify sites within each region or sub-region by [a]
urban road-class B, C or UC and [b] within those classes by high /medium /low
pedal cycle flow. Because rural cycle flows are very low, the selection of rural
sites would be much more aggregated, being based only on road-class.
12.2 For the urban minor roads stage [a] is straightforward because road-class is
known [although OS maps unfortunately do not distinguish C and UC roads].
The classification by flow level is much more difficult because we do not know
the flow until we have conducted a count. However, there exist historical data
based upon samples of minor roads. These could be used to establish a first
estimate of the proportion of sites having high/medium/low cycle flows, with
the dividing lines between these levels defined within each road class. This
initial estimate could be refined by taking a random sample of sites within each
region or sub-region and conducting short counts at these sites.
12.3 We would aim to set up the three tables below. Table B is established from
the historical and other site data, as described above. Table C then applies
the percentages in Table B to the known total mileage in each road-class, to
derive the road-lengths. Table D is an estimate of the average flow, again
obtained from the historical and other data. These figures do not have to be
obtained with great accuracy at this stage, since they will be used only as a
guide to the balance of the panel sample.
12.4 Stratified sampling theory tells us that the optimum weighting for our panel of
urban minor road sites is then governed by the ‘size’ of each cell multiplied by
the standard deviation of the estimate based on that cell. The ‘size’ of our cell
is the volume of cycle traffic, ie the cycle-kilometres or the average flow in
Table D multiplied by the length of road in Table C; this is given by Table E.
Earlier work by SR2 suggested that typically the variance of a daily flow on
minor roads was of the order [flow] 0.75; hence the standard deviation would be
of the order [flow]0.4. If we take the product of the cycle-km and the standard
deviation of daily flow, we get [cycle-km] * [flow]0.4 which is equivalent to
[average flow]1.4 [length of road]. This formula could be used as a guide to
the proportion of cycle monitoring sites that should be in each cell of our panel
for urban minor roads [see Table F]. This means that the number of sites in
one of the flow-level groups would be influenced mainly by the average flow,
but also by the total length of road with this flow-level.
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 112
12.5 For rural minor roads it would probably be sufficient to repeat the process
without the subdivision by flow-level, because the cycle traffic, and hence the
number of sites, will be much smaller. [As already recommended, the
estimates of cycle traffic for major roads could continue to be based upon the
AADFs derived from the standard one-day manual counts].
13
How the Cycle Panel could be used
13.1 The panel would first be established on the basis of data that were already
available, and it must be recognised that the percentages in Table B would be
approximations. However, we know that SR2 did design a panel of cycle
monitoring sites in 2001, and these could be a useful starting point.
13.2 Further information on flow levels on the urban minor road network might
come to hand from LA sources, enabling these figures to be refined. But the
main source of information about the pattern of cycle flows would be
benchmarks undertaken to underpin the monitoring of general minor road
traffic [see Task 7g]. We can therefore expect the balance of counts between
the different strata to be improved year by year, as new sites are added to the
panel and more up-to-date information on flows becomes available.
13.3 We may find that the panel contains a large number of very low flow cycle
sites, although the weights in Table F for the low-flow sites would tend to be
small relative to the high and medium flow sites. We would not expect to
conduct 12-hour counts at these very low flow sites, because they would
represent poor value for money and the changes from year to year would be
very erratic. The best approach may be to take short [eg one-hour] counts in
the morning peak, in order to establish whether the flows continue to be very
low. To establish year on year trends at these sites we might have to assume
that these were the same as for either trends for the other [higher-flow] sites
within the low-flow category or, if there too few of the latter, trends within the
medium flow category.
13.4 In general the approach would be to establish AADFs at the panel sites, based
upon a 12-hour count taken in a neutral month, together with some counts at
weekends and in summer and winter months; expansion factors from the Core
ATC census would also be used in establishing AADFs. Various methods
could be considered to reduce the costs of one-day counts: for example, using
‘rotating’ cycle detection equipment; counting for 6 minutes in every 15 [to
allow a 6-minute count at another nearby site in the same 15-minute period];
or counting only in the peak periods at some sites.
13.5 Having established the AADFs in each minor road ‘cell’, we then multiply by
the relevant road-length to establish cycle-kilometres. At this stage we can
then update Table F to improve the balance of the panel for the following year.
14
SUSTRANS monitoring
14.1 SUSTRANS has already established a panel for sites on the National Cycle
Network [NCN]. It would be of interest to compare trends on this network with
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 113
trends based on a Cycle Panel similar to that described above. There would
be some differences, because some of the SUSTRANS sites are off-road and
the emphasis is on higher-flow sites. Nevertheless we would expect to see a
general coherence between similar categories of sites, and any discrepancies
would need to be investigated, with a view to correcting any imbalances that
existed in either the SUSTRANS or DfT panels.
14.2 In the medium term the aim could be to integrate the SUSTRANS and DfT
panels, to provide a representative panel for the whole cycle network, covering
both public roads and the off-road network [although ‘informal’ tracks such as
those in parks and by rivers and canals could not be covered systematically].
Whether this integration was worthwhile would depend upon [a] whether DfT
policy interest was sufficiently strong to extend the monitoring to tracks off
public roads and [b] whether SUSTRANS felt confident that an integrated
system would provide proper coverage for the NCN.
15
Local Authority Monitoring
15.1 We can expect the number of pedal cycle monitoring sites counted by LAs to
increase, if LAs follow government advice to encourage cycling within their
localities. However, this expected growth in cycle monitoring may only
continue if the planned growth in cycling actually occurs on the ground.
15.2 If the majority of LAs produce annual monitoring reports, there would be a role
for DfT in collating these reports and correlating their findings with those from
its own panel. The LA reports will tend to be based on higher-flow sites, rather
than the full range of locations, so the main findings of interest from a national
viewpoint will be the year-on-year trends. Trends for LAs could be collated
and plotted against trends derived from the DfT’s own cycle Panel. Given
that the emphasis of local monitoring will vary from one LA to another, there
are likely to be contrasting trends.
15.3 It is possible that growth at LA sites will be higher than at more general
locations, because there will be a tendency to monitor sites where
improvements have been made and cycling has been made more congenial.
If so, the comparisons will be of value in demonstrating the potential growth
that can be achieved by local initiatives.
15.4 It seems unlikely that the set of LA monitoring sites could be used directly to
supplement the DfT panel, unless an LA had decided to count at a randomly
chosen set of sites. Possibly in some cases the LA sites could be taken to
represent the ‘high-flow’ cells in the national panel, in which case these sites
could be ‘adopted’ within the national panel and their data used to improve the
national estimates.
15.5 Whether DfT should undertake this role of collating and comparing the LA
trends depends upon continuing policy interest in these comparisons. Given
staff resourcing constraints there could be a trade-off between time spent on
improving national cycling estimates through the national Cycle Panel and in
co-ordinating the reports from the LAs. One point to be borne in mind is that
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 114
information from the LA reports will inevitably be less timely than information
from the national panel, because the LA dataset will depend upon reports
being received from a large number of individual authorities.
16
Summary of Recommendations
16.1 Given a requirement to improve the accuracy and precision of cycle
monitoring and given resources to do this, we would recommend that a
National Panel of up to 2000 cycle sites be established for minor roads. In
particular

this Panel would be a stratified sample based upon flow-levels within
each road-class on urban roads in each region. For rural roads it
would be based upon road-class alone. The flow data will be obtained
from historical benchmark surveys, such as the 1999 benchmark and
the 2006 benchmark we have recommended for 2006;

cycling on major roads would continue to be covered by the general
traffic estimation procedures, ie using AADFs for each link of the
network;

DfT should consider extending the minor road Panel to cover cycle
tracks on the National Cycle Network which are not on public roads,
thus encompassing the existing SUSTRANS monitoring;

provided that sufficient local authorities undertake regular cycle
monitoring, DfT should collate these statistics, comparing trends with
those based on the national Panel. But it seems unlikely that a large
proportion of the LA monitoring sites could be incorporated in the
national Panel.
16.2 All the above recommendations carry cost implications, both for processing
within the Statistics Divisions at DfT and for the dedicated manual counting at
Panel sites. The cost per vehicle counted would be much higher than for
general traffic, and by implication the policy interest in developing the statistics
would have to be strong.
16.3 If the costs of a dedicated Cycle Panel were deemed to be too high, then the
compromise approach would be to rely entirely on improvements to the
general traffic estimates for minor roads. Minor roads are the subject of
discussion in Annex 5.4. Under this option the SUSTRANS programme would
continue independently, because the sites away from public roads would
continue to be covered.
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 115
Table A
Comparisons of Approaches to Pedal Cycle Monitoring
Data Source
Positive Aspects
Negative Aspects
1. ‘Standard approach’. Works reasonably for This approach has proved
A set of ‘rotating’ sites on general traffic on major unreliable for minor roads:
each
road-class
is roads. Can also be used for general traffic and
counted for on-day each for cycle traffic, given that especially for p/c traffic
year and an AADF is only 17% of p/c traffic is
estimated.
Average on major roads.
AADF is multiplied by
road length to give overall
vehicle-km.
2. Use Core ATC data This approach was, until Not satisfactory for p/c
alone to estimate trends recently, being used for all trends, because flows at
based
on
previous minor road traffic trends.
ATC sites are low and
benchmark.
volatile.
3. Use a fixed panel of Would improve coverage Sites suitable for general
minor road sites, as of minor road sites and minor road traffic are not
recommended
under this would benefit p/c in general suitable for p/c
Minor Roads
traffic, 83% of which is on traffic.
Also does not
these roads.
cover cycle-tracks etc off
public roads.
4. Use fixed panel of Would
allows
proper Expensive. Would require
pedal-cycle sites on balance of sites based special p/c sites where
minor roads, counted for upon p/c flow. Could be only cycles were counted,
one day each year.
extended to cover NCN and therefore poor valueoff-road sites.
for-money. The one-day
counts
would
miss
weekends when much
cycling occurs.
Core
Census expansion factors
for cycling would not be
very robust as estimators
of seasonal effects.
4a. As above but with Would
provide
a Would increase costs at
additional weekend and comprehensive picture of least threefold, depending
summer/winter counts
cycling trends, if 2000 or on what was covered.
more sites in the panel.
5.
SUSTRANS Covers sites off public Sites will tend to represent
monitoring
roads. Provides a panel higher-flow locations and
that can be monitored therefore cannot be used
regularly.
to estimate total cyclekilometres
6. LA monitoring
If all LAs participate, If mainly high-flow sites
provides a vast amount of chosen, trends may be
data to estimate trends. biased. And it would be
Would indicate effects of difficult to gross up to
local initiatives.
provide
estimates
of
overall cycle kilometres.
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 116
Table B: % of sites by P/c flow-level and Urban road-class in Region A
Road-class
% of high % of medium % of low flow
All sites in
flow sites
flow sites
sites
region A
B
%BH
%BM
%BL
100
C
100
UC
100
Table C: Length of road by P/C flow level and Urban road-class in Region A
Road-class
High flow
Medium flow
Low flow
All sites in
sites [km]
sites [km]
sites [km] region A [km]
B
LB_H =
LB_M =
Etc
LB
LB * %BH
LB * %BM
C
LC
UC
LUC
Table D: Average P/C flow by flow-level group and Urban road-class in Region
A
Road-class
High flow sites
Medium flow
Low flow sites
[cycles/day] sites [cycles/day]
[cycles/day]
B
FB_H
FB_M
FB_L
C
UC
Table E: First estimate of Cycle-kilometres by flow-level group and Urban roadclass in Region A
Road-class
High flow sites
Medium flow
Low flow sites
sites
B
CB_H
= FB_H
CB_M = FB_M CB_L = FB_L *
* LB_H
* LB_M
LB_L
C
UC
Table F: Target ‘weights’ for Urban road-class panel in Region A
Road-class
High flow sites
Medium flow
Low flow sites
[cycles/day] sites [cycles/day]
[cycles/day]
B
WB_H = [cycle- WB_M = CB_M *
WB_L = CB_L *
km] * [av.flow]0.4 =
[FB_M]0.4=
[FB_L]0.4=
0.4
1.4
CB_H * [FB_H] =
[FB_M] * LB_M
[FB_L]1.4 * LB_L
[FB_H]1.4 * LB_H
C
UC
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 117
Annex 5.6
Road length and vehicle kms by class of road 2005
Percentage
of roadlength
Percentage of
vehicle-km
1%
Total million
vehicle-km
p.a.
97,049
Road Class
Motorway
Length
(kms)
3,523
Rural A roads
Urban A roads
35,530
11,138
9%
3%
141,532
81,779
28%
16%
Rural B roads
Urban B roads
24,639
5,550
6%
1%
21,450
15,873
4%
3%
Rural C roads
Urban C roads
73,581
10,878
19%
3%
23,046
18,304
5%
4%
109,426
28%
22,309
4%
113,757
29%
78,299
16%
All major roads
All minor roads
50,192
337,832
13%
87%
320,361
179,280
64%
36%
All roads
388,023
100%
499,640
100%
Rural
unclassified
Urban
unclassified
19%
QUALITY REVIEW: KEY STATISTICS TO INFORM THE SAMPLING
STRATEGY
Vehicle kms
Length
Average
flow
Road Class
(millions)
(kms)
Motorway
97,049
Rural A roads
Urban A roads
(veh/day)
Number of
automatic
counters
Number of
manual count
points1
3,523
75468
60
141,532
81,779
35,530
11,138
10914
20115
Rural B roads
21,450
24,639
Urban B roads
15,873
Rural C roads
Vehicle-km covered
per ATC
per MCC
1072
1617
91
45
30
7880
8941
3145
2726
18
9
2385
8
2681
31
5,550
7835
18
694
404
882
39
23,046
73,581
858
9
2561
29
Urban C roads
18,304
10,878
4610
5
803
459
3661
40
Rural unclassified
Urban
unclassified
22,309
109,426
559
7
736
3187
30
78,299
113,757
1886
10
2017
7830
39
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 118
Sub-totals
Major roads
Minor roads
320360.5144
179279.7
50191.71
337831.5
17487
1454
135
57
17893
5113
2373
3145
18
35
All roads
499640.2
388023.2
3528
192
23006
2602
22
1 Only a sub-set of these are counted every year.
Rural
Urban
207645.7
180378
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 119
Annex 5.7
Vehicle kms 2005: proportions by vehicle and road type
Vehicle-type
Pedal cycles
Motorcycles
Cars and taxis
Vans
Rigid 2-axle lorries
Rigid 3-axle lorries
Rigid 4+ -axle lorries
Articulated 3 and 4-axle
lorries
Articulated 5-axle lorries
Articulated 6+ -axle lorries
All Lorries
Buses/coaches
Percentage of Vehicle-km on
Major
Minor
Roads
Roads
15%
85%
48%
52%
64%
36%
60%
40%
78%
22%
75%
25%
83%
17%
93%
98%
96%
87%
49%
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 120
7%
2%
4%
13%
51%
Percentage of Vehicle-km on
Urban
Rural Roads
Roads
23%
77%
42%
58%
51%
49%
54%
46%
62%
38%
68%
32%
72%
28%
79%
86%
84%
71%
32%
21%
14%
16%
29%
68%
Annex 5.8
DEMAND FOR DISAGGREGATED STATISTICS
1.
Introduction
2.1
In the course of consulting users about information needs, some respondents
suggested that more detailed results should be published by DfT.
2.2
In this note we consider the benefits and costs involved, taking account of the
reliability and precision of the results, and the resource implications for SR2.
2.
AADFs for weekends and weekdays
2.1
Currently SR2 produce AADFs which are literally Annual Average Daily
Flows, based upon every day of the year, including weekdays, weekends and
public holidays. This has the merit of summarising all traffic for the year, and
can be translated directly into vehicle-kilometres by multiplying the AADF by
the total length of road. But because weekdays and weekends are so
different in their traffic make-up and profile, the AADFs themselves do not
usually represent typical travel days.
2.2
Annual Average Weekday Flows [AAWDFs] could be calculated using the
ATC data. This would involve SR2 in deriving a separate set of expansion
factors, based upon weekdays only. These factors would relate the average
annual weekday flow to the 12 hour count on the counting day; for these
factors we might choose to exclude weekdays that were bank holidays. The
additional costs involved would be the resources involved within SR2 in
deriving a separate set of factors.
2.2
AAWDFs would probably be more reliable [have smaller standard errors] than
AADFs, because the variable element resulting from weekend traffic would be
removed from the expansion factors.
2.3
Annual Average Weekend Flows [AAWEFs] could then be derived
algebraically from the AADFs and the AAWDFs, given that we know the
numbers of days of each type, using the relationship
AADF = {Number of weekdays * AAWDF + Number of weekend days *
AAWEF}/365.
2.3
However, a more reliable way to estimate AAWEFs would be to derive special
weekend expansion factors from the ATC data, based upon the ratio of the
annual average weekend flow to the average flows on ‘neutral month’
weekends. We would then have two sets of expansion factors, one based
upon weekdays and one based upon weekends, and these would replace the
current set of factors covering both types of day. Each of these expansion
factors would be more reliable than the single expansion factors because they
would each be based on more homogeneous data. But this would also imply
that we carried out manual classified counts on Saturdays and Sundays at the
sites where we required AAWEFs, implying additional costs for these counts.
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 121
It seems unlikely that these counting costs could justify the improved
precision of the AAWEFs.
3.
16-hour AADFs
3.1
Traffic engineers and planners frequently operate with 16-hour flows [06002200]. It is feasible to produce 16-hour AADFs [AA16DFs], by calculating a
simple adjustment factor. This adjustment factor is the ratio of the annual
average 16 hour flow and the average 24-hour flow, both of which can be
derived directly from ATC data. It has to be derived entirely from the ATC
data because the manual count data cover only the period 0700-1900.
3.2
Special adjustment factors could be derived to derive 16 hour AAWDFs and
AAWEFs, again using the ATC data.
4.
Estimates of annual average flows for each hour
4.1
Traffic management and road safety considerations sometimes require an
hourly profile of traffic for a link on the network. For hours during the weekday
manual counting period [0700-1900] the manual counts themselves provide
the best estimate of the hourly profile. For the other, night-time, hours [19000700] the only information comes from the annual average weekday profile
derived from the ATC data.
4.2
So the procedure would be to start with the Annual Average Weekday Flows
[AAWDFs] and divide this into the flow during 0700-1900 and the rest, using
the proportions given by the annual ATC data. Then the total flow during
0700-1900 would be split into single hours using the manual count profile.
And the rest would be split using the annual ATC data.
4.3
This procedure would be worthwhile only if we also produced AAWDFs,
because the individual hourly flows would differ markedly between weekdays
and weekends; for example, the average flow from 0800-0900 represents the
morning peak for weekdays but a quiet period for Saturdays and Sundays.
There is no prospect of producing hourly traffic estimates for weekends
unless we were to carry out manual counts at weekends at the sites
concerned.
5.
More detailed geographical and locational breakdown of traffic volumes
5.1
Some users have asked for more detailed geographical breakdowns,
such as those produced in Scottish Transport Statistics [STS].
5.2
STS provides

[a] major road vehicle-kilometres in Scotland by vehicle-type and roadclass;

[b] major road vehicle-kilometres by over 30 LA s and 6 road-classes;

[c] average flows for each of the Scottish ATC sites and each month.
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 122
5.3
[a] is already published by DfT at the GB level. Could this be extended
reliably to the regional level, bearing in mind that there are 11 regions,
including Scotland? On the whole the answer should be yes, with the
possible exception of some minor road categories in some regions. Standard
errors of these estimates should be calculated and published.
5.4
[b] is calculated by DfT but the figures for LAs are not broken down by roadclass in DfT publications. No doubt some of the individual cells would have
wide standard errors and the estimates might vary considerably from year to
year; these would need to be examined carefully. But the presumption would
be in favour of publication, together with the standard errors, if necessary with
‘health warnings’.
5.5
[c] is done by DfT and Highways Agency, but there is no published
breakdown by month. Since the ATC data are validated site by site, it seems
reasonable to show the data month by month. The question arises whether
the figures should relate to weekdays, weekends or all days in the month.
6.
Publishing hourly count data for each counting point
6.1
Some users have asked for the raw count data obtained at ATC sites or in
manual counts, so that they can undertake their own specialised analyses of
the data.
6.2
There are good reasons to support such requests, because DfT undertake
quality checks on the data and can therefore release them to public scrutiny.
However, we must emphasise that they relate only to the single day [in the
case of MCCs] or some longer period [for ATCs] when the data were
collected.
6.3
Obviously there would be costs involved for DfT staff. Given that release of
the raw data is not a normal output in standard DfT publications, it would be
reasonable to make a charge for such data; this would be the full cost of
retrieving and delivering the data to the enquirer, not their collection cost.
7.
Summary of Recommendations
7.1
The main recommendations on publishing disaggregated results are

Annual Average Weekday Flows [AAWDFs] should be produced, provided
that the resource implications within SR2 can be accommodated. ;

Annual Average Weekend Flows [AAWEFs] could be obtained as a byproduct of AADFs and AAWDFs. But it would not be worthwhile refining
AAWEFs by undertaking manual counts at weekends;

16-hour AADFs, AAWDFs and AAWEFs can be derived straightforwardly
from the corresponding 24-hour figures;

As with AADFs, standard errors of the above disaggregated figures should
be produced;
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 123

Annual average hourly flow profiles for weekdays [and weekends] can be
derived straightforwardly for individual links, provided that AAWDFs [and
AAWEFs] are produced;

More detailed geographical breakdowns, such as those produced in
Scotland, should be produced and published along with their standard
errors;

Raw data from ATC and manual sites should be made available on
request, provided that users pay for the marginal costs of retrieving and
distributing these data.
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 124
Annex 6.1
Department for Transport
Road Traffic and Road Length Statistics Quality Review
LA Data Sources: Questionnaire
YOUR DETAILS
Name
Organisation
Job title
Address
Telephone
Fax
E-mail
Your Local Authority:
(mark as appropriate)
a) collects most/all of its own traffic data?
b) outsources the collection of most/all traffic data
c) neither collects nor outsources the collection of traffic data
If you answered "a" please continue with the questionnaire
If you answered "b" please liaise with the company who does your counts to complete the questionnaire
If you answered "c" please return the questionnnaire to us with the box for "c" marked.
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS
125
1
If you collect traffic data regularly for what purpose do you use it?
(mark as appropriate)
DfT road traffic estimates
Monitoring for LA purposes
For other LA purposes (please describe)
2
If you collect traffic data on an ad-hoc basis; for what purpose do you use it?
RAW DATA COLLECTION
3
What method/s do you employ to count vehicles?
(mark as appropriate)
Manual Counts
Automatic Traffic Counters(ATCs): fixed
Automatic Traffic Counters(ATCs) :'mobile' tube
Video counts
Pedal cycle counters
Other (please specify)
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS
126
4
Do you conduct:
Yes
No
Cordon counts?
Link counts?
Other types? (please specify)
5
What type of counts do you conduct?
Fixed means that the same sample is counted every year
Rotating means that different parts of the sample are counted every year covering the whole sample over a number of years
One-off means that the site is counted, not as part of a sampling strategy, but for other purposes
Please mark all applicable boxes
Fixed
Motorways
'A' Roads
'B' Roads
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS
127
Rotating
One-off
counts
None
'C' Roads
'Unclassified' roads
Urban
Rural
How do you define urban and rural?
6
How do you count on the following road types, if at all?
Motorways
'A' Roads
'B' Roads
'C' Roads
'Unclassified' roads
Urban
Rural
Please complete questions 7 and 8 if you have
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS
128
Fixed
Mobile
Manual
ATCs
ATCs
Counts
Do not count
any Automatic Traffic Counters (ATCs)
Please fill in the relevant number e.g. 30
7.a
How many ATCs do you have?
How many, if any, of your ATCs
classify by vehicle type?
7.b
Which of the following vehicle types are classified separately?
(mark as appropriate)
Cars
Buses/ Coaches
Light Vans (up to 3.5 tonnes gross vehicle weight)
Two Wheeled Motor Vehicles
All Heavy Goods Vehicles
Heavy Goods Vehicles - Articulated
Heavy Goods Vehicles - Rigid
Pedal cycles
Other types/ combinations? (please specify)
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS
129
Please fill in the relevant number e.g. 30
7.c
How many, if any, of your
ATCs are fixed counters
7.d
How many, if any, of your
ATCs are mobile counters
7.e
Please describe the scheme for your mobile counters e.g. One counter is used for one week per annum on
each of 25 sites
Please fill in the relevant number of days
7.f
If applicable, how many days per annum do
most of your fixed and/or
Fixed
Mobile
ATCs
ATCs
mobile counters count for?
8
In which of the following intervals do you normally store your
(mark as appropriate)
count data?
15 minutes
30 minutes
Hourly
Other (please specify)
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS
130
Plese complete questions 9,10 and 11 if you do any Manual Counts
9.a
How often do you count at each site?
9.b
Why do you do it this way?
10
In which of the following intervals do you normally store your
(mark as appropriate)
manual count data?
15 minutes
30 minutes
Hourly
Other (please specify)
11
Vehicle type: Do you count
Yes
All motor vehicles only?
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS
131
No
Disaggregated by individual vehicle types?
If individual vehicle types, which of the following do you count separately:
(mark as appropriate)
Cars
Buses/ Coaches
Light Vans (up to 3.5 tonnes gross vehicle weight)
Two Wheeled Motor Vehicles
All Heavy Goods Vehicles
Heavy Goods Vehicles - Articulated
Heavy Goods Vehicles - Rigid
Pedal cycles
Other types/ combinations? (please specify)
Automatic Traffic Counter and Manual Count Sites
12
Please describe how you select which road links to count
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS
132
13
Please describe what the distribution of sites is (e.g. 50 sites on A roads
30 sites on B roads, 20 sites on unclassified roads)
14
What do you define to be a road link?
15
Do you count traffic in both
directions (of a road link)
combined or in each direction
separately?
Both directions combined
Each direction separately
16
Do you geographically reference your Count Points by:
(mark as appropriate)
Ordnance Survey Grid reference
Name of Road
Road Number
Other (please specify
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS
133
Yes
17.a
18
No
Does the data you collect fully meet your requirements for traffic
information?
Please describe any planned major changes to the way in which you collect traffic data.
OTHER DATA/ INFORMATION MANAGEMENT
19
Yes
No
Yes
No
Do you have SCOOT ?
(Split Cycle Offset Optimisation Technique)
20
Do you use Automatic Number Plate Recognition technology?
21
Do you collect the following data (including those derived from
real time traffic information systems)
Incident data (road accidents etc)
Roadworks (planned and unplanned)
Planned large scale events (e.g. concerts)
Traffic management interventions (e.g. diversions due to
roadworks or an accident)
Inventory data (road gradient, condition, number of lanes etc.)
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS
134
Weather
STATISTICS CALCULATED FROM THE RAW DATA
Yes
22
No
Do you calculate aggregate statistics from the data collected?
If "Yes" please describe what aggregate statistics you calculate from the raw data?
e.g. Annual Average Daily Flows, Total Vehicle kms
Yes
23
No
Do you calculate separate aggregate statistics for Minor roads
(B, C and Unclassified roads)?
If "Yes" please describe the statistics calculated in more detail (if necessary)
than may have been described in question 22.
24
Yes
Do you have a system or process that provides for the aggregation and
analysis of different data sets, and if so, how developed is it? (The DfT and
HA are developing a Roads Information Framework that will aggregate journey
time, asset and external events data, enabling performance management
reporting and analysis at operational and strategic levels.)
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS
135
No
If there were any questions where you felt you hadn't been given the opportunity to respond fully, please feel free to do so now, making
clear refererence to the Question number.
Please return completed questionnaires as soon as possible, by 18 February 2005 at the latest. Electronic copies would be greatly preferred, email to:
roadtraff.qualityreview@dft.gsi.gov.uk, with "LA Questionnaire Completed" in the subject heading.
Alternatively, completed paper copies of the questionnaire should be returned to Martin Trevor, Department for Transport, Zone 2/14, Great
Minster House, 76 Marsham Street, London, SW1P 4DR.
Contact Louise Buckle for enquiries at:
E-mail:
Telephone:
roadtraff.qualityreview@dft.gsi.gov.uk
020 7944 6389
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS
136
Annex 6.2
Road Traffic Statistics Quality Review
Local Authorities Data Consultation
Introduction
43. This note reports on the results of a questionnaire sent out to Local Authorities in
England, Scotland and Wales, to establish what data they collect.
44. It is intended that the findings, covering a total of 64 responses received, will
prepare the groundwork for future discussions with Local Authorities on data
sharing, with a view to Local Authority data contributing to DfT estimates and DfT
data being accessible for Local Authorities uses.
Questionnaire Findings
3. 64 responses were received and these fall into the following groups:
English LAs
County
London Borough
Metropolitan County
Metropolitan District
Unitary Authority
Transport for London
Scottish LA
Welsh LA
16 responses
7 responses
2 responses
7 responses
19 responses
1 response
8 responses
4 responses
Of these, one of the LAs neither collects nor outsources the collection of traffic
data. Thus for the purposes of this analysis a total of 63 responses will be used.
As only one LA out of the 64 neither collects nor outsources the collection
of traffic data it is possible that the responses might not be representative
of LAs as a whole. Moreover it is possible that the response was biased to
those LAs with efficient and established traffic data collection or
outsourcing programs.
4. A map illustrating from which areas responses were received highlights that no
responses were received from the South West Government Office Region, but
they were sent the questionnaire and a reminder. There are relatively few LAs in
the South West. The detailed breakdown of all responses received are held in a
database for further analysis as required.
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 137
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 138
6. Of the 63 responses :
a. 32 (51%) of LAs use the data they collect exclusively for monitoring
their own traffic levels,
b. 7 (11%) use the data for DfT estimates and their own monitoring
purposes, i.e. all LAs who collect data for the DfT, use those same data
for their own monitoring purposes,
c. while11 (17%) LAs use the data for monitoring and other LA purposes
and,
d. 10 (16%) use the data for DfT road traffic estimates as well as
monitoring LA traffic and for other LA purposes.
e. 3 (5%) use the data they collect for their own ad hoc purposes only, eg
assessing the impact of a traffic scheme
These results show that of the 63 LAs who responded 97% (60 LAs) monitor their
own traffic levels of which 33% (21 LAs) use the data for other LA purposes. Of
the 60 LAs who use the data that they collect, 43 (70%) do so independently of
the DfT.
7. Of those who collect traffic data on an ad hoc basis the overwhelming use is for
scheme feasibility studies, scheme specific monitoring, requests and
investigations (eg into speeding or rat-running, location of pedestrian crossing).
8. There is a wide spread of methods used to count traffic
 92% (58) use Manual counts
 87% (55) use fixed ATCs
 89% (56) use mobile tube ATCs
 51% (32) use video counts
 67% (42) use pedal cycle counters
 21% (13) use other methods
Of the other methods, Radar (5), Surveys and/or interviews (4) and SCOOT
(3) were mentioned. One mention was made of pedestrian surveys.
9. Many of the LAs conduct Cordon and Link counts (73% and 79% respectively).
There were a variety of other methods mentioned of which Screenline, Turning
and Junction counts were frequently mentioned.
10. 62 of the LAs count on each of A roads, B roads, C roads and Unclassified roads,
55 count on Urban roads and 40 count on Rural roads in some manner. The most
common types of counts appear to be a combination of two or three of Fixed,
Rotating and One-off counts. One of the potential benefits to DfT would be data
collected on Minor roads and it appears as though there is a source of this data
collected by LAs. Few LAs count on Motorways (12) and the majority (9) of those
counts are conducted manually.
11. The manner in which the LAs count on these road types seems to be
concentrated on a combination of Fixed ATCs, Mobile ATCs and Manual counts.
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 139
12. Responses from the LAs on how many ATCs (fixed and mobile) they have, how
many days they are used for and how the data from them is stored were
comprehensive and form part of the database referred to above. Comprehensive
answers about how the mobile counters are used are also in the database.
13. Similar responses about the manual counts also form a substantial part of the
database.
14. The majority of the LAs distinguish between the different vehicle types although
question 11 which asked this directly may have been slightly misunderstood as
some LAs ticked both options in response to the question: "Vehicle type: Do you
count
All motor vehicles only?
Dissagregated by individual vehicle types?"
This is not necessarily an issue as it is clear to see that the great majority
collect disaggregated statistics. A breakdown follows:
ATCs








Cars
Buses/Coaches
LGVs
Motorcycles
Heavy Goods Vehicles Total
Heavy goods vehicles articulated
Heavy goods vehicles rigid
Pedal cycles
36
37
28
28
29
31
31
21
Manual counts
 Cars
 Buses/Coaches
 LGVs
 Motorcycles
 Heavy Goods Vehicles Total
 Heavy goods vehicles articulated
 Heavy goods vehicles rigid
 Pedal cycles
54
55
50
54
39
41
41
54
Although this clearly shows that not all LAs collect data on all the vehicle
types, it does indicate that there are substantial sources of data available on
all the vehicle types used in DfT's traffic estimates.
15. Discussions on technical issues such as definitions of road links and sampling
distribution of sites are recorded in the database.
16. Many of the LAs have SCOOT (A traffic management system) which may prove
to be a source of data for DfT. Some further investigation into the possibilities of
SCOOT will have to be carried out.
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 140
17. 58 of the LAs do not use ANPR technology. Of the 5 that do, some may prove a
source of information on the possibilities and challenges of this technology.
18. There is also a rich source of additional data such as: Incident data, Roadworks
(planned and unplanned), and Inventory data that may prove very useful in
validating ATC and Manual counts.
19. 76% of the LAs (48) calculate aggregate statistics from the data that they collect,
and 33% (21) of LAs calculate separate aggregate statistics for Minor roads.
20. There are 4 LAs who have "… a system or process that provides for the
aggregation and analysis of different data sets…" None of these LAs was Leeds
which is acting as a pilot for the RIF project. Although this is a small number,
these authorities may have useful advice to offer on how a system for sharing
data may be set up.
21. Although a lot of the data collected by the questionnaire is not immediately
useful, it is clear from these results that there is a great deal of data collection
and traffic monitoring going on, and that there is in fact data which could be used
in DfT estimates.
April 2005
SR2
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 141
Annex 6.3
CLIP-TS
Paper on potential for Local Authority data sharing
Prepared by John Marriott Leicestershire CC
Introduction
This paper is based largely on my own experience working for Leicestershire CC and a
personal interest in new technologies. A number of LA traffic data collection officers in the
midlands, including me, meet from time to time to share experiences through a group known
as LATDAG. I have taken some soundings from members of that group.
Over the last 2 to 3 years the County Council and Leicester City Council have spent several
hundred thousand pounds conducting surveys collecting traffic data. A significant proportion
of that has been used to create and validate traffic models for Leicester and the surrounding
area and other county towns.
It is evident that most local authorities spend very considerable time and resources collecting
traffic data. It seems likely that expenditure by LAs on traffic data will far exceed that spent
by central government on collecting traffic information for statistical purposes.
While much of the data could theoretically be used to supplement data collected for DfT
there are a number of quite significant issues that would need to be addressed before local
data could be used for this purpose. Some of these include: the wide variety of purposes for
which data is collected; the different types of technology employed; choice of sites; and data
format issues. Clearly, this would require some centralised co-ordination which would be
extremely difficult, if not impossible, to achieve through LA organisations so the lead would
inevitably have to be taken by DfT. From a LA perspective, the process could involve
considerable additional work and time and LAs may be very reluctant to join in any scheme
on a voluntary basis unless there are seen to be some real benefits to them and/or adequate
financial compensation.
Purpose
National or local
It appears that the main purpose of the Quality Review is to improve traffic data for National
Statistics purposes. Most local authorities appear to collect a reasonable amount of data
subject to the resources available to them, which is often dictated by local priorities. It is
evident that there is quite a disparity in the way LAs collect data and some of this can be
attributed to historic reasons and levels of funding. The LTP process has encouraged a
fundamental review of data collection in many authorities.
Local authority data can be targeted at specific schemes or it can take the form of more
general monitoring to establish general trends. In many cases a knowledge of absolute traffic
levels or growth is not necessarily regarded as that important because it is not major cause for
concern. Even where it is a cause for concern, there may be comparatively little that can be
done about it. Monitoring small year on year changes that may be very hard to detect with
any reliability may not be regarded as a high priority.
Motor vehicles flows and traffic
Nearly all authorities are likely to collect volumetric flow data as this is relatively easy to
collect and avoids problems of vehicle classification. The extent to which authorities collect
additional information on speeds and vehicle classes seems to vary considerably. Estimation
of traffic, in terms of veh.km, requires flows to be applied to network links and some
knowledge of the lengths of those links. The estimation of traffic on M, A and B roads is
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 142
relatively straightforward since there are only a limited number of links and junctions. It is
much more problematic to estimate traffic levels on lower category roads and it seems likely
that a number of authorities do not attempt to do this.
Cycles
Monitoring cycle flows appears to be problematic for all authorities. The current generation
of cycle counters all seem to have some technical problems to a greater or lesser degree in
terms of reliably identifying and counting cycles. This applies particularly to counters in
mixed traffic lanes. It is also difficult to find sites where all cycles will be intercepted and the
majority of cycle counters seem to be located in off-road sites. Their seems to be scope for
more research aimed at improving the accuracy and reliability of cycle counters in order to
try and detect what are often relatively small changes, usually in fairly low flows. This would
provide a better basis for understanding why such changes are (or are not) occurring.
Format of surveys
Automatic
Most authorities collect data automatically and many have moved to using loops rather than
tubes, particularly for permanent or semi-permanent sites. Different automatic counters offer
different options for classifying vehicles. Some attempt to determine classes by reference to
an electro-magnetic “profile”, axle spacing or length. There are many different ways of
classifying vehicles and local authorities tend to choose either a consistent methodology or
vary the classes according to particular requirements or the limitations of available
equipment. The number of automatic counters in each LA appears to vary considerably and it
seems likely that the extent to which the sample is representative of the road network will
also vary.
Some flow data is now being derived from traffic signal SCOOT loops and this could provide
a rich source of data, particularly in urban areas.
There seems to be very little published data on the accuracy of different types of counters in
terms of total traffic or accuracy of classification. Some types of detectors e.g. radar are
primarily designed to measure speeds but can provide some volumetric or classification
information that while not being 100% accurate can provide an indication of change if used
consistently or can be calibrated to provide a better estimate.
Manual.
Manual surveys offer great flexibility in the way data is collected but most are confined to
single days usually for a maximum of 12hours during the working weekday. It seems likely
that few LAs would choose to use this type of survey for long term monitoring purposes
because of the day to day variability. While it is possible to factor up such data using data
from automatic counters this may or may not provide good results. It relies heavily on the
skills and local knowledge of individuals and a number of other factors.
Vehicle Classification
A number of LAs find it useful to use a Medium Goods class which is not separately
identified in the DfT 12hr surveys. Some authorities are more interested in the physical size
of a vehicle, rather than the load it is carrying so they do not observe whether all axles are
actually lowered onto the road.
In theory, if a large number of classes are counted initially then the data could be aggregated
to provide a smaller number of more consistent classes. However, the variation between
authorities means that classes, even when aggregated, may not be entirely consistent. Moving
from volumetric to classified data will increase data volumes significantly and this has
implications for data storage. With some systems the time taken to collect and download the
data can become extended.
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 143
Coverage and analysis
Road network
For many authorities it could be a considerable challenge to devise a statistically robust
sampling methodology. Not all authorities are likely to need this sophistication and the
monitoring network may have been devised to monitor specific local issues.
The road network can be described in many ways but only M, A and B roads have a
nationally recognised numbering system. (Nevertheless, there are some A roads in different
parts of the country with the same number!) For un-numbered classified roads and unclassified roads each LA may have adopted its own referencing system. Such systems may
not embrace all roads and are unlikely to be included in any national data sources, including
those produced by the OS.
Computerised road networks are split at different points depending on what the particular
network is used for. The NSG is broken at District boundaries whereas for other purposes it is
convenient to split links at speed limits. The OS centreline network (OSCAR) has link and
node references but these appear to have change from issue to issue and so cannot be used as
the basis for a local referencing system. The OSCAR reference numbers are also long and
cumbersome to use. It is not possible to use the OSCAR based road network to calculate road
lengths because of the way the product was originally specified. When creating new networks
based on products such as OSCAR it would be desirable to achieve consistency regarding
numbering, the level of complexity used for junctions, the extent to which the public have
access as well as information on the maintaining authority.
The would appear to be a case for establishing a definitive national link and node traffic
network which could be used specifically for traffic statistical purposes, so that it could be
used to calculate road lengths, flows, accident rates and speed limits. Some authorities have
already devoted considerable time and effort to the creation of such networks and are unlikely
to want to change unless it offers some very considerable benefits. The specification for the
OS’s Integrated Transport Network (ITN) suggests that it will show many traffic related
features including signs, speed limits, traffic calming and restrictions. However, it would
seem that there are no mechanisms in place to ensure that such data meets LA users’
requirements in terms of accuracy or definitions. A more integrated approach involving data
sharing between the OS and mainstream users, such as Highway Authorities, would appear
beneficial to everyone but may require a lead from Central Government.
Analysis
A particular problem with automatic data arises when data is missing or appears erroneous. If
data is not collected regularly from automatic sites several months may go by and problems
will only become evident when data is next collected. One of the advantages of telemetry is
that errors might be noticed far more quickly and actions taken to rectify problems thereby
considerably reducing the amount of suspect data that needs to be patched or rejected. There
would appear to be a need for better software tools to clean or reject suspect data.
Smaller authorities may lack staff with the necessary statistical or software skills to carry out
robust analysis of data. This would not be a problem if the raw data was to be transferred to
others for subsequent analysis. Any transfer process would require some form of
standardisation of the data format. Software tools would need to in place to automate the
process to reduce the time taken and minimise the need to involve skilled staff.
Technology opportunities
ANPR
The technological advances in video and software processing would appear to make this an
increasingly attractive technology not just for counts but also for providing speed and
routeing information. There are, however, some concerns about Data Protection issues which
need to be clarified. The linking of registration numbers to the DVLA vehicle database
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 144
allows a far more detailed breakdown of vehicle types than is possible using other
methodologies. It could also be used to monitor the age of vehicles in particular areas, which
might be useful for estimating Air Quality emissions. With regard to foreign registered
vehicles it is understood that vehicles are already logged using ANPR at ports. It should be
possible to link such data to the national vehicle database, preferably including some
information on vehicle type. This would provide additional information on cross border
traffic and have advantages in terms of enforcement of traffic offences and security.
Image recognition
Military image recognition systems appear to be quite advanced but do not seem to have
found their way into mainstream traffic collection systems. Assuming that the price of
cameras and video storage and transmission systems continues to fall, their would seem to be
an opportunity to consider mass producing the necessary recognition systems to reduce costs
and open up new collection opportunities.
Radar
The emergence of moderately priced radar detectors that are claimed to be able to count and
classify (mainly by length) offers the potential to avoid some of the limitations of loop
counters. While these might not be as accurate as some other methods, no methodology
appears to be 100% accurate and in many cases it is the change rather than the absolute
number that may be more useful so good consistency is potentially just as useful as accuracy.
Electronic number plates
ENPs would appear to be another technology that would offer the ability to count and classify
vehicles, assuming this is still under consideration.
Road User / Congestion charging
There would appear to be the potential to build in traffic data collection systems into any
national charging scheme. This would seem to depend on whether such information was
retained within the vehicle “black box” or processed centrally. (It is noteworthy that various
recent articles on RUC, which appear to be based on DfT press releases, refer to vehicles
being “tracked” implying that data would be collected centrally. Surely it would be easier “to
sell” a system to the public if the tracking and fee calculations were carried out entirely
within the vehicle). However, the latter approach would appear to limit or rule out RUC for
many traffic data collection purposes.
DfT role on technology
LAs are probably spending quite a bit of time and resources on evaluating new technologies.
It would It would be useful for DfT to take a more pro-active lead in researching new
technologies and assisting LAs in the development of future data collection strategies.
Potential for LAs to share data with DfT
The main problems regarding data sharing appear to be : Much LA data is not collected for purposes that could be used for consistent national
monitoring although clearly there is quite a lot that could be used.

Multiple technologies which classify vehicles differently.

Unreliable counters and data patching / rejection issues.

Referencing and categorising the network and achieving a statistically robust sampling
methodology for it.

Limited technical and statistical skills in some authorities.

An apparent reluctance within LAs to share what they have learnt (good or bad).

Data stored in various formats in each authority.

Lack of standardised data transfer protocols.
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 145

Limited analysis software to do data aggregation.

Time and resources, particularly given all the other pressures being put on LAs by Central
Government.
Suggested way forward
 DfT to consider the needs of national and local monitoring and commission, collate and
disseminate research into the accuracy, reliability and costs (equipment and human
resources) for existing and new technologies.

Instigate a programme of trials and controlled monitoring of promising new technologies
to determine the most appropriate solutions for the medium and longer term.

It is suggested that central government should fund the development of any new
technologies taking into account the benefits to all levels of government.

Some LAs should be encouraged (and assisted) to work with DfT on piloting some
elements of this process.
John Marriott
June 05
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 146
Annex 6.4
LOCAL AUTHORITY DATA: POTENTIAL USE BY DfT
1.
The table below sets out the various types of traffic data that may be collected
or compiled by Local Authorities. For each type of data we suggest ways in
which DfT can use the information.
2.
The ideal situation is where the LA has a full year’s ATC data at a major road
site, classified by vehicle type, so that a precise AADF can be obtained and
there is no need for DfT to mount its own survey. This can also apply when
the LA has estimated the AADF from shorter-period counts.
3.
In cases where the data are not fully compatible with DfT vehicle-types, it is
suggested that DfT estimates the AADF using a combination of its own data
and the LA data. In such cases the AADF will be a weighted average of the
two sources. The weights to be applied would ideally be the inverses of the
variances of the two estimates, but if the LA source does not have a standard
error calculated, then a subjective estimate will have to be substituted.
4.
It does not matter if the major road sites surveyed by the LAs are not a
random selection, because DfT estimates AADFs for every major road link
each year. Those links that are surveyed by the LAs will potentially have
tighter confidence limits than those surveyed only by DfT, because they can
be based upon two sets of data. The fact that the LA sites might have higher
than average flows, or higher than average growth-rates, would not induce
any bias in the overall estimates of vehicle-kilometres.
5.
It seems unlikely that DfT will be interested in LA traffic data for minor roads,
unless there is an attempt to estimate AADFs for a representative sample of
links on the minor road network; in most cases the LA would probably have
selected the minor road sites for local traffic management purposes, rather
than as part of a statistical sample.
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 147
Type of data
AADFs
Potential use by DfT
For Major roads, derived If AADFs are classified
from continuous ATCs
using DfT vehicle-types,
over full year
no need for DfT to
mount its own one-day
count.
AADFs
For Major roads,
established from ATCs
for periods less than
one year, or from
manual counts
ATCs for periods shorter For Major roads
than one year.
Manual count data, in
neutral months
For Major roads
If AADFs are not
classified or if they use
non-standard vehicletypes, DfT can use the
AADFs in conjunction
with its own one-day
counts to derive AADF
by vehicle-type.
DfT to check
methodology; if
compatible with DfT
approach, can proceed
as above.
DfT can estimate
AADFs based upon
‘neutral month’ data
from the ATCs, factored
by usual expansion
factors. If the ATC data
are not classified, still
need to undertake the
one-day manual count.
If classified using DfT
vehicle-types, can
estimate AADFs.
If not classified or uses
non-standard vehicletypes, can use in
conjunction with DfT
one-day counts to
derive AADF.
Other data
Traffic counts for minor
roads
Cordon and screenline
counts
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 148
Probably of little interest
to DfT, unless part of a
random sample.
Can use the counts for
individual major road
sites on the cordons and
screenlines
Vehicle-kilometre
Major or minor roads
estimates for local areas
Counts of cycle traffic at
a collection of sites
Major or minor roads
MD
28 March 2005
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 149
If sites selected
randomly by road-class,
can use to improve
estimates of average
flow in LAs.
Need to establish basis
of monitoring. If aim is
year-on-year trends, can
use along with similar
data from other LAs to
compare trends at LA
cycle-monitoring sites
with national trends
based upon DfT panel.
Annex 6.5
POTENTIAL USE OF EXISTING DATA SOURCES
1.
Introduction
1.1
This annex considers how some existing, or potentially, data sources might
be used to improve estimates of annual traffic.
Potential use of Highways Agency traffic data
2.1
2.2
2.3
What HA produce
HA currently produce volumetric counts [ATCs] at some 1500 motorway and
trunk road sections, generally for 365 days * 24 hours. HA and its contractors
validate and patch the data, though on occasions the data can be declared
‘bad’ and are not used in the subsequent analysis.
Traffic is not classified by vehicle-type [VT] but usually there is some
classification by vehicle-length [VL]. The number of classes varies with the
contractor but typically there are 4 or 2 VL-classes, and in some cases no
split at all.
Data are held in the Highways Agency’s Traffic Flow Database System
(TRADS). The usual site-specific output is the AADF plus the percentage of
Heavy Goods Vehicles. The latter is derived from the nearest DfT manual
counts.
Potential use by DfT
2.4
Instead of using 12 hour manual classified counts [MCCs] and factoring these
to produce AADFs, DfT might take the AADFs directly from TRADS. This
would be possible at a very large number of sites, though it would not directly
provide the VT-split.
2.5
The TRADS data might provide a more representative set of sites from which
to derive counting-day-to-year expansion factors. Currently the set of ATC
sites is biased towards uncongested sites. With more soundly based factors,
from a more representative sample, DfT could produce better AADFs for the
sites where MCCs had to continue.
Preliminary investigations required
2.6
Before HA data could be exploited a number of preliminary trials would be
required. First there would be check counts to test the reliability of traffic
counts. Of particular interest are whether there are under-counts due to lanechanging, whether the counts fail when the traffic is very slow-moving, and
whether the VL classification is reliable. HA may be able to obtain results of
video comparisons from one Design, Build, Finance, Operate (DBFO)
scheme, but other checks would be needed over a range of sites.
2.7
Provided that the basic HA counts were found to be reliable, crosscomparison counts would be taken to establish relationships between VL and
VT, ie the probability that a vehicle recorded in VL category j belongs to VT
category i. These relationships might vary by hour of day and by type of site;
and of course by the different types of VL-classification. The aim would be to
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 150
establish robust ‘transition probabilities’ that could be used with confidence at
a large number of sites. [Obviously there would be no point in establishing
separate ‘transition probabilities’ for each site, because this would be more
onerous than the current MCCs].
2.8
An alternative approach to deriving VT breakdowns would be to ‘twin’ an HA
site with a nearby DfT site and use the latter’s VT-classification. But such
twinning would need to be justified on a case by case basis, with rigorous
cross-checks on its reliability. Perhaps a more promising approach would be
to exploit the many ANPR camera sites that are springing up on motorways
and trunk roads, but this would depend inter alia on the data owners’
willingness to share these data.
2.8
No doubt the results of these trials would depend on particular site conditions.
The objectives would be to build upon the positive findings, ie the cases
where the counts were reliable and where robust transition probabilities could
be established.
How DfT could derive traffic estimates for motorway and trunk roads in
England
For sites that were covered directly by TRADS, DfT would take the AADFs from
TRADS and apply the ‘transition probabilities’ to derive the split by VT.
However, if there was a need to apply an adjustment for undercounting in
certain hours, or if the ‘transition probabilities’ varied by time-period, then
there would be a need to apply these adjustments to the data for individual
days.
For those [comparatively few] sites not covered by TRADS, DfT would continue to do
12-hour MCCs, although the possibility of longer counts could be considered
for the busier sites. DfT would use data from a range of HA sites, and from its
own ATC sites, to establish expansion factors and growth factors; to derive
these factors by VT, adjustments would be applied to the HA data. The
factors would then be applied to the MCCs to produce AADFs.
Thus AADFs would be obtained for all sites on the motorway and trunk road
system, based upon data for the latest year; there would be no need to
update figures from earlier years using growth factors.
Furthermore the AADFs would be more reliable than those currently produced, being
based upon many times more data and on sites that covered the full range of
conditions. However, whether this is a practicable proposition depends
crucially on [a] whether the HA volumetric counts are sufficiently accurate and
[b] whether robust ‘transition probabilities’ can be established to determine the
split by VT.
More ambitiously, it might be possible to envisage using the HA ATC data to extend
the core census monitoring, in particular to sites which the current core
census is unable to cover. Whether this was feasible would again depend
upon the trials. One difficulty could be in obtaining the vehicle-type ‘transition
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 151
probabilities’, because these would be likely to vary over the year. ANPR
data might provide the best means of obtaining these data.
3.
Potential use of mileage data from annual MOT tests
3.1
There is a likelihood that the mileometer readings on MOT test forms will be
recorded centrally at DVLA for all non-goods vehicles. It should therefore be
possible to deduce the mileage undertaken over the previous 12 months, and
hence the overall mileage by these vehicles, split by vehicle-type. This could
be a useful cross-check on the DfT vehicle-kilometre figures, although there
would be no breakdown by road-type, month, time of day or region [though
area of licensing would be available].
The main problem in using centrally-held MOT mileometer readings is that
newer vehicles under 3 years old are excluded from MOTs, so that the total
mileage derived from this source would significantly undercount the overall
total. However, the mileages from these cars would be available when they
had an MOT test, and there might be opportunities for some
modelling/assumptions about mileages in the initial 3 years.
There would also be some concerns about the reliability of the mileometer
readings, given the possibility of tampering. We could not be certain that all
vehicles in use had MOTs conducted regularly every 12 months; also when
vehicles are scrapped, this final period’s mileage would not be covered by
MOTs.
In conclusion, therefore, it seems unlikely that any aggregate mileage
statistics derived from MOT records could provide a substitute for the DfT
traffic statistics. It would nevertheless be useful to compare the two sets of
figures: the percentage shortfall between the DfT figures and the MOT-based
figures might indicate the proportion of overall mileage covered by newer and
very old vehicles.
3.2
3.3
3.4
4.
Potential use of satellite images
4.1
At some stage satellite imagery may be able to present a snapshot
‘photograph’ of traffic on all roads in GB. Given that this would cover all
roads, could it provide a more efficient and effective substitute for point-based
traffic counts?
4.2
Let us assume that such images could readily be converted into numbers of
vehicles on each road link [or in the case of minor roads, perhaps the
numbers of vehicles on the totality of these roads], and that this process can
weed out parked vehicles [as opposed to vehicles sitting in stationary traffic
queues]. We could then derive traffic densities, ie vehicles-per-kilometre. If
we knew the average vehicle speeds on each road, we could then derive the
vehicle flows from the formula
Vehicles–per-hour = Vehicles-per-km * Km-per-hour.
4.3
Using assumed speeds for different classes of road, one could therefore
deduce approximate estimates of hourly flows during the period when the
images were obtained. Given that the relevant data were available in the right
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 152
form, this would provide an effective geographical profile of traffic flow levels
across the regions and across the different road-classes.
4.4
5.
However, it seems unlikely that this process could be extended to all hours of
the day, given the costs involved and the ‘data-explosion’ involved in
processing the information. Also the need to make assumptions about
average speeds may limit the accuracy of any derived traffic flows, though
knowledge about the distribution of speeds over the network is being
extended as new survey techniques become available.
Summary of Recommendations
5.1
The main recommendations on the use of Highways Agency data are:
5.2

That trials be undertaken of HA data, to test their reliability for DfT
monitoring;

If the trials are successful, some cost-effective means of disaggregating
HA data by vehicle-type must be found. These might involve special
manual counts, ‘twinning’ with nearby DfT sites, or use of existing ANPR
data, if the latter could be made available;

DfT should aim to maximise its use of HA data, subject to the above trials.
The aim would be obtain comprehensive annual data on AADFs on
motorways and trunk roads. The more ambitious aim would be to use HA
data to extend the coverage of continuous counting, so that the Core
Census extended its reach to the more ‘difficult’ sites.
As regards other data sources:

The main use of annual mileages derived from comparisons of MOT test
would be to compare annual mileages with the RTS figures. The MOT
figures would relate mainly to older vehicles;

Satellite imagery might provide comprehensive information at some stage,
but there are a number of practical requirements to be satisfied first.
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 153
Martin Dale
Annex 7.1
Road Traffic Statistics Quality Review
Exploring New Technologies
Introduction
1. The road traffic estimates are currently derived from two sources of data:
manual counts and automatic counters. Whilst these are tried and tested data
collection instruments, recent years have witnessed a number of new and
emerging technologies that might also be used in the estimation process.
Background
2. The DfT traffic estimates are currently calculated using data from some 190
automatic traffic counters outside London and 56 counters in London, together
with manual traffic counts. The automatic counters outside London are the
Piezo sensor and inductive loop type which can classify by 21 vehicle types to a
good level of accuracy. The London counters can only distinguish between
“short” and “long” vehicles, therefore supplementary data from manual counts is
required for more detailed breakdowns.
3. In addition to counting the number of vehicles, the automatic counters are used
to provide data about vehicle speeds, published by the Department. It costs
around £0.9m pa to maintain and run the automatic counters compared with
some £2.6m pa to undertake the manual traffic counts.
4. The attached note describes the main current and new technologies. It
concludes that each technology has a number of advantages and
disadvantages with none best for all situations.
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 154
5. A number of factors need to be taken account of in considering which
technologies to employ, including:
(a) User requirements for traffic statistics, including:
 flows at different geographic levels, eg individual links, and
for individual time periods for congestion monitoring
purposes
 breakdown by vehicle type
 breakdown by road class
(b) Methodological and sampling issues, ie appropriate combination of
data from automatic counters, manual counts and other sources
(c) Data from other providers, including Local Authorities and the
Highways Agency
(d) Consistency with data used for traffic management purposes
(e) Costs: purchase, installation, and subsequent maintenance
(f) The level of accuracy required
(g) Reliability
(h) The practicalities of installation and maintenance
(i) The storage and processing of data collected
(j) Timescales for introduction
6. Some work has already been undertaken to consider the potential of new
technologies in the context of road freight activity. Road freight data are
currently collected through the Continuing Survey of Road Goods Transport and
related surveys. The DfT surveys request hauliers to record details of their
journeys over a week or shorter time period, including the origin, destination
and distance travelled. Unlike the road traffic estimates, data are not required
about the route, though this would be useful for some purposes. Annex B
summaries a consultancy project undertaken for the Department on the
potential use of Intelligent Transport Systems (ITS) to provide road freight data.
The study concluded that the current ability of ITS applications to deliver useful
data for statistical purposes is extremely limited.
Conclusions
7. Clearly, the new and emerging technologies described offer a number of
opportunities to enhance the availability of data used to produce the traffic
estimates. In order for such technologies to be exploited effectively, decisions
will need to be made about how many to employ and where to locate them.
This, in turn, depends to a large extent on the review’s conclusions relating to
methodological and sampling strategy issues.
Further work will be needed to assess the new technologies, perhaps against the
criteria listed at paragraph 6, together with any others identified. Timing will need to
be carefully considered. Whilst some technologies might be introduced in the short
or medium term, others will have a longer timeframe.
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 155
ANNEX A
AUTOMATIC TRAFFIC DATA COLLECTION
CURRENT TECHNOLOGY AND FUTURE DEVELOPMENTS
Synopsis
1. This paper is submitted to the review board as part of the Review into the Quality
of National Road Traffic Statistics. This paper aims to provide a relatively brief view
into the technologies available and recent developments in the field of automatic
traffic counting.
2. It is generally assumed that inductive loop data counting is the most widely used
of the currently available technology and is considered to be highly accurate, only
beginning to drop off in performance when traffic speeds drop below 5kph.
3. As a result, most evaluations of advanced or newer non-intrusive detectors
compare with inductive loops because loops are a proven, successful and mature
technology, and when properly installed, serve as an adequate benchmark for test
purposes. The main downfall in the use of inductive loops is the expense of traffic
management and the danger in exposing installation crews to traffic.
Available Technology
Inductive Loop
4. As discussed above, this is the most widely used and generally accepted
method of automatic traffic data collection. The pure loop only system is
capable of collecting volumetric and speed data, and depending on the
sophistication of the data recorder connected to the loops, can drop vehicles
into a number of different categories based on the length of the vehicle. The
vast majority of these length based bins work on the ‘short’ and ‘long’ vehicle
types, i.e. +/- 5.2m in length as used in the London core census.
5. Recent developments in this field have focused on the need for loop
profiling from the counters attached to the loops. This is a method of
ascertaining the ‘magnetic shape’ of the vehicle from underneath as it passes
over the loop. So far trials are proving to be reasonably accurate, although
there is still a need for a higher degree of refinement before it can be said to
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 156
good enough for accurate classification purposes. This could certainly be
looked upon in the future as a relatively ‘cheap’ method of getting automatic
classified traffic counts.
6. Loops are highly reliable in terms of maintenance and operational
performance. Most failures originate in the loop wire, but the wire itself is not
necessarily the initiating cause of failure. Generally the failure arises because
of improper sealing, pavement deformation or foreign material in the slot. An
example of the reliability of such equipment is that some of the inductive loops
used on the current DfT automatic traffic count system described below, have
been in place, untouched since 1986.
Piezo sensor and inductive loop
7. This system is the one currently being used by the Department in its
collection of data with automatic traffic data collection system. This system
has all the benefits of the inductive loop system, described above, plus the
added benefit of two piezo axle sensors installed in combination with the loop.
This allows an accurate count of the number of axles and a higher degree of
accuracy in calculation of vehicle speed. Depending on the type of classifier
attached to the array, through a complex set of algorithms it can provide a
high degree of vehicle classification to a good level of accuracy.
8. Again, as mentioned above, it is hoped that with the new developments
with loop profiling it will be possible to include profiling into the algorithms to
improve classification further. The aim will be to iron out the current ‘grey’
areas of classification such as the separation of the shorter ‘city hopper’ type
buses from the rigid 2 axle lorries and the larger transit type vans from the
rigid 2 axle lorries.
Pneumatic tube counting
9. This is probably the cheapest and simplest method of counting on an
automatic basis. The system relies on either one or two hollow rubber tubes
placed across the road, relying on axle impacts from the vehicles as they pass
over. This system is only as intelligent as the counter attached to them, so
they can either be reasonably reliable as a straight volumetric traffic counter
or more a sophisticated method employed from a counter such as the
Australian manufactured Metrocount system. This relies on a complex
algorithm written into the manufacturers software that allows a reasonable
degree of classification for the user.
10. Due to the relatively crude method of deploying the tubes, these counters
can only be used for temporary traffic surveys of periods up to a month or so.
Their ideal target locations would be for relatively low trafficked roads such as
country lanes. The counters can be deployed for a designated period to
provide counts over a 24-hour period in situations where a similar manual
count would either be impractical or prohibitively expensive.
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 157
Passive infrared sensor
11. This type of sensor can usually only monitor one lane at a time and can
be either mounted overhead or slightly to the side of the road. It must also
face oncoming traffic. The alignment needs can cause problems in obtaining
optimum performance, so installations should normally prefer overhead
mounting.
12. Tests have found that it is apparently simple, straightforward, small and
easy to mount. The detection accuracy was better during free-flow conditions,
but it undercounted by 10 per cent during heavy traffic. The device also
consistently underestimated speed by 10 per cent on average.
Passive infrared and pulse ultrasonic
13. Again, a single lane detector that incorporates the two technologies of infrared and
ultrasonics. This can be deployed either overhead or from the side, and is designed to
detect vehicles at a short distance (no more than 13 metres). Ideally it should be installed
at no higher than 6 metres above the lane and 6 metres to the side. Tests undertaken have
shown that it undercounts by approximately 9 per cent from overhead, but only 1 per cent
from the side. However, it did demonstrate unstable performance during parts of the
sidefire tests. Unfortunately, speed comparisons are unavailable for this method.
Passive infrared, pulse ultrasonic and Doppler radar
14. As can be assumed by the title, this employs a combination of three technologies.
This type of counter must be mounted overhead as sidefire is not an option. Tests have
shown that it was easy to set up, taking only 30 minutes. Count results from these tests
were good, showing an absolute percent difference between sensor and baseline of 2.8% at
a height of 7 metres and 4.9% at a height of 5 metres. For speed accuracy, its absolute
average percent difference between sensor and loops was 4.4% at 7 metres and 3% at 5
metres. It appears that height of the mounting is therefore crucial, with a higher accuracy
in speed at lower heights but greater accuracy in counting when placed higher.
Video imaging
15. Cameras can be mounted either overhead or from the side. Tests have shown when
mounted at 9 metres over the centre of the lanes that performance was good. The absolute
performance difference between the sensor data and the loops were under 5% for all three
lanes. The detector also performed well for speed detection. The absolute average percent
difference was 7% in lane 1, 3.1% in lane 2 and 2.5% in lane 3. When checked at
mountings from other locations, either overhead or from the side, the equipment
performed best when mounted high and closest to the road.
Active infrared sensor
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 158
16. This is a system that can monitor only one lane at a time and must be mounted at a
height between 6 and 7 metres. Volumetric tests indicated excellent agreement with the
baseline loop system. The absolute percentage difference averaged 0.7%, which falls
within the normal accuracy tolerance of inductive loops. On a 24-hour test, the percentage
difference of average speed between sensor and loops was 5.8%. Speed was consistently
overestimated.
Microwave detector
17. Tests carried out on behalf the Federal Highway Administration in the USA indicated
that it did not function properly. It is a relatively new form of detection and needs further
development.
Passive acoustic detector
18. This type of detector monitors vehicular noise, (primarily tyre noise) as vehicles pass
the detection area. The detector can monitor as many as 5 lanes and it must be mounted
from a sidefire position. Precise alignment is not critical because the sensor can cover a
wide area. Tests carried out found this type of equipment provided good results under
free-flow conditions but undercounted during congested flow with slow speeds. During
free-flow the difference between from the baseline count provided by loops ranged
between 0 and 5% but during congested periods this figure rose to between 10 and 50%.
Speed accuracy was reasonable during free-flow conditions where they were usually
within 5mph of the baseline but this rose to 20-25mph during some peak periods. In
summary, it performs well in free-flowing traffic but its slow-speed accuracy need to be
addressed.
Non-invasive microloop probe (magnetometer)
19. Microloop probes measure the distortion in the earths magnetic field caused by the
passage of a vehicle and then convert the change to an inductance change. These can then
be used to monitor traffic from a 3 inch non-metallic conduit 18 to 34 inches below the
road surface or from underneath a bridge structure. Installers must use a magnetometer
underneath bridges to determine proper placement of the probes, otherwise it becomes a
case of trial and error. One of the requirements of this system is that the probes remain
relatively vertical so keeping the horizontal bores straight is critical to ensure accurate
detection. In tests when compared with the baseline loop system, the difference in volume
was under 2.5%, which is within the accuracy capability of most loop systems. For
speeds, the absolute average speed difference varied between 1.4 to 4.8%.
Satellite tracking (GPS)
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 159
20. GPS was originally developed for defence purposes. Satellites orbiting
the earth at 12,500 miles emit signals to track military ships, aircraft and
ground vehicles. The 24 satellites monitor location, direction and speed. The
navigation advantages of GPS have now expanded into the field of transport
field technology. To work satisfactorily, every vehicle would need to be
equipped with a GPS receiver to pick up signals from the satellites. The
positional information determined from the GPS signals is then transmitted to
a control centre to display the real-time position of the vehicle. It is likely that
the sheer quantity of data and current economical level of computing facilities
would make this prohibitive at this stage. However, this could well be a viable
option some years down the line.
Conclusions
21. Each detector technology and particular device has its own limitations,
specialisations and individual capabilities. No single device is best for all
applications. The successful application of detector technologies largely
depends on proper device selection. Within the same technology, the
performance and capabilities of each device can be quite different. A number
of factors impact upon detector selection, including data type, data accuracy,
ease of installation and calibration, cost, reliability, maintenance,
communication, power and installation site. Therefore, selecting an
appropriate device should be viewed upon as more important than choosing a
specific technology.
22. A summary table is shown below, for ease of reference, to compare the
advantages and disadvantages of the available detector technologies.
Technology
Inductive loop
and piezo
sensor
Advantages
 Flexible design to satisfy large
variety of applications
 Mature well understood technology
 The equipment cost is lower when
compared to non-intrusive detector
technologies
 Provides basic traffic parameters
(e.g., volume, presence,
occupancy, speed, headway, gap,
length and axle count)
 High frequency models provide
classification data
 Operability in harsh environment
 Quality of data high
Pneumatic tube 
counting
Quick installation for temporary
data recording
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 160
Disadvantages
 Disruption of traffic for
installation and repair.
Installation and
maintenance require lane
closure.
 Failure associated with
installation in poor road
surfaces
 Resurfacing of roads and
utility repair can also
create the need to reinstall
 Subject to stresses of
traffic and temperature
 Can decrease pavement
life
 Routine maintenance
requirement
 Inaccurate axle counting
when traffic volume is high
Infrared



Low power usage
Low cost
Simple to maintain


Active sensor transmits multiple
beams for accurate measurement
of vehicle position, speed and
class
Multi-zone passive sensors
measure speed
Multiple lane operation available


Multiple lane operation available
Easy installation

Ultrasonic

leading to poor data
quality
Cut tubes resulting from
vandalism and wear
produced by vehicles
Operation of active sensor
may be affected by fog or
blowing snow
Passive sensor may have
reduced sensitivity to
vehicles in its field of view
in rain and fog.
Data quality low
Some environmental
conditions such as
temperature change and
extreme air turbulence can
affect performance.
Large pulse repetition may
degrade occupancy
measurements when
vehicles travelling at
moderate to high speeds
Data quality poor
Inclement weather,
shadows, vehicle
projection in adjacent
lanes, occlusion, day to
night transition, water and
other detritus on lens can
affect performance
Requires a certain
mounting height for
optimum performance
Strong winds can cause
camera shake
Cost effective only I many
detection zones are
required in the field of
view
Antenna beam width and
transmitted waveform
must be suitable for
application
Doppler sensors cannot
detect stationary vehicles
Poor quality of data

Cold temperatures may






Video








Monitors multiple lanes
Easy to add and modify detection
zones
Rich array of data available
Provides wide-area detection when
gathered at one camera
Locations can be linked

Quality of data good


Microwave



Insensitive to inclement weather
Direct measurement of speed
Multiple lane operation available


Passive

Passive detection
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 161
acoustic


Insensitive to precipitation
Multiple lane operation available


Magnetometer





GPS



Can be used where loops are not
feasible (e.g. bridges)
Less susceptible than loops to
stresses of traffic
Some models transmit data over
wireless link
Less disruption to traffic flow than
inductive loop
Data quality within loop accuracy
capability
Will pick up every vehicles position
Clean and non-invasive technology
Data quality high







MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 162
affect data accuracy
Some models are not
recommended with slow
moving vehicles in
congested areas
Data quality low in
congested areas
Installation requires road
cut
Installation and
maintenance require lane
closure
Some models have small
detection zones
Induction magnetic
detectors cannot detect
stationary vehicles
Sheer quantity of data
could prove unwieldy
Cost and processing
required to deal with
volume would prove
prohibitive at present
Potential "big brother"
political issues
Annex B
The use of Intelligent Transport Systems to provide data on road freight
activity
1. DfT currently undertakes a number of continuous surveys in order to collect data
on both the domestic and international road freight activity of UK registered heavy
goods vehicles. The main data collected relate to distance travelled, weight and type
of goods carried, and origin and destination of journey. These surveys are carried
out by sending questionnaires to road haulage companies and to companies
operating in other sectors who own their own vehicles.
2. The surveys are fairly resource-intensive, both in terms of the compliance burden
on respondents and their administration by DfT. Therefore, DfT contracted TRL Ltd
to investigate the scope of Intelligent Transport Systems (ITS), particularly vehicle
tracking systems, for collecting data from such sources rather than the current
paper-based methods. The study, in 2004, involved a combination of a detailed
desk-based review of the telematics industry and a data gathering exercise, based
on both the use of questionnaires and on face to face interviews with senior staff in
telematics system providers and logistics companies.
3. The main conclusion of the study was that the current ability of ITS
applications to deliver useful data for statistical purposes is extremely limited.
A number of relevant specific key findings were identified.
4. In particular, the UK telematics industry covers only a small proportion of the UK
HGV fleet *. The stock of vehicles in which ITS systems have been installed is
unlikely to be representative of the stock of all HGVs in the United Kingdom.
Therefore any data collection methodology based upon these vehicles only would
not produce representative results. An option might be to collect the ITS-based data
from this subset of the HGV stock, and to continue paper-based methods for other
HGVs. However, organising such a dual approach and then reliably combining data
obtained from different ITS systems with other data could present considerable
difficulties.
* This low take-up is reflected in a survey of domestic hauliers
undertaken for the Road Freight Statistics Quality Review in 2002. Of
the 396 respondents, 21 per cent said that they currently used either
ITS or telematics and 6 per cent planned to do so in the future. A
further 25 per cent said they would consider its possible use but 48 per
cent had no plans to do so.
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 163
5. The data currently available will allow only partial analysis of road freight journeys
due to the limited nature of the data collected. For example, a number of the systems
investigated could not provide information about the weight or type of goods carried.
These are key variables in DfT road freight surveys. Some respondents considered
this information of low importance, suggesting that it was unlikely that in future much
effort would be devoted to collecting it. For Road Traffic Estimates purposes it is
important to note that all systems investigated provide the origin and destination,
though in some cases this is just the place name. However, not all systems routinely
record the route, road type, and distance travelled.
6. The study found it difficult to predict future long term developments in the industry
because of a number of unknown variables, including technical and financial issues.
7. In view of the above broad conclusions, DfT has no immediate plans to change
the way in which it carries out its road freight surveys, although it will continue to
monitor ITS developments in order to identify future opportunities.
SLAM4
January 2005
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 164
Annex 8.1
INDICATORS OF QUALITY
1.
Introduction
1.1
This note suggests a number of indicators of traffic data quality, together with
indicators of the quality of the resultant traffic estimates. The aim has not
been specifically to define performance indicators that might be used to
compare progress against some absolute standard or from one year to
another. Rather the aim is to summarise the results of quality checks, to
indicate that these checks have been done and by implication that the
appropriate remedial actions have been taken.
1.2
For convenience the suggested indicators are divided into three groups:

Data quality

Coherence of data-sources

Effectiveness of estimates and statistical reliability.
2.
Data Quality
2.1
The ATC data undergo two main sets of checks:

First, as the data are summarised as hourly totals each day, these totals are
compared with the ‘expected’ values based upon the previous five weeks’
data for the same days and hours. Where the figures lie outside the
acceptable range [which varies according to the vehicle-type], they are
patched automatically. Each time this occurs a flag is set. This automatic
patching can occur when a site [or lane] is not producing data at all, and this
situation could last a number of months before engineers are able to access
the loops or axle-sensors; when eventually ‘real’ data are obtained, the
patched data may or may not prove acceptable, and might then be revised;

Secondly, when the ATC data are used to derive expansion factors, further
patching may occur if the resultant factors appear to be implausible. At this
stage the data are graded in terms of plausibility; only if the data appear to be
at the extremes of implausibility are they patched.
2.2
Appropriate indicators for the ATC data could be

the proportion of hourly data figures to be patched during the year, for each
vehicle-type. This shows where the data-collection system has been deemed
to have failed;
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 165

the proportion of hourly figures to breach the ‘warning’ levels. These cases
are not ‘failures’ but may indicate a general instability in the data.
2.3
Both indicators would of course be monitored on a site-by-site basis, as a
means of diagnosing problems.
2.4
The Manual Counts are also subjected to validity checks, based upon figures
obtained in earlier counts, sometimes several years before. Given the
comparatively thin track-record for these counts, the confidence ranges are
comparatively wide. When the data are deemed to be extremely implausible,
this may indicate that the wrong site has been surveyed, and usually a recount is specified.
2.5
Appropriate indicators could be similar to those for the ATCs:

The proportion of manual counts requiring a re-count. This indicates an
‘inefficiency’ in the system, in the sense that it required a second count.
However, if the re-count tends to confirm the original count, then the AADF
can be based upon both sets of counts and will have a better standard error;

The proportion of sites where various levels of warning were triggered. This
does not indicate erroneous data, but a general lack of stability. In a year with
a high proportion of warnings we might expect some lack of coherence
between the trends indicated by the Core ATC sites and the more general
trends from the one-day counts. We explore this further below.
3
Coherence of data-sources
3.1
The Core ATC data and the one-day counts are used together to estimate
AADFs at sites on the network and overall vehicle-kilometres. The Core data
provide direct evidence of day-to-day variation and of year-on-year changes,
whilst the large set of one-day counts provide evidence of the absolute levels
of traffic flow. However, these two functions are not entirely distinct: for
example the average flow at the one-day count sites will provide some
indication of the change from the previous year. It seems reasonable to
provide indicators measuring the extent to which the two sets of data cohere;
the greater the consistency, the more likely we are able to present a
consistent and reliable picture of traffic levels and trends.
3.2
We have recommended elsewhere that there be annual checks on the
coherence of the ATC and one-day count data. This can be done by
computing, for each one-day count site, the ‘expected’ AADF, based upon the
previous year’s AADF, trended forward by the ATC data.
The overall
expected vehicle-kilometres are of course an example of this, where
provisional annual estimates are made before the full set of manual counts
become available.
3.3
Occasionally the majority of one-day count AADFs may be found to be
significantly higher [or lower] than expected, and this might lead us to revise
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 166
some of the vehicle-kilometre estimates for earlier years. The same could
occur when we undertake occasional benchmark surveys for minor roads.
Such revisions of past estimates are clearly undesirable, but sometimes
inevitable. We may also discover mistakes that were made in past years [eg
that the wrong road was counted in 2001] which will cause us to revise our
figures.
3.4
Suitable indicators could therefore be

The proportion of AADFs in each year found to be outside the confidence
limits for their expected values. These proportions could be worked out for
each vehicle-type. The calculations would be based upon the confidence
limits for the difference between the latest-year AADF and the expected
AADF, bearing in mind that each quantity has confidence limits. If the latter
are 95% confidence limits and if the data are consistent, then we would
expect no more than 5% of the differences to be significant;

The percentage difference between the provisional and final annual vehiclekilometre estimates for each vehicle-type. Some differences are of course
expected, so that this can be regarded as an indicator of how useful the
provisional estimates were;

The percentage revisions that have been made to ‘final’ data for past years. If
the revisions affect only some vehicle-types, or some road-classes, this
should be indicated.
4.
Effectiveness of the estimates
4.1
In statistical terms there are two key requirements. The first requirement is for
accuracy, ie that the traffic estimates be unbiased, ie not tending consistently
to over- or under-estimate the vehicle-kilometres for each vehicle-type or
road-class. The second requirement is for precision, ie that the confidence
limits surrounding those estimates be as tight as practicable.
4.2
To address accuracy one needs independent figures against which the survey
estimates can be checked. In practice such independent sources may also
be subject to bias. Examples for the traffic data are the National Travel
Survey [NTS], the Continuing Survey of Road Goods Transport and similar
surveys of vans and foreign goods vehicles, Highways Agency ATC data, and
Local Authority counts. Given the uncertainties involved, we cannot always
draw definitive conclusions from simple cross-comparisons; for example, we
may suspect that NTS and the CSRGT respectively miss some pedal cycle
and goods vehicle journeys in their travel diaries, or that the HA data miss
vehicles that straddle lanes. But there are good reasons for making these
cross-comparisons every year, to expose problems in all the alternative
sources and to assess whether differences tend to be consistent from year to
year.
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 167
4.3
Some indications of bias in our estimation procedures may also emerge from
the indicators on coherence discussed above. For example, if our provisional
estimates for car traffic are always too high, then our methods of estimating
expansion- and growth-factors may be deficient.
2.1
4.4 To address precision we need to produce, and publish, confidence
limits for all our principal outputs. This has been recommended in the Notes
on methodology. These are not in themselves performance indicators, but the
overall performance could be indicated by the 95% confidence limits for all
motor vehicle traffic level and all motor vehicle year-on-year changes, both
expressed as percentages. Initial work to explore the calculation of such
confidence intervals suggests this will be a complex exercise in view of the
number of stages and variables involved. It is therefore recommended that
DfT commission work to develop these.
5.
The following sets of indicators are recommended for each year’s data:
5.1









5.2
Conclusions
1. The proportion of ATC hourly data that are patched
2. The proportion of ATC hourly data that breach the different ‘warning’ levels
3. The proportion of one-day manual counts requiring a re-count
4. The proportion of one-day manual counts that breach the different ‘warning’
levels
5. The proportion of AADFs found to be significantly different from their
expected values, based upon trends applied to the previous AADFs
6. The percentage difference between the provisional annual estimates and
the final estimates
7. The percentage revisions applied to past years’ estimates, identifying the
vehicle-types and road-classes involved
8. Percentage differences in estimated vehicle-kilometres between RTS and
other sources [NTS, CSRGT, surveys of van operators and foreign operators,
HA and LA]
9. Confidence limits [expressed as percentages] for all motor vehicle traffic
levels and year-on-year changes.
For indicators 1 to 7, these should be produced by each vehicle-type.
MD
27 March 2005
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 168
Annex 8.2
THE DEPARTMENT FOR TRANSPORT
TRANSPORT STATISTICS ROADS DIVISION
OPERATIONAL REQUIREMENT DOCUMENT
FOR
THE MANUAL TRAFFIC DATA COLLECTION
(ENGLAND) 2005 & 2006
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 169
CONTENTS
PAGE NO
Background details
1
INTRODUCTION
3
2
METHOD OF DATA COLLECTION
3
Requirements
3
AREAS TO COUNT
4
4
DATES AND LOCATIONS OF SITES
4
5
COLLECTION OF TRAFFIC DATA
5
6
SUPPLY OF DATA TO THE DEPARTMENT
6
7
QUALITY OF DATA
6
8
PRICE
7
9
SUBMISSION OF DATA AND DEADLINES
8
10
SUBMISSION OF INVOICES
8
11
SAFETY AT SITES
9
12
EVALUATION CRITERIA
9
13
CROWN COPYRIGHT
9
CONTACT POINTS AT DFT
10
ANNEXES
Annex 1
How the annual traffic estimates are made
11
Annex 2
Days of the neutral weeks in 2005 and 2006
14
Annex 3
Areas for which bids are sought in each tranche
16
Annex 4
Timetable for each tranche
19
Annex 5
Detailed definitions of vehicle class
20
Annex 6
Detailed information on junction counts
22
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 170
Annex 7
National Manual Traffic Counts link information data
23
Annex 8
Format of data to be supplied
24
Annex 9
Format of itemised estimates to be supplied
25
Annex 10
Average number of sites to be counted in each year split
by type of site
26
Evaluation criteria for DfT Traffic Count 2005-2006 tenders
28
Annex 11
Background details
1. Introduction
1.1
Estimates of national traffic are made using data from two main sources: manual
counts and automatic counts. Both sources of data are used to determine annual average
daily flows on the road network. Traffic estimates are then produced by taking account of
road lengths by class of road for each authority. For 2003, the estimate of all GB motor
vehicle traffic was 493 billion vehicle kilometres.
1.2
Data collection of the continuous Manual Traffic Counts Survey is currently carried
out by Local Authorities and private contractors. The new contract is to start in January 2005
and will cover counting in all local authorities in England and will be for two years.
2. Method of data collection
2.1
The manual counts consist mainly of a series of 12-hour vehicle counts taken on the
links of the major road network (motorways and ‘A’ roads) and at a sample of minor road
sites. These counts are conducted between 7am and 7pm on weekdays during the ‘neutral’
periods of mid-March to mid-July and from mid-September to end-October, excluding Easter
and school half-term weeks. Information from the limited number of DfT automatic counters
are then used to gross up these 12-hour manual counts into annual average daily flows for
each site. This information on traffic flows is then combined with information on road lengths
to estimate total traffic in Great Britain.
2.2
There are 14,000 or so major road links in England. 1,900 of these sites are noncounting ones, which take their traffic flow values from neighbouring links. This could be
because, for instance, they are unsafe to count. About 40% of the remaining 12,100 links are
counted each year. The most important sites are counted every year and the least important
ones are counted once every eight years.
2.3
In addition to this, each year there is a sample of about 4,000 English minor road
sites counted, mostly during the neutral periods described above. These constitute a panel
sample of sites, which are counted on the same week number and day of each year, with
adjustments for the Easter period. Those counts which are done outside the neutral weeks
period are termed Summer-Winter counts. Thus the total number of manual counts
undertaken throughout Great Britain each year is about 9,000. A detailed description of the
current system is given at Annex 1.
2.4
The summer-winter counts of minor roads include a few sites where weekend
counting between 09:00 hours and 19:00 hours will be required i.e. 10 hours instead of the
usual 12 hours for weekdays. (see also section 8.3)
2.5
The Scottish Executive and Welsh Assembly are responsible for counts in their own
countries. The counts in Wales are currently carried out by the Welsh local authorities, but
they generally follow the same guidelines as those issued to contractors in England. The
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 171
counts in Scotland are currently carried under the same guidelines as those issued to
contractors in England.
Requirements
3.
Areas to count
3.1
England has been divided into 63 areas of approximately the same size 8 in terms of
numbers of major road links. These areas have been divided into three groups and they are
listed in Annex 3.
3.2
To assist tenderers in their planning, it is intended to notify all those invited to tender
of the results of the tranche one competition before the deadline for the submission of bids
for tranche two areas. Similarly, tranche two awards will be notified before the tranche three
deadline for the submission of bids. The exact dates can be seen from the timetable given in
Annex 4.
4.
Dates and locations of sites
4.1
The contractor will be asked to carry out counts for both directions of traffic and by
vehicle type at locations specified by the Department. The Department will also specify the
date on which the count should be conducted and the expected minimum and maximum
traffic flow to be found at each site, both in total and after weighting (see footnote to the table
after paragraph 7.2). Each year, the contractors will be supplied with a schedule of data and
locations for each count no later than end-October for the Summer-Winter survey and midJanuary for the neutral months survey. This schedule will reflect any changes in the major
road network. The locations will be given as 12 digit grid references from the Ordnance
Survey 1:50,000 Landranger Map series, though road names will also be provided to confirm
the location.
4.2
All vehicle counts will be scheduled for a specific date, which will normally be a
weekday during the period mid-March to mid-July and mid-September to end-October. The
contractor must ensure that sufficient enumeration teams are available to undertake all
counts scheduled by the Department within the prescribed period.
4.3
Every effort should be made to conduct counts on the scheduled date. However,
where this is impracticable, because of extensive disruption to traffic caused by planned
road works or some other factor, the contractor can reschedule counts for the same day of
the week up to 2 weeks later (providing it is still within the counting season) without
contacting the Department at the time. However, when a new date is selected the
Department must be informed within 4 weeks of the new date being selected. This rescheduling should be carried out even on the day of counting, when a sudden, major change
in traffic levels, e.g. due to a road accident, would make the daily total unrepresentative of
normal traffic flow (see paragraph 8.5 for details of compensation for lost days of
counting).The contractor must provide the appropriate local authority and Vanessa
Kovacevic at the Highways Agency with details of the dates and expected locations of the
counts to be undertaken within that authority’s area, not less than one calendar month prior
to the date of the first count. DfT will work towards providing a list of suitable contacts for this
purpose within each local authority in due course.
-796063 or Email: vanessa.kovacevic@highways.gsi.gov.uk)
4.5
At frequent intervals, the contractor must ascertain from the local highway authority
(and with neighbouring authorities for planned motorway roadworks,) whether they are
8
For administrative reasons, the Isle of Wight and the Scilly Isles are exceptions to this rule. Also, the
London areas are smaller than those outside London, because there are so many different authorities
within this region.
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 172
aware of any factors which might affect traffic flows at count sites on the dates scheduled for
counting. Temporary closures of roads on which counts are scheduled will necessitate rescheduling of such counts. Further details about planned roadworks and other disruptive
events are being made available on the National Traffic Control Centre’s website:
http://www.highways.gov.uk/knowledge/tcc/nat_cen/index.htm
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 173
5.
Collection of traffic data
5.1
All weekday counts are to be for a twelve-hour period, between 0700 and 1900
hours. Vehicle flows should be totalled by hour, by vehicle type and by direction. Separate
directional totals are not, of course, required for counts on one-way roads.
5.2
The following vehicle classifications are to be separately counted in each hour:
Pedal cycles
Motor cycles
Cars and taxis
Buses and coaches
Light vans
Rigid 2-axle goods vehicles
Rigid 3-axle goods vehicles
Rigid 4 or more-axle goods vehicles
Articulated 3 or 4-axle goods vehicles
Articulated 5-axle goods vehicles
Articulated 6 or more-axle goods vehicles
5.3
Detailed definitions of vehicle class are provided in Annex 5.
5.4
The contractor should record pedal cycles on the main carriageway and on footways,
shared use paths and segregated cycle lanes adjacent to the Census Point.
5.5
In the case of a new site without an expected traffic flow figure, the contractor is
expected to inform the Department if they have reason to believe that the likely flow will be
less than 1,000 vehicles per day, when surveying the site for the link information data. The
Department will then decide whether or not the count should proceed. In sparsely populated
areas where it is difficult to find minor road sites with flows over 1,000 vehicles, junction
counts may be carried out instead with the permission of the Department. An explanation of
how to count traffic at junction sites can be found at Annex 6.
5.6
If any count on a major road delivers a flow of less than 65 vehicles in the first three
hours of counting the count should NOT be abandoned as all major road sites need to be
counted (unless this is due to a road accident, or other such incident, in which case the
count would need to be rescheduled). However, if any count on a minor road delivers a flow
of less than 65 vehicles in the first three hours of counting the count should be abandoned
and the Department notified, unless (a) the count point is already known to be a low-flow
site (identifiable from the minimum flow data supplied by DfT), or (b) there is good reason for
the low flow, eg introduction of home zones, speed bumps, pedestriasation methods. In the
instance of (a) or (b) occurring, the count should proceed.
5.7
Video counts should only be considered as required in exceptional circumstance such as complex junctions on motorways where it is unsafe to count and only with
permission from the Department who will supply Contractors with details of Count Points
already known to require a video count. Contractors will need to contact the Department by
email or letter to confirm that they have investigated the three possibilities below before a
video count will be considered by the Department:



It is not possible to find a safe place to do a manual count by moving the S
reference to another point on the link. For example the start or end where there is
a junction that may offer a vantage point overlooking the link.
The contractor has checked that there is not a building, bridge or other structure
on an adjacent road from which traffic on any part of the link in question could be
counted.
It would be more expensive to count the traffic on one of the adjacent links and at
the same time to also count the traffic on slips roads leading onto and off of the
link in question so that flows for the link under query could be ascertained.
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 174
The contractor will also need to inform the Department of any extra costs involved in carrying
out a video count so that a price can be negotiated and agreed between the contractor and
the Department before the count is undertaken.
5.8
The Department will provide the contractor with the schedules and network
information for all the count sites in the contractor's area. All new sites, as well as existing
sites without maps, whether or not they are due to be counted in the year, will need to be
fully surveyed and checked. For the remaining sites whether they are due for counting or
not you will only need to visually check the Network Information sheet in the office for any
obvious discrepancies –e.g. where carriageways are recorded as 2 and width as 7m. New
sites are coded as normal two-way roads as the default setting and contractors will need to
notify the Department of any new sites, which are skewed or one way sites. These checks
should be carried out between February and end-April. The completed network information
sheet should be returned to the Department by end-May with information being correct as at
1 April. Details of the information to be supplied on the sheet and what needs to be done are
shown at Annex 7. The contractor is required to:
confirm the information supplied is correct, or
provide corrections and additions as necessary. Fields that have been altered should be
coloured, using red text.
5.9
Maps prepared by previous contractors, will be provided to new contractors carrying
out counts at sites that were counted during the period 1999 to 2004. Contractors who
carried out the counts in the previous contract period (2002/2004) will be expected to use
their own copies of their maps in these cases. For new sites, the contractor will be asked to
provide DfT with a detailed map themselves. Maps will also need to be prepared by the
contractor for all sites that have had major changes since the last map was provided. All
maps should give details of the actual location of the enumerators (marked by a line across
the road), place to park (if applicable), road number (if applicable), road name, local authority
title, nearby refreshment and toilet facilities (if available) and any other information needed
by an enumerator to carry out the work. All contractors will be provided with a list of Local
Authority codes, which should be used to label the maps. This is needed by the Department
to enable spot checks to be carried out without prior warning and to ensure that counts in
future years are conducted from the same spot. The Department will supply maps to any
new contractor to ensure continuity of data.
6.
Supply of data to the Department
6.1
All count data should be sent by email to the Department’s ‘manual traffic count’
address (roadtraff.manual@dft.gov.uk) in Excel format. Contractors covering more than one
area should send a separate file or separate spreadsheet for each area. The data supplied
should include all the information specified at Annex 8. Any data received not in the required
format will be returned to the contractor for amendment. Explanations should be given (in the
comments column of the itemised estimate) for abnormal hourly fluctuations, especially on
high flow sites, and for 12-hour flows that are significantly outside the expected minimum
and maximum figures.
7.
Quality of data
7.1
The contractor should provide details of training provided to enumerators, equipment
used, level of supervision applied, the length of shifts, method of allocating enumerators to
sites and anything else related to ensuring that data supplied to the Department is of a high
quality. As well as describing what steps are taken to obtain good data, contractors should
also specify the checks that they will undertake on the data to ensure that accurate and
complete hourly vehicle count data is supplied with link information sheet.
7.2
Sites will be allocated to one of six road types (6-lane motorway with 90,000+ AADF
to minor road site with less than 5,000 vehicles per day). The minimum number of
enumerators counting at any one time, will be :
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 175
Site category
Weighted9 annual
average daily flow
1. V. busy dual
carriageway
2. Busy road
90,000+
Minimum no. of enumerators
counting at any time of the
day
6 (7 for peak hours)
60,000-89,999
4 (5 for peak hours)
3. Quite busy road
40,000-59,999
4
4. Ordinary road
20,000-39,999
2 (3 for peak hours)
5,000-19,999
2
Under 5,000
1 (2 for peak hours/safety
reasons)
5. Fairly quiet road
6. Quiet road
9
Cars are allocated a weight of 0.8 and other vehicles are given a weight of 2. For most roads, the
weighted flow will equal the unweighted flow, but those with a more difficult mix of traffic, with extra
goods vehicles, will have a higher weighted flow.
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 176
Annex 1
7.3
Whilst only one enumerator may be required for counting at any one time on a quiet
road, it is expected that a second enumerator will be present for most of the day for health
and safety reasons. For category 1,2,4 and 6 roads, DfT expect more enumerators to be
used during the peak hours (7.00 hrs to 9.59hrs and 16.00hrs to 17.59hrs).
7.4
Contractors may specify higher levels of staffing than that shown in the table above.
This should be accompanied in their tender by an explanation of why they feel this is
necessary and how they will be deployed. This is needed so that the presumed better data
quality from the arrangement can be judged against the likely extra expense of the proposal.
7.5
Spot checks will be carried out by DfT staff. If the specified number of enumerators
are not counting at the site or the count data is later to be found to be frequently inaccurate,
DfT has the right to terminate the contract with the contractor. In practice, there would be a
warning to be followed up by extra spot checks before a final decision is taken.
8.
Price
8.1
For each area of England, tenders will need to be expressed in terms of a price per
site category (rounded to the nearest pound). These prices will then be matched with the
number of different types of sites in the bid area, as given in Annex 10, in order to determine
the overall price of the bid. Annex 10 gives details of the sites to be surveyed in an average
year over the coming two years in each area, so that potential contractors can assess the
mix of sites that will need to be counted. Contractors should also provide a price per site
(rounded to the nearest pound) for both types of network information checks, on the
assumption that about 5% of all sites will need a full check in any year.
8.2
Occasionally sites not previously counted will need to be counted. In these cases
payment will be made on the basis of the weighted flow of traffic obtained from the actual
count carried out. For this reason also, a price should be given for all types of roads, in case
a new or relocated site has a weighted flow unlike those expected for the area (this need not
be done for the areas of the Scilly Isles and the Isle of Wight).
8.3
Counting at weekend sites will be reimbursed at the normal fixed price per site, even
though the counting period on these days is 10 hours not 12 hours.
8.4
Contracts will be awarded either for individual local highway authority areas or
combinations of areas.
8.5
Where counting is disrupted and has to be abandoned or rescheduled for reasons
outside the contractor’s control, the following percentage compensation payments will be
made by the Department:
Cancelled/Resched DfT/Other
uled
circumstances
Either
Major incident after 5
hours counting
completed
Notification from DfT
n/a
%
compensation
100
Either
Major incident before 5
hours counting
completed
n/a
50
Cancelled
<65 vehicles in 1st 3
hours
n/a
50
Cancelled
DfT decision
>2 weeks
0
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 177
Annex 1
Cancelled
DfT decision
Cancelled
DfT decision
Rescheduled
Planned roadworks, etc.
preventing counting
beginning
>=3 working days
and <2 weeks
<3 working days
25
100
n/a
0
n/a
50
Emergency roadworks,
etc.
Preventing counting
beginning
Rescheduled
DfT decision
>2 weeks
0
Rescheduled
DfT decision
0
Rescheduled
DfT decision
>=3 working days
and <2 weeks
<3 working days
20
8.6
Payment for counting will not be made, if the count data or link information sheet data
are incomplete or contain obvious errors. Price per site increases will only be made with
prior consent from the Department.
8.7
The compensation for counts rescheduled because of roadworks preventing counting
beginning will only be paid if the contractor can demonstrate that they took sufficient steps to
check that counting was possible on the scheduled date.
9.
Submission of data and deadlines
9.1
Data for counts carried out between January and end-April should be returned to the
Department by end June. For counts carried out from May to mid-July, data should be sent
by mid-September. For counts between mid-July and mid-December, data should be
returned by end-December. Count data should be accompanied by an electronic itemised
estimate, as described in Annex 9.
9.2
Contractors can provide data more quickly than this, on a monthly basis, if they wish.
10.
Submission of invoices
10.1
Electronic itemised estimates for billing purposes should be submitted three times a
year with the count data as described in paragraph 9.1 for approval by the Department before
any invoices are issued. The estimate should include all the information specified in Annex 9.
The Department will normally respond to estimates within 3 weeks providing the data and
estimate have been sent in the correct format. Once an agreement has been reached between
the Department and the contractor regarding the estimate, an invoice may be issued and will be
paid within 30 days of receipt of the invoice.
10.2
Contractors can provide estimates more quickly than this, on a monthly basis, if
they wish providing the data is submitted at the same time.
11.
Safety at Sites
11.1 The Contractor shall at all times ensure that all staff employed on site are aware of and
comply with, the Health and Safety at Work Act.
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 178
Annex 1
11.2 The Contractor shall comply with all safety instructions issued to the Contractor by the
Department, the relevant Highway Authority or the Police.
11.3 The Contractor's attention is particularly drawn to the hazards of working on motorway
sites. The recommendations contained in guidance documents issued by the Department shall
be followed at all times.
11.4 High visibility jackets must be worn at all times when working outside a vehicle.
Vehicles must clearly show official notice or identification in car window.
12. Evaluation criteria
12.1
area.
Annex 11 gives details of the criteria to be used in selecting the suitable agent for each
13. Crown Copyright
13.1 Raw un-validated data collected by contractors is the property of the Department.
Contractors must not supply this to anyone other than the local authority in whose area the
counts have been conducted. Private contractors may levy a nominal charge to cover the
marginal costs of providing this information to the local authority if they require the data prior
to the validated data being available from the Department. But they should bear in mind that
they are dependent on the goodwill of authorities who provide them with information about
forthcoming road works etc.
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 179
Annex 1
Contact Points at DFT
Further information, if required, may be obtained from:
Ms Ann Seal ( 020-7944 6555)
or
Mrs Hashmia Cole ( 020-7944 6393)
or
Mr Brian Corbett ( 020-7944 6603)
Department for Transport (DfT)
TSR2
Zone 2/16
Great Minster House
76, Marsham Street
London, SW1P 4DR
Fax no: 020-7944 2164
Email: roadtraff.manual@dft.gov.uk
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 180
Annex 1
How the annual traffic estimates are made
1.
Introduction
24. In 2002, motor vehicles travelled about 485 billion kilometres while pedal cycles
travelled 4.4 billion kilometres along the public roads of Great Britain. How do we
know this? We know this from the traffic counts conducted about many different
types of roads and information on road lengths. However it is not easy to convert
count data to total traffic data
25. The road network consists of about 50,000 kilometres of motorways and class "A"
roads, with a further 340,000 kilometres or so of minor roads. The road system as
a whole is thus much too extensive to allow the collection of comprehensive
traffic data for every part. Moreover, the density of traffic carried (and the mix of
traffic by vehicle type) varies enormously from place to place and from hour to
hour. Flows are less than 100 vehicles a day on many minor roads but exceed
150,000 a day on some motorway links. Even within a road class, one site may
easily carry ten times as much traffic as another. Flow also varies by time of day,
by day of week and by month of year. Furthermore, there is variation within the
variation - car traffic levels, for example, may change relatively little over the
seven days of the week, but goods traffic is usually at far lower levels on
Saturdays and Sundays than on other days. These characteristics mean that
estimating national traffic volumes requires a fairly complicated sampling design,
the collection of substantial volumes of data and complex computational
procedures. This note describes the new methodology used for the traffic
estimates from 1993 onwards.
26. Estimation of traffic levels uses information from both manual and automatic
counts. They are each described briefly below.
The manual counts (previously referred to as the rotating census and the
biennial counts)
27. These counts operate somewhat differently for major and minor roads. The major
roads are split into five road classes: motorways, trunk roads and principal roads
with the latter two divided into urban and rural roads. Urban10 roads are defined
as those within the boundaries of the Urban Area polygons for settlements of
10,000 population or more, based on the 2001 Population Census. On the
outskirts of urban areas, bypasses are normally treated as rural even if part of the
road may lie within the urban area polygon. Conversely, roads between urban
areas with short lengths outside the polygons are normally treated as urban.
Minor roads are divided into 6 classes: B class, C class and U (unclassified)
roads, each sub-divided into urban and rural.
28. For major roads (motorways and A-roads), the traffic on every link - normally a
section of road between consecutive junctions with other major roads - must be
10
Before 8 May 2003, roads were, instead, classified as built-up and non built-up. Built-up roads were
those with a speed limit of 40mph or less. Non built-up roads were those with a higher speed limit.
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 181
Annex 1
regularly assessed. This is done by counting the traffic at a statistically random
point on most links at regular intervals – traditionally, once every three years in
England and Wales and once every six years in Scotland11. It is recognised that
with the exception of motorways, traffic levels will vary along the length of a link.
However, the procedure of counting at a statistically random point on each link
can be expected to lead to good estimates at national level although estimates on
some individual links may be less reliable.
29. In total, about 5,300 major road sites are scheduled to be counted in 2004. In
addition to traffic count data, information is collected about the characteristics of
each link, such as its length and the road class and road width at the place of the
count. At each chosen point, trained enumerators count vehicles of each of
eleven types (pedal cycles, two-wheeled motor vehicles, cars and taxis, buses
and coaches, light vans, and six separate categories of goods vehicle) for the 12
hours from 7am to 7pm. These counts are all scheduled to take place on
weekdays, but not on or near to public holidays or school holidays. To minimise
the effects of possible seasonal factors, counting is confined to the so-called
"neutral weeks". These are namely most weeks in March, April, May, June,
September and October.
30. Some major road links are unsafe to count or are too short to be worth counting
in the normal way. In these cases, traffic estimates are derived from the judicious
use of flow data on adjacent links. These are called derived links. Further,
because all links are now defined as ending at a local authority boundary, some
links are treated as dependent links. In these cases, it is assumed that the flow is
the same along all of the link. So, a count in one local authority can be used as a
proxy for the flow on the dependent link. In 2003, there were 15,500 normal links,
1,200 derived links and 1,000 dependent links. Complete coverage of the minor
road network is not attempted as it is too extensive. It is not practicable to define
the minor road network in terms of individual links; even if all the links could be
identified, their number would be far too great to allow traffic data for each link to
be collected. Minor road traffic estimates are therefore made by grouping minor
roads into one of the six road classes12. An attempt is then made to measure the
average flow on each of these road types by carrying out a number of counts
along them. A random sample of approximately 4,500 sites across GB is visited
each year. These same sites are counted each year. Most of these counts are
carried out in neutral weeks. However about 200 counts per year, the "summerwinter counts", are carried out in non-neutral weeks and on weekends. These
200 sites are visited each year and are mainly used to provide extra information
about two-wheeled traffic throughout the year. This is because pedal cycles and
motor cycles are not always accurately identified by automatic counters (see
paragraph 9).
11
Since 1999, the frequency of counting has been broadly based on the variance of the traffic of the
link (variance of average flow * link length). Intervals are every one, two, four and eight years in
England and Wales. In Scotland they are two, four, eight and sixteen years. However, Scotland is
proposing to increase its budget for traffic counting and so may come into line with level of counting in
England and Wales.
12 These include: urban and rural 'B' roads, 'C' roads and unclassified roads
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 182
Annex 1
31. The manual counts have the advantage over automatic counts of complete
coverage of major road sites and moderately good coverage of minor roads.
However, the data (hourly by vehicle type), are very sparse since traffic is
counted for only 12 hours on each visit. Thus these counts give no information
about traffic at night, at weekends, over public holiday periods, and little about the
seven non-neutral months. In calculating national traffic estimates, therefore, use
must be made of data from automatic counters.
The automatic counts (previously referred to as the core census)
32. The automatic counters fill the gaps left by the manual counts. There are some
160 sites in GB outside of London where traffic is monitored continuously using
automatic sensors, which classify the traffic into vehicle type. The numbers of
vehicles of each type detected are combined into hourly totals and stored on-site
until it is downloaded during the night to a computer in the DfT headquarters
building. The automatic counting equipment recognises 22 different types of
vehicle; these are then combined to provide estimates for the eleven vehicle
types13 used by DfT.
33. The automatic counters do not give 100% accuracy. For example, the equipment
cannot classify vehicles into their different types when the traffic is moving very
slowly (5mph or less). They also have a tendency to malfunction, though the new
sensors recently introduced, are more reliable than the previous ones. The
equipment cannot distinguish between cars and car-based vans and can also
have difficulty distinguishing between some types of buses and coaches and
goods vehicles having similar axle spacing and chassis height. The equipment is
also prone to failing to identify two-wheeled vehicle traffic, both bicycles and
motorcycles14. Nevertheless, they do have the big advantage over manual counts
that they operate continuously and so can give a complete picture of traffic at the
points where they are sited.
34. The automatic counters in London are slightly different to those outside London.
There are 56 automatic counters in London and they are "volumetric" classifiers
that only distinguish between short (up to 5.2 metres) and long (greater than 5.2
metres) vehicles. They need 24-hour manual counts every three months to
provide estimates of the breakdown of traffic by vehicle type in each hour of the
day. These counters suffer from similar problems as those outside London, but
are more reliable than the automatic counters used outside London.
Annual Average Daily Flows (AADFs)
35. The data for all manual counts done in neutral months15 are combined with
information from automatic counters on similar roads to provide an estimate of
13
These include: Pedal cycle, two-wheeled motor vehicle, car, light goods van, bus, rigid 2 axle lorry,
rigid 3 axle lorry, rigid 4 or more axle lorry, 3 axle or 4 axle articulated, 5 axle articulated and 6 axle or
more articulated.
14 This deficiency underpins the need for the extra sites for manual counting on minor roads in the
summer and winter months
15 Counts done in the summer and winter months are not grossed-up: the sites are normally counted
at the same time of the year each year and so are compared directly with each other.
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 183
Annex 1
the AADF at that site. This is normally done by multiplying the raw count data by
factors derived from the automatic counts in that same year. There are a large
number of such expansion factors since there are separate factors for each
vehicle type, day of counting and expansion factor group16. Because these
counts are done in neutral weeks, the expansion factors used do not usually vary
too much from year to year, except when bad weather has restricted traffic during
the winter months. For cars, the factors are usually between 1.15 and 1.25
(except on motorways and in London where the factors are higher) while for
goods vehicles the factors vary between 0.75 and 0.9 - lower because of the
greater drop in traffic at weekends.
36. The automatic counters provide a reasonable guide to changes in traffic over
time. This information is used in two ways. Firstly it is used to provide provisional
quarterly estimates of traffic, which are published on the sixth Thursday following
the end of a quarter. Secondly they are used to provide growth factors between
consecutive years. These growth factors are used for links not counted, or not
counted satisfactorily in the latest year. In these cases, the AADF for the previous
year is multiplied by the appropriate growth factor to give a reasonable estimate
of the AADF for the latest year.
Use of AADFs in Calculation of Annual Traffic Estimates
37. Different procedures are used for major and minor roads in converting AADF data
to traffic estimates. The difference arises because the link concept cannot be
applied to minor roads.
38. Major roads: A major road link of length 2 km with an AADF of 50,000 has a
traffic figure of 100,000 vehicle-kilometres (2*50,000). This equates to 36.5
million vehicle kilometres a year. Because every major road link is counted, in
principle, total traffic on major roads can be obtained by summing the traffic
figures for every link.
39. As mentioned in paragraph 6, some links are not counted. In these cases, the
traffic flows are derived from adjacent links using suitable formulae (derived links)
or using the flow of the adjacent link as a proxy (dependent links).
40. Minor roads: In the base year (currently 1998), for each minor road class in each
region a mean AADF is calculated as a simple average of all the available
AADFs, including those projected forward from counts done in earlier years.
These averages are then multiplied by the total road length for the relevant minor
road category to give an estimate of traffic for that road category.
41. Traffic for the latest year is then obtained by calculating changes in traffic flows
derived from the fixed sample of minor road sites, after taking into account any
changes in road length. It is assumed that new minor roads have the same
average flows as that of other minor roads. This is plausible, since some of the
newest will be quiet roads on housing estates whilst others will be busy roads
recently declassified from major road status.
16
There are now 22 expansion factors groups. These are based on type of area (from holiday area to
Central London), road category and, in some cases, traffic flow level.
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 184
Annex 1
Quarterly and annual estimates of traffic
As mentioned in paragraph 12, the automatic counters are used to provide
provisional quarterly estimates throughout the year. A first estimate for the year is
published in early February, and this is largely based on automatic count data. A
final estimate for the year is normally published in early May and this is produced by
putting together the estimates for major and minor road traffic as detailed above.
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS 185
Annex 2
Days of neutral weeks in 2005
I.
Dates for counting will be determined by internationally agreed week numbers17. This will mean
that counts in each year will be related to particular days of particular weeks. Some local
authorities are considering changes to term lengths and times. Contractors should check with
the appropriate local authorities and inform the Department if counts are scheduled during such
rearranged holidays, as appropriate.
Week
Exclusions
numbe
rs
Description
of counts
1
No counts
2 to
10.8
5 days - Week 8 or
9 (half-term week)
Summer/Wint
er counts
10 Jan to
17 Mar
12 to
21
11 days - Week
before and week
after Easter, Early
May Bank Holiday
Monday
-
Neutral month
counts up to
late May Bank
Holiday
week18
Summer/winte
r counts
Neutral month
counts
18 Mar and
4 Apr to 27
May
22
18
-
29 to
37
1 day - August Bank
Holiday Monday
Summer/Wint
er counts
18 Jul to 16
Sep
38 to
43.4
5 days - half term
week
Neutral month
counts
19 Sep to 1
Nov
43.6 to
50
-
Summer/Wint
er counts
2 Nov to 16
Dec
No counts
Week 1 of a year is the week that contains the first Thursday of the year.
May Bank Holiday week in non-leap year is always in week 22.
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS
186
Number
of
neutral
days
counted
40
30 May to 3
June
6 Jun to 15 30
Jul
23 to
28
51 to
end
17
2005 dates
27
Annex 2
Days of neutral weeks in 2006
I.
Dates for counting will be determined by internationally agreed week numbers19. This will
mean that counts in each year will be related to particular days of particular weeks.
Week
Exclusions
numbe
rs
Description
of counts
1
No counts
2 to
10.8
5 days - Week 8 or
9 (half-term week)
Summer/Wint
er counts
9 Jan to 16
Mar
12 to
21
11 days - Week
before and week
after Easter, Early
May Bank Holiday
Monday
-
Neutral month
counts up to
late May Bank
Holiday
week20
Summer/winte
r counts
Neutral month
counts
17 Mar to 7
April and
24 Apr to
26 May
22
20
-
29 to
37
1 day - August Bank
Holiday Monday
Summer/Wint
er counts
17 Jul to 15
Sep
38 to
43.4
5 days - half term
week
Neutral month
counts
18 Sep to
31 Oct
43.6 to
50
-
Summer/Wint
er counts
1 Nov to 15
Dec
No counts
Week 1 of a year is the week that contains the first Thursday of the year.
May Bank Holiday week in non-leap year is always in week 22.
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS
187
Number
of
neutral
days
counted
40
29 May to 2
June
5 Jun to 14 30
Jul
23 to
28
51 to
end
19
2006 dates
27
Annex 3
Areas for which bids are sought in each tranche
Region
First tranche
Second tranche
Third tranche
North East
Durham exc.
Darlington
Cleveland inc.
Darlington UA
Cumbria
Cheshire
Northumberland
North Tyne and Wear
South Tyne and Wear
North Gt. Manchester
South Gt. Manchester
Lancashire
North West
Merseyside West Merseyside
East Merseyside
Yorkshire & West Yorkshire - East
Humber
North Yorkshire
West Yorkshire West
South Yorkshire North
Humberside exc. E
Riding
Derbyshire
South Yorkshire South
E Riding inc. Goole
West Midlands - West
Warwickshire
Shropshire
Staffordshire
Herefordshire
South West Avon
Cornwall
Devon
Scilly Isles
Eastern
Essex
Suffolk
Wiltshire
Dorset
Somerset
Gloucestershire
Bedfordshire
Hertfordshire
Cambridgeshire
Norfolk
South East
Hampshire
Oxfordshire
Berkshire
Surrey
E Sussex
Buckinghamshire
W Sussex
Isle of Wight
Kent
Greater
London
NW Outer London
South Outer London
Central London
NE Outer London
East Inner London
SW Outer London
South Inner London
SE Outer London
North Inner London
East
Midlands
Lincolnshire
Nottinghamshire
West
Midlands
Worcestershire
West Midlands - East
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS
188
Leicestershire
Northamptonshire
Annex 3
TSR Area names
Authorities included
Avon
Bath & North East Somerset, City of Bristol, North
Somerset, South Gloucestershire
Luton, Bedfordshire excl. Luton
Bedfordshire
Berkshire
Buckinghamshire
Cambridgeshire
Central London
Bracknell Forest, West Berkshire, Reading, Slough,
Windsor & Maidenhead, Wokingham
Milton Keynes, Buckinghamshire excl. Milton Keynes
Cheshire
Peterborough, Cambridgeshire excl. Peterborough
Hammersmith and Fulham, Kensington and Chelsea,
Westminster
Halton, Warrington, Cheshire excl. Halton & Warrington
Cleveland inc.
Darlington UA
Cornwall
Hartlepool, Redcar & Cleveland, Middlesborough,
Stockton-on-Tees, Darlington
Cornwall
Cumbria
Derbyshire
Devon
Cumbria
Derby City, Derbyshire excl. Derby City
Plymouth, Torbay, Devon excl. Plymouth & Torbay
Dorset
Bournemouth, Poole, Dorset excl. Poole & Bournemouth
Durham exc. Darlington
UA
East Inner London
East Merseyside
East Riding inc. Goole
East Sussex
Durham exc. Darlington
Essex
Gloucestershire
Southend, Thurrock, Essex excl. Southend & Thurrock
Gloucestershire
Hampshire
Portsmouth, Southampton, Hampshire excl. Portsmouth &
Southampton
Herefordshire
Herefordshire
Hertfordshire
Humberside exc. East
Riding
Kent
Lancashire
Leicestershire
Lincolnshire
NE Outer London
Norfolk
Hackney, Newham, Haringey, Tower Hamlets
Knowsley, St. Helens, Sefton
East Riding inc. Goole
Brighton and Hove, East Sussex excl. Brighton & Hove
Hertfordshire
Kingston-upon-Hull, North East Lincolnshire, North
Lincolnshire
The Medway Towns, Kent excl. The Medway Towns
Blackburn with Darwen, Blackpool, Lancashire excl.
Blackburn & Blackpool
Leicester City, Rutland, Leicestershire excl. Leicester &
Rutland
Lincolnshire
Havering, Barking & Dagenham, Enfield, Redbridge,
Waltham Forest
Norfolk
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS
189
Annex 3
North Gt Manchester
North Inner London
Bolton, Bury, Oldham, Rochdale, Wigan
Camden, Islington, City of London
North Tyne and Wear
North Yorkshire
Newcastle, North Tyneside
North Yorkshire excl. York, York
Northamptonshire
Northumberland
Nottinghamshire
NW Outer London
Oxfordshire
Northamptonshire
Northumberland
Nottingham, Nottinghamshire excl. Nottingham
Barnet, Brent, Ealing, Harrow
Oxfordshire
Scilly Isles
SE Outer London
Shropshire
Somerset
Scilly Isles
Bromley, Bexley, Greenwich
The Wrekin, Shropshire excl. The Wrekin
Somerset
South Gt Manchester
Manchester, Salford, Stockport, Tameside, Trafford
South Inner London
South Outer London
Lambeth, Lewisham, Southwark
Croydon, Sutton, Wandsworth, Merton
South Tyne and Wear
Gateshead, South Tyneside, Sunderland
South Yorkshire - North
South Yorkshire - South
Barnsley, Doncaster
Rotherham, Sheffield
Staffordshire
Stoke-on-Trent, Staffordshire excl. Stoke
Suffolk
Surrey
Suffolk
Surrey
SW Outer London
Warwickshire
Kingston-u-Thames, Richmond-u-Thames, Hounslow,
Hillingdon
Warwickshire
West Merseyside
Wirral, Liverpool
West Midlands - West
Sandwell, Dudley, Walsall, Wolverhampton
West Midlands- East
West Sussex
West Yorkshire - East
West Yorkshire - West
Wiltshire
Worcestershire
Birmingham, Coventry, Solihull
West Sussex
Wakefield, Leeds
Bradford, Calderdale, Kirklees
Swindon, Wiltshire excl. Swindon
Worcestershire
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS
190
Annex 4
Timetable For Each Tranche
TRANCHES
DATES
Invitation to tender for all tranches sent out
11 June 2004
TRANCHE 1
Deadline for receiving bids of first tranche only.
26 July 2004
Results of first tranche posted to all invited to
tender.
20 August 2004
TRANCHE 2
Deadline for receiving bids of second tranche
only.
06 September 2004
Results of second tranche posted to all invited
to tender.
01 October 2004
TRANCHE 3
Deadline for receiving bids of third tranche.
18 October 2004
Results of third tranche posted to all invited to
tender.
19 November 2004
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS
191
Annex 5
The following vehicle classifications are to be separately counted in each hour:
Pedal cycles
Two-wheeled motor vehicles
Cars and Taxis
Buses and coaches
Light goods vehicles
Rigid 2-axle goods vehicles
Rigid 3-axle goods vehicles
Rigid 4-axle goods vehicles
3 or 4-axle goods vehicles (articulated or with trailer)
5-axle goods vehicles (articulated or with trailer)
6 or more axle goods vehicles (articulated or with trailer)
The definitions for the above vehicle types are given below:
Pedal cycles
Includes all non-motorised cycles.
Two-wheeled motor vehicles
Includes motorcycles, scooters and mopeds and all motorcycle or scooter combinations.
Cars and taxis
Includes estate cars, all light vans with windows to the rear of the driver's seat and other passenger vehicles
with a gross vehicle weight of less than 3.5 tonnes, normally ones which can accommodate not more than 15
seats. Three-wheeled cars, motor invalid carriages, Land Rovers, Range Rovers and Jeeps and smaller
ambulances are included. Cars towing caravans or trailers are counted as one vehicle.
Buses and coaches
Includes all public service vehicles and works buses with a gross vehicle weight of 3.5 tonnes or more, often
vehicles with less than 16 seats.
Light goods vehicles
Includes all goods vehicles up to 3,500 kgs gross vehicle weight. This includes all car delivery vans and those
of the next larger carrying capacity such as transit vans. Included here are small pickup vans, three-wheeled
goods vehicles, milk floats and pedestrian controlled motor vehicles. Most of this group are delivery vans of
one type or another and it includes those vehicles that might previously have been coded as medium goods
vehicles.
This picture shows an example of a light good vehicle which has no side guards fitted, that can be
difficult to classify. Goods vehicles over 3,500 kgs gross weight have sideguards fitted between
axles.
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS
192
Annex 5
Heavy Goods vehicles:
Rigid with two axles
Includes all rigid vehicles over 3,500 kgs gross vehicle weight with two axles. Includes larger
ambulances, tractors (without trailers), road rollers for tarmac pressing, box vans and similar large
vans. A two axle motor tractive unit without trailer is also included.
Rigid with three axles
Includes all non-articulated goods vehicles with three axles irrespective of the position of the axles.
Excludes two axle rigid vehicles towing a single axle caravan or trailer. Three axle motor tractive units
without a trailer are also included.
Rigid with four or more axles
Includes all non-articulated goods vehicles with four axles, regardless of the position of the axles.
Excludes two or three axle rigid vehicles towing a caravan or trailer.
Articulated with three axles (or with trailer)
Includes all articulated vehicles with three axles. The motor tractive unit will have two axles and the
trailer one. Also included in this class are two axle rigid goods vehicles towing a single axle caravan or
trailer.
Articulated with four axles (or with trailer)
Includes all articulated vehicles with a total of four axles regardless of the position of the axles, i.e. two
on the tractive unit with two on the trailer, or three on the tractive unit with one on the trailer. Also
includes two axle rigid goods vehicles towing two axle close coupled or drawbar trailers.
Articulated with five or more axles (or with trailer)
This includes all articulated vehicles with a total of five axles regardless of the position of the axles.
Also includes rigid vehicles drawing close coupled or drawbar trailers where the total axle number
equals five and articulated vehicles where the motor tractive unit has more than one trailer and the
total axle number equals five.
Articulated with six or more axles (or with trailer)
This includes all articulated vehicles with a total of six or more axles regardless of the position of the
axles. Also includes rigid vehicles drawing close coupled or drawbar trailers where the total axle
number equals six or more and articulated vehicles where the motor tractive unit has more than one
trailer and the total axle number equals six or more.
Articulated goods vehicles
Please note, it is important when a goods vehicle is travelling with one or more axles raised
from the road (sleeping axles or hobos) that the vehicle is classified according to the number
of axles on the road, and not the total number of axles.
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS
193
Annex 6
Detailed Information on Junction Counts
2.2 We are expanding our minor road sample to include counts undertaken at junctions, so that
we can count at some quite remote locations but still achieve reasonable vehicle flows. A
Junction Count should only be undertaken once it has been agreed with the Department
that the existing location has flows that are too low, and there are no other suitable
alternatives in the vicinity. A suitable junction should then be chosen, as close as is
practical to the originally selected site location – i.e. within 20kms. The junction could be a
T-junction or a crossroads.
Characteristics



Should only be needed for UR and CR roads
The junction must be two roads of the same road class, and hopefully same speed (unless in
exceptional circumstances DFT agree that the junction can be where a unclassified road
meets a C road)
Count any vehicle passing the centre of the junction in any direction, [so this is not a turning
count which is a type of count some LAs commission for their own purposes]
Recording sites on the network






As Junction Counts only apply to CR and UR roads, we assume that most of the details will be
the same for both of the roads at the junction. For example the : road class, speed, number of
lanes, and number of carriageways are unlikely to differ
A Junction Count should be indicated using the two new Types of Count codes that have been
created - i.e. NJ & SJ. These show that the site is a Junction Count but still records whether
the time of the count is a Neutral Week or a Summer Winter.
The S ref should be located at the centre of the junction and needs to be as accurate as
possible.
Two road names should be recorded if they are known
The width of both roads should be measured and recorded as an average of the two.
Road numbers should be recorded if known, the first one [matching the number of the road
that was originally selected] in the Road Number field and the second in the CP Location field)
Returning data



There should only be twelve rows of data.
A new code J should be used in column E to identify a junction count instead of N, S, E or W
as the count will be combined not directional.
As there are only 12 rows of data for junction road counts column F should always
contain the code 1 (even though the roads are not actually 1 way !).
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS
194
Annex 8
National Manual Traffic Counts Link Information Data
21
Local Highway Authority
Road Number
Road Name (if available)
Census Point Grid Reference
A-REF Grid Reference
(start of link)
B-REF Grid Reference
(end of link)
[Motorways and major roads only]
[Motorways and major roads only]
Junction Number ( or identifier )
for start of link
[Motorways only]
Junction Number ( or Identifier )
for end of link
[Motorways only]
Road Class*
Length of link built-up
[Motorways and major roads only]
(in kilometres to nearest 0.1 km)
Length of link non built-up
[Motorways and major roads only]
(in kilometres to nearest 0.1 km)
Number of Carriageways at Census Point
Width [in metres] of Carriageway (s) at Census Point
Number of Lanes at Census Point
Speed Limit at Census Point [in MPH]
* ROAD CLASS
The contractor will need to confirm or provide a road class from one of the following:-
Trunk Motorway
Principal Motorway
Trunk urban (formerly known as 'built-up')
Trunk rural (formerly known as 'non built-up')
Principal urban (formerly known as 'built-up')
Principal rural (formerly known as 'non built-up')
]
]
] “A” roads
]
]
“B” urban (formerly known as 'built-up')
“B” rural (formerly known as 'non built-up')
“C” urban (formerly known as 'built-up')
“C” rural (formerly known as 'non built up')
Unclassified urban (formerly known as 'built-up')
Unclassified rural (formerly known as 'non built-up')
21
]
]
] Minor roads
]
]
]
Subject to change in the light of legislation
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS
]
]
]
] Major roads
]
]
]
]
195
Annex 8
Format of the data to be supplied
EXCEL COUNT DATA FORMAT
When returning hourly data in EXCEL format please ensure that column contents are as follows:-
COLUMNS
DESCRIPTION OF FIELD CONTENTS
A
B
C
D
E
F
CP number
Area code 22(see footnote)
Day (see below)
Date/month/year e.g. 15/04/02
Direction (N, S, E or W)
If count is one-way insert the digit 1 here
If count is two-way insert the digit 2 here
If count is skewed insert the digit 3 here
Hour on 24 hour clock, e.g. 7, ...., 18.
Pedal cycles
Two-wheeled motor vehicles
Cars and taxis
Buses and coaches
Light vans
Rigid 2-axle goods vehicles
Rigid 3-axle goods vehicles
Rigid 4-axle goods vehicles
3 or 4-axle goods vehicles (articulated or with trailer)
5 axle goods vehicles (articulated or with trailer)
6 or more axle goods vehicles (articulated or with trailer)
G
H
I
J
K
L
M
N
O
P
Q
R
Format of Days
Mon Thur Sun
Tue Fri
Wed Sat
22Area
numbers will be supplied along with schedules.
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS
196
Annex 9
Format of itemised estimates to be supplied
When submitting itemised estimates for billing purposes in EXCEL format please ensure that column
contents are as follows:-
COLUMNS
A
B
C
D
E
F
DESCRIPTION OF FIELD CONTENTS
CP number
Date Scheduled
Date Counted
Road Number
Cost to be invoiced
Comments (This should include explanations for rescheduling and cancellations, higher or
lower flows than expected or any other information that the Department should be aware
of
regarding the count).
Contractors covering more than one area should send a separate file or separate spreadsheet for each
Area clearly marked with the TSR2 area number.
MANUAL TRAFFIC DATA OPERATIONAL REQUIREMENTS
197
Annex 10
Average number of sites to be counted in each year split by type of site
The table below shows the number of sites likely to be needed to be counted each year in each
area. They are provided here only for indicative purposes.
TSR2 area
Very busy Busy
dual
road
carriageway
Quite
busy
road
Ordinary
road
Fairly Quiet
quiet road
road
All
roads
AVON
BEDFORDSHIRE
BERKSHIRE
BUCKINGHAMSHIRE
CAMBRIDGESHIRE
CENTRAL LONDON
CHESHIRE
CLEVELAND inc.
DARLINGTON
CORNWALL
CUMBRIA
DERBYSHIRE
DEVON
DORSET
DURHAM exc.
DARLINGTON
EAST INNER LONDON
EAST MERSEYSIDE
EAST RIDING inc.
GOOLE
EAST SUSSEX
ESSEX
GLOUCESTERSHIRE
HAMPSHIRE
HEREFORDSHIRE
HERTFORDSHIRE
HUMBERSIDE exc.
EAST RIDING
ISLE OF WIGHT
KENT
LANCASHIRE
LEICESTERSHIRE
LINCOLNSHIRE
NE OUTER LONDON
NORFOLK
NORTH GT.
MANCHESTER
NORTH INNER
LONDON
3
3
7
8
1
6
13
1
9
2
2
8
9
4
4
6
3
12
13
15
10
6
24
13
23
23
14
34
22
17
65
39
37
49
62
16
93
31
61
27
50
61
66
14
76
55
168
85
131
143
164
94
218
114
5
1
1
1
2
2
6
18
12
11
5
10
8
12
33
21
5
46
32
65
54
51
28
47
55
69
99
58
36
103
101
170
201
141
77
3
1
-
7
-
11
8
4
28
11
3
18
28
22
23
33
30
90
81
59
5
14
10
-
2
22
5
10
5
-
3
9
1
15
4
6
24
30
16
35
3
35
7
43
76
44
104
12
56
28
60
140
51
134
26
76
37
132
282
117
312
41
186
78
13
3
7
9
10
4
5
2
8
3
17
14
12
3
9
7
1
36
32
29
5
28
22
32
6
99
89
49
47
27
59
71
9
112
88
71
70
34
73
54
16
281
231
170
125
115
154
177
1
6
8
34
24
14
87
TRAFFIC CENSUS OPERATIONAL REQUIREMENTS
26
Annex 10
All
roads
TSR2 area
Very busy Busy
dual
road
carriageway
Quite
busy
road
Ordinary
road
Fairly Quiet
quiet road
road
NORTH TYNE AND
WEAR
NORTH YORKSHIRE
NORTHAMPTONSHIRE
NORTHUMBERLAND
NOTTINGHAMSHIRE
NW OUTER LONDON
OXFORDSHIRE
1
1
8
4
23
34
71
4
1
5
2
4
5
1
2
7
6
11
10
8
7
8
12
23
6
34
26
7
68
44
21
62
29
61
84
51
31
60
29
53
179
137
59
167
103
137
S INNER LONDON
SCILLY ISLES
SE OUTER LONDON
SHROPSHIRE
SOMERSET
SOUTH GT.
MANCHESTER
SOUTH OUTER
LONDON
SOUTH TYNE AND
WEAR
SOUTH YORKSHIRE –
NORTH
SOUTH YORKSHIRE –
SOUTH
STAFFORDSHIRE
SUFFOLK
SURREY
SW OUTER LONDON
WARWICKSHIRE
WEST MERSEYSIDE
WEST MIDLANDS EAST
WEST MIDLANDS WEST
WEST SUSSEX
WEST YORKSHIRE EAST
WEST YORKSHIRE WEST
WILTSHIRE
WORCESTERSHIRE
Grand Total
5
15
1
4
1
7
17
1
1
4
13
31
25
10
12
47
10
25
33
65
81
9
2
30
56
45
46
68
2
90
100
127
209
2
3
5
35
19
31
95
1
5
11
14
37
45
113
3
4
8
6
36
32
89
5
1
4
21
34
43
108
5
12
9
5
8
3
1
8
7
8
4
10
10
8
8
3
6
14
26
7
42
25
11
18
41
77
44
67
25
36
38
53
79
56
76
36
55
40
76
200
118
213
110
118
102
196
3
1
3
33
50
44
134
1
8
4
11
9
11
25
31
49
50
63
56
151
167
5
-
3
22
58
64
152
6
231
4
3
230
2
1
448
12
23
1,313
55
39
2,867
56
56
3,358
129
128
8,447
TRAFFIC CENSUS OPERATIONAL REQUIREMENTS
27
Evaluation criteria for DfT Traffic Count 2005-2006 tenders
I.
Price
II.
Quality of enumerators
A. Recruitment policy
B. Training;
C. Experience of staff carrying out the counts.
III. Quality of data collected on-site
A. Number of enumerators employed during the day on each type of site;
B. Shift patterns employed;
C. Supervision (number and timing of visits by supervisors);
D. Site safety procedures.
IV. Quality of final data
A. Validation procedures at headquarters;
B. Timeliness of sending data to DfT;
C. Timeliness of sending invoices to DfT.
V.
Comprehension of requirement
A. Satisfying requirements;
B. Showing good understanding of what is required and how the service can best be
delivered.
28
Annex 8.3
TRANSPORT STATISTICS ROADS DIVISION
WORKS REQUIREMENT DOCUMENT
FOR THE MAINTENANCE OF THE OUTSTATIONS OF THE
AUTOMATIC TRAFFIC DATA COLLECTION (ATDC) SYSTEM
(PPAD 9/63/68)
Crown Copyright
29
CONTENTS
Section
No
Page
1.
INTRODUCTION
4
2.
CONTRACT PHILOSOPHY
4
3.
SITE LOCATIONS
5
4.
EQUIPMENT TO BE MAINTAINED
6
5.
SCOPE OF CONTRACT
8
6.
ROUTINE MAINTENANCE
12
7.
UNSCHEDULED MAINTENANCE AND ADDITIONS TO THE CONTRACT 13
8.
GOOD DATA
16
9.
PROCEDURES
17
10.
CONTRACTORS PERFORMANCE
18
11.
DOCUMENTATION AND TRAINING
22
12.
MEETINGS
23
13.
TRAFFIC MANAGEMENT
23
14.
SAFETY MEASURES
24
15.
CONTRACTORS VEHICLES
25
16.
PAYMENT TERMS
25
30
APPENDICES
APPENDIX A
Table of Site Details
APPENDIX B
Site Map
APPENDIX C
List of Specialist Tools
APPENDIX D
List of Spares
APPENDIX E
Checks and Works during Routine Site Visits
APPENDIX F
Replacement of Axle Sensors
APPENDIX G
Electronic Documentation Procedure
APPENDIX H
Routine Site Safety Inspection Pro-forma
APPENDIX I
Site Fault Visit Pro-forma
SCHEDULES
2.3 SCHEDULE A Bills of Quantities (fixed costs)
2.4 SCHEDULE B Bills of Quantities (additional costs)
2.5 SCHEDULE C Bills of Quantities (materials)
2.6 SCHEDULE D Bills of Quantities (equipment mark up)
2.7 SCHEDULE E Bills of Quantities (engineers rates)
2.8
Bills of Quantities (traffic management arrangement fee)
2.9 SCHEDULE F Depot Locations and CVs
2.10
31
1.
INTRODUCTION
1.1
This document details the maintenance requirements for the outstations of the
“Automatic Traffic Data Collection (ATDC) System”.
1.2
The system comprises 183 core census classifiers (outstations) and 5 weighin-motion classifiers at 155 locations throughout Great Britain, linked via telecom
lines to an instation located in London. At most motorway and some major road
locations there are separate outstations covering each direction. Each outstation
contains a single automatic classifier and associated equipment capable of collecting
data. Of the 183 core census classifiers, 10 additionally collect data for weigh-inmotion purposes.
1.3
The contract shall be for the maintenance of the outstations, which are
described in section 4.
1.4
The services required are detailed in sections 5 to 7 inclusive.
1.5
The site locations are described in section 3 and are given in Appendix A.
1.6
The evaluation criteria the Department will use to determine the successful
tender are as follows:
Proposed project maintenance procedures
Quality of the bid and understanding of the works requirement document
Technical competence and expertise of staff
Proposed resources to carry out the work
Price
1.7
The Department will evaluate the tenders on an 80 - 20 ratio in favour of
overall quality and project management ability against price.
2.
CONTRACT PHILOSOPHY. (The contained requirements form part of the
Works Requirements).
2.1
The object of the Automatic Traffic Data Collection System (ATDC)
Outstation Maintenance Contract is the continual production of accurate
classified traffic data by all the outstation sites. The amount and quality of the
data gathered is dependent upon the ATDC outstation sites being maintained to a
high standard. The Department for Transport (The Department) wishes to pass the
responsibility for the maintenance of the ATDC sites to a Contractor.
2.2
The Works Requirement Document contains information on the levels of
maintenance performance, which the Contractor is required to achieve. The
Contractor will provide performance records, which will be periodically reviewed
against the contract requirements. There will be a major review after the first nine
months and yearly thereafter.
32
2.3
It will be for the Contractor to determine, and employ, sufficient dedicated staff
for the Contractor to meet the demands of the contract. As a minimum the
contractor should have a project manager who is solely dedicated to this contract,
who can also be an engineer, but does not have to be. Initially there should also be
a minimum of three dedicated engineers assigned to the project, excluding the
project manager. The Department require tenderers to quote for the number of
dedicated and trained staff they intend to employ in their bids.
2.4
The contract will be charged according to the B of Q (Bills of Quantities) rates
in Schedules A to F for all the maintenance (including wear & tear, fault and
preventative maintenance) and also includes all associated costs with the exception
that the cycle and axle sensor costs, over and above the provision included in the
contract, will be reimbursed in accordance with Supplementary Condition of Contract
56 (Variation of Price). The contract prices will be fixed for the first two years and
indexed for the remaining three years.
The Department wishes to stress the importance of documentation throughout the term
of this contract. The Contractor is expected to record details of all work, including
serial numbers of items used and provide these to the Department. Details of the
documentation required are given in section 10.2.
3.
SITE LOCATIONS
3.1
The sites are located on all types of roads. Breakdown for the number of sites
for each class of road is indicated below:
ROAD TYPE
Motorways
Total No. of
Site Locations
31
WIM Motorways*
Class A Roads
58
2
67
WIM Class A Roads*
Class B Roads
Total No.
of Classifiers
68
3
26
26
33
Other Minor Roads
31
31
Total
155
188
Note: those sites marked *, the locations are adjacent to an ATDC site.
A table giving further site details is included in Appendix A and a map is included in
Appendix B.
3.2
At each site there is electronic equipment housed in a roadside cabinet linked
to sensors in the carriageway. A description of the site hardware is included in
section 4.
3.3
Hut Locations
At 6 of the sites, the roadside cabinet is installed inside a wooden hut. The huts are
located as follows:
Melmerby, Sparkford, Weston, Merthyr Tydfil, Prickwillow and North Nibley.
4
EQUIPMENT TO BE MAINTAINED
4.1
Each outstation site has a set of roadside equipment and a set of inroad
equipment.
4.2
The roadside equipment comprises:
a)
A roadside cabinet
b)
A classifier
c)
A mains supply and distribution board
d)
A battery
e)
A heater and thermostat
f)
An earth spike and strap
g)
A modem and telephone line
h)
Up to 3 cabling looms
i)
At motorway sites the telecom line and electricity supply are fed from
intermediate 609 cabinets in the fence line. These cabinets also form part of the
outstation
j)
At six sites the equipment is located within a wooden Hut.
4.3
The roadside equipment is linked to the second category of equipment (inroad) installed at each site.
34
In each lane there is:
a.
One inductive loop, which is wound in two halves, shaped similar to two “D”s
back to back.
b.
The vast majority of sites are equipped with non-removable class 2 piezo
electric axle weighing sensors.
c.
At most non-motorway sites, there are also two non-removable cycle piezo
sensors fitted in the near side lanes.
d.
A number of minor road sites, are currently equipped with two axle sensor
housings, which are fixed using an epoxy resin into the road surface along with two
removable piezo electric axle sensors, which fit into the housings. These are being
currently used instead of the items specified in b.) above.
e.
At 13 sites there is an additional array of weigh-in-motion sensors and a single
2m square inductive loop in one lane.
f.
At two sites there is an additional array of weigh-in-motion sensors and a single
2m square inductive loop in two lanes.
There is ducting and cabling between the sensors/loops and the roadside cabinet
and on motorways also between the roadside cabinet and the 609 cabinets which
also forms part of each outstation.
This equipment is subject to wear and the axle sensors have a finite life dependent
upon the traffic experienced. A gradual migration from the items at d. above to
item b. will be completed over the period of the contract.
Because of the harsh environment the in-road equipment is also more subject to
premature failure than normal electronic components. However, it is expected that
less failure will occur as the migration process, referred to above, is completed.
4.4
Any of the equipment detailed in 4.2 to 4.3 may also require maintenance
because of loss or damage caused by third party actions, accidents, road repairs,
public utility excavations etc.
4.5
At the 6 sites, where the roadside cabinet is installed inside a wooden hut, the
huts are to be included as part of the equipment to be maintained in a secure
manner.
4.6
Different materials have been used at some of the sites, e.g. ducting is usually
made of a plastic material but sometimes metal tube is used. Armoured cable is
used at some sites for loop cabling and/or power cabling dependent upon the nature
of the site.
4.7
Steps have been taken to maintain the equipment in good order but these are
not exhaustive and the equipment to be maintained may have defects, which are not
35
known to the Department at the commencement of the contract. It shall be the
responsibility of the Contractor to correct such defects.
5.
SCOPE OF CONTRACT.
5.1
As part of the fixed price element of the contract the maintenance Contractor
shall fully maintain all the outstation sites within the area of the contract (however
this number may be subject to additions or reductions). This maintenance shall
include the engineering routine service and safety site inspection (RSI) visits, with
each site being inspected and serviced twice a year.
5.2
Additions to the Contract.
The Contractor, if required in writing by the Department, shall: 5.2.1
install and commission modifications.
5.2.2
reinstall equipment at an outstation site.
5.2.3
remove an outstation(s).
5.2.4
relocate an outstation(s).
5.2.5
install and commission a new outstation(s).
5.2.6 repair damage to an outstation site(s).
5.2.7 amend existing outstation(s) either by addition or removal (or both) of
equipment.
The Contractor may also be requested, by the Department, to assist with trials of
equipment or maintenance facilities.
For all additional works the costs shall be in accordance with those specified in
Schedule B of the B of Q.
5.3
The Contractor shall supply from his own resources or through his
subcontractor all necessary, labour, materials, equipment, tools, transport,
subsistence and test and monitoring equipment for the rapid restoration of service
and the repair of faulty units.
5.4
All equipment and spares supplied, by the Contractor, for this contract shall
be to the latest version, available at the time of purchase.
5.5
The Contractor shall keep an itemised record of all equipment, including
location, reference numbers, serial numbers and type, used on the contract both in
36
storage and where installed at outstations. A copy of this list is to be provided to the
Department on a quarterly basis after the commencement of the contract.
5.6
The Contractor shall provide a project engineer and a team of site engineers.
This team and any specialist or replacement staff must be technically competent and
appropriately trained. All engineering staff must be willing to work unsociable hours,
e.g. nights and weekends, if required to do so. The charges for work irrespective of
the time of day and day of week will be specified in Schedule B of the B of Q.
5.7
Tenderers shall complete the appropriate table in Schedule F, which gives
details of depot location(s).
5.8
The costs quoted in Schedule E, for engineers, shall include all overheads
and back-up support for the engineers.
5.9
The Contractor shall supply the Department with a telephone number, which
shall receive calls, to be called at any time for fault reporting, the answering of which
shall be considered as being the time of reporting of the fault.
5.9.1 Telephone automatic answering and recording systems may be used
providing the Contractor accepts the time of recorded messages as that
stated by the Department.
5.9.2 The Contractor shall give the Department at least 14 days notice of any
change to the telephone number supplied, including temporary
numbers if required for use during public holidays.
5.9.3 The Contractor and its engineering team shall have the facility to send
and receive electronic mail.
5.10 The Department shall normally report faults to the Contractor. They may also
be identified by the Contractor during RSI visits or when visiting a site because of
another reported fault. However, if the Contractor identifies a separate fault at an
outstation this must also be recorded as a fault.
5.11 When the Contractor is notified of an outstation fault then the outstation will be
deemed not to be providing “good data” (see section 8) from and including the day of
the notification. If the fault is such that loss or corruption of the classifier data has
not occurred and it is subsequently recovered by the in-station (after the fault has
been repaired) then no loss of “good data” will be deemed to have occurred.
5.12
Response to Calls
5.12.1
The Contractor shall respond to calls, reporting outstation faults, as
described below and shall liaise directly with any public utility operator, Highway
Authority ( or their agents) and the Department as appropriate.
37
5.12.2
Where a fault or damage to equipment is causing a potentially
dangerous situation the Contractor shall attend on site within 8 hours of
notification to make the site safe.
5.12.3
Subject to paragraph 5.12.2 the Contractor shall attend on site, within 2
working days of the notification for a minimum of 90% of reported site faults, and
achieve 100% attendance within 3 working days. The percentages shall apply
over a period of a calendar month and for the annual periods. The Department
may relax these criteria for a few of the more remote sites where exceptional
circumstances prevent the Contractor attending within the above time scale.
5.13 When attending to a reported fault, the contractor will repair all off-road faults
during the first visit (with the exception of B.T./Power faults that are diagnosed as
being the responsibility of a third party.) The site engineer must carry equipment
tools and materials, not only relating to the nature of the call but also to enable
repairs to other replaceable equipment to be undertaken. The Contractor shall
provide, at his own expense, a car/portable telephone, portable computer and
portable printer for each site engineer and for the project engineer when they are
engaged in site visits. If it is found a fault was not repaired when an engineer
attended site because he/she was not carrying sufficient spares for the task, no
payment will be made by the Department for any subsequent visit required to rectify
the fault.
5.13.1
All test equipment used for maintenance and measurement and setting
up shall be re-calibrated against traceable standards at periods not exceeding 1
year. The dates of the last and/or the next re-calibrations shall be clearly marked
on the equipment.
5.14 Upon completing any work the Contractor shall leave the site in a safe and
tidy state, this shall include as applicable:
a. Sweeping all debris from the carriageway.
b. Levelling and firming any ground that was dug up, replacing slab paving and
making good any damage to the site area.
c. The collection and disposal, in the proper manner, of all waste material and
debris.
5.15 The telecom line provided to each outstation shall only be used by the
Contractor, or his subcontractor(s), for e-mailing information to the Contractors
office, the Department, testing or in an emergency.
5.16 All fault visits determined as "No fault found" calls will normally be paid at the
standard attendance rate.
5.16.1 No additional payment will be made for a call for service if no fault is found
and the Contractor will be expected to visit the same outstation as normal if
subsequent faults are reported. However, if no fault is again found (when the
second or subsequent visit is requested within 28 days following the previous call)
38
then the Contractor shall be entitled to a compensation payment. This will be made
as per the price schedule rates for travelling time, time on site, subsistence and
mileage, for the second and subsequent visits. Any other consequential costs will
not be reclaimable.
5.16.2 The Department may at any time instruct the Contractor not to attend on a
particular site(s).
5.16.3 If a fault, reported at the time as “No Fault Found”, is later (within 3 months of
the initial report) determined as a valid fault, then any claims for additional payment
(paragraph 5.16.1) shall be cancelled and any payments made shall be refunded to
the Department.
5.17 The maintenance charges for the outstation(s) shall be directly proportional to
the time the outstation(s) is under the maintenance and at the rates included in the B
of Q.
6.
ROUTINE MAINTENANCE (Fixed Price Contract - see also Schedule A).
6.1
For a single fixed price (detailed in Schedule A) the Contractor shall be
responsible for all routine maintenance of outstation equipment, including in addition
the supply and replacement of all faulty components.
6.2
Routine maintenance of the outstation equipment shall entail: -
a.
Undertaking RSI visits; each outstation shall be visited twice a year, and
carrying out the work detailed in Appendix E. An allowance (see paragraph 10.4.2.1)
of one day will be made, for each outstation, for the loss of “good data” consequent
upon each routine site visit. RSI visits should still be carried out even if a site has a
long-term fault to check on the safety of the equipment and site. Replacement of
sensors is not part of the fixed element of the contract and will need to be invoiced
separately. At some sites, where it is possible to do so safely, replacement of faulty
sensors should be carried out during the RSI visit.
b.
RSI visits shall be carried out at regular intervals and the contractor shall
prepare a schedule. This shall be presented to the Department within two weeks of
the commencement of the contract. No RSI visits will be carried out during the
months of August and February and the contractor shall complete a minimum of 30
visits per month and no more than 50 per month. No site shall receive an RSI within
4 months of its previous RSI.
c.
The scheduled replacement of classifier batteries as detailed in Appendix E.
d
Any small repairs, which can be done during the RSI, visit should be
conducted. The Department will reimburse the cost of any materials used
under the B of Q rates, along with a labour charge for the additional time
required to undertake the repairs, in accordance with the relevant rates quoted
in B of Q Schedule E.
6.3
Prior to every RSI visit to each outstation the Contractor shall contact the
appropriate Highway Authority, and enquire if any road works are planned which
39
might affect the in-road equipment at the outstation. The Contractor shall also ask to
be kept informed about any road works, which will affect the in-road equipment. The
Contractor shall inform the Department about such road works.
6.4
The Contractor shall maintain the equipment, including spares and test
equipment in good working order. He shall ensure that the level of performance and
condition of the equipment is, as a minimum, not less than the level of performance
and condition of the equipment at the commencement of the contract, having due
regard to its current age and expected life span; (see also 7.1.1.4).
6.5
The Contractor shall provide facilities so that all classifiers, replacement
boards and modems can be fully functionally checked at the depot.
6.6
Spares and Storage Facilities
6.6.1 The Department will make available an initial quantity of free issue spares at
the start of the contract. Appendix D lists all the spares that will be provided. An
equivalent quantity of spares to those free issued at the commencement of the
contract shall be returned to the Department at the completion of the contract.
6.6.2 The Contractor shall, at all times, ensure that sufficient spares are available to
fully meet the maintenance requirements; at no time shall the stock level of those
spares which were free issued fall to less than half the initial stock level prior to reordering.
6.6.3 The Contractor shall provide secure indoor storage facilities for the spares. A
minimum floor space of 6m by 4m will be required with sufficient racking to a height
of 2m along the complete length of one of the 6m sides to accommodate all the
sensors.
6.6.4 Items returned as faulty or suspect shall be so labelled and stored separately
from other spares items.
6.6.5 Faulty equipment identified under paragraph 6.6.4 may be repaired or
replaced on application to the Department. On request the Department will raise an
order number for the defective items either to be repaired or replaced at the B of Q
rate in Schedule C.
7.
UNSCHEDULED MAINTENANCE AND ADDITIONS TO THE CONTRACT
7.1 The following items are excluded from the fixed price of the contract and are
chargeable according to the costs quoted in B of Q Schedule B. Costs quoted shall
include labour, civil works, travel and subsistence and other incidentals.
7.1.1 Unscheduled maintenance - includes the repair of all faults occurring on any
outstation equipment, followed by re-commissioning. The contractor’s attention is
drawn to appendix F that gives details relating to sensor replacement.
40
7.1.1.1 Where practicable faults shall be identified and immediately cleared on site
by the replacement of printed circuit boards, sub-units or assemblies and if relevant
shall be chargeable under the costs quoted in the B of Q Schedule B.
7.1.1.2
The Contractor shall ensure that defective units are returned to the
depot and are promptly repaired or, if irreparable, replaced with new equipment.
Equipment covered by the B of Q will be re-charged to the Department
7.1.1.3
The Contractor himself may carry out, repairs to defective units
removed from site, wholly or in part, provided that proper facilities are available and
that the quality of the repair is equivalent to that which would be provided by the
original manufacturer. Repaired items shall be separately inspected and marked
prior to being made available for spares usage. If proper repair facilities are not
available the units shall be returned to the original manufacturer for repair.
7.1.1.4
Digitised format photographs should be used to record and advise the
Department of the specific nature of problems found during site visits.
7.1.2 When the Contractor has advance information that an outstation will be
damaged by road works he shall take measures to protect the in-road equipment. In
cases when the equipment cannot be protected he shall take appropriate action to
recover salvageable equipment before the road works commence.
7.2
Upon satisfactory completion of any additional works the equipment installed
shall, at the request of the Department, become part of the system to be maintained
and be included in the fixed price maintenance contract and a variation to the fixed
price portion of the maintenance costs shall be made.
The amount of the variation shall be either as quoted in the bills of quantities, or if
the work was carried out by a third party, the variation shall not exceed 10% of the
total cost of the equipment which has been installed, less the current cost of any
removed equipment.
7.3 Removal/Relocation of an Outstation /Equipment
7.3.1
The Contractor may be required to relocate an outstation either wholly
or in part. This involves the removal of equipment from site and perhaps placing
the equipment into storage, then installing the equipment at a different site. The
opportunity may be taken when installing the relocated equipment to change the
configuration of the equipment.
7.3.2
When equipment is removed, this shall be carried out carefully and,
where it is economical, the equipment shall be salvaged for future use. All
equipment shall be functionally checked before being reinstalled.
7.3.3
Relocation shall not be deemed to have occurred where the in-road
equipment and any associated ducting is reinstalled in a different position at the
same geographical outstation site location.
7.4 Install New Outstation Site(s)
7.4.1
The Contractor may be required to install and commission, wholly or in
part, additional outstation sites.
41
7.4.2
It will be the duty of the Contractor, when asked to install a new site, to
arrange for all the BT (British Telecom) and power connections to the site.
7.4.3
The Department will pay the Contractor a sum of £1,000 to cover for
the administrative costs involved when arranging for the installation of a new site.
This sum will be payable upon successful commissioning of the site.
7.4.4
With reference to paragraph 7.3.3 above, a new site will be deemed to
have been installed if a re-location is more than 100 metres from the existing
position of the site.
7.5 Repair of Damage to an Outstation
7.5.1 Where accidental damage has occurred the Contractor shall: a. Make the site safe from dangerous electrical potentials, and physical hazards.
The Contractor shall inform the Department of his actions and when the site is
safe.
b. Collect damaged equipment and retain pending instructions on salvage or
disposal from the Department. The Contractor shall advise the Department that
he is retaining damaged equipment.
c. The Contractor shall repair the damaged site. If third party damage results in
the loss of “good data”, from the damaged outstation, the site will be deemed not
to be producing “good data” only after the third working day following the
Contractor becoming aware of the damage. Other relevant allowances for the
suspension of loss of “good data” given elsewhere in the contract documents
shall also apply.
d. After the restoration of the site equipment the Contractor shall re-commission
the site, by the end of the next working day.
7.5.2
Where damage has been caused by a Third Party, the Contractor shall
take all reasonable steps to recover the cost of rectifying the damage from the
Third Party.
7.6 Amendments to Existing Site(s)/Equipment
7.6.1 The Contractor may be required to carry out alterations to an outstation
site(s). The Contractor may also be asked to assist with trials. B of Q rates shall
apply where relevant.
7.7
Payments for un-scheduled maintenance and fault rectification.
7.7.1 Payments for response to notified faults will be in accordance with B and Q
quoted costs in Schedule E. A fixed payment will be made for attendance at a
fault with further fixed payments being made for any subsequent visits required
to the site for the completion of the fault repair i.e. attendance at required Traffic
Management. The fixed payment will be at a reduced rate for attendance in
excess of 3 working days as B of Q costs in Schedule E. Additional fault
rectification off site (which will require prior approval by the Department) will be
billed in accordance with B of Q rates in Schedule E.
42
7.7.2 Payments for the rectification of notified faults will be in accordance with
the B of Q equipment costs in Schedule B. These costs are for the
Removal/Install/Commissioning of complete units, the rectification of a fault by
the replacement of an outstation discrete component or printed circuit board
shall be covered by the fixed payment fault response charge. If the
removal/replacement of a discrete component or printed circuit board requires
the site to be re-commissioned this will be chargeable at the rate quoted in the B
of Q costs in Schedule B.
7.7.3 A single payment will be made at the lower agreed visit rate if more than
one fault is being attended to during a single fault visit.
7.8
Additional works, which are not covered by the B of Q rates shall be identified by
the Contractor and notified to the Department. The Department will require the
Contractor to submit relevant costings, which may require a variation in the contract to
implement such works
7.9
It is important that the Department is made immediately aware when a loop
related fault has been repaired. This is irrespective of whether it is the loop or the loop
board. Any delay in informing the Department will count against the contractor.
8
GOOD DATA
8.1 The purpose of the ATDC outstation system is to provide continuous “good
data” from each of the outstations to the in-station. The Department will advise
the maintenance contractor of new sites, which have been successfully
commissioned and ask them to be included in the contract (as per paragraph 7.2)
8.2 The data from each outstation is collected each night by the in-station and
inspected for fault messages and for changes in the pattern of the traffic data that
also indicate faults at the outstation. It is upon this information that the
Department may report a fault to the contractor.
8.3 The criteria used for the inspection of the data from the outstations, which will
be used to assess if the data is "good data" or to initiate a fault report to the
Contractor(s), are as follows:
For each outstation, for each day, midnight to midnight.
8.3.1
The number of hours in which the proportion of High N's (merely
classification by axle count) exceed 5% of the total vehicle count, shall not be
more than 2 in any 24hr period.
8.3.2
There shall be no negative values in any of the data.
8.3.3
The total for all motor vehicles for each day shall be greater than zero.
43
8.3.4
In any 24 hour batch of data there shall be no more than 2 hours of
corrupted or missing data.
8.4 Delays may occur before traffic management can be implemented. Where
the delay is outside the contractor’s control (see paragraph 10.4.2) then the site
will be deemed to be producing “good data” during the period of that delay.
8.5 The Contractor shall ensure that every classifier provides 350 days of "good
data" in each contract year, subject to the allowances provided for in paragraph
10.4.2.
9
PROCEDURES
9.1 All work shall be carried out according to procedures laid down in
manufacturer’s handbooks and the documentation supplied by the Department.
Any deviation from these procedures shall be agreed with, and demonstrated to,
the Department prior to being implemented.
9.2 Upon arriving at a faulty outstation site the site engineer shall determine
precisely the nature of the fault and take the appropriate action as indicated
below: 9.2.1
Clear any problems/faults associated with the roadside cabinet
equipment.
9.2.2
Make any traffic management arrangements, which are required.
9.2.3
If traffic management can be arranged without undue delay then the
engineer shall remain on site and complete the repairs to the in-road equipment.
9.2.4
Report time of arrival on site, findings, action taken, any further actions
which are required, and when work is completed or the time when he departs
from site.
9.2.5
Notify any third party of faults, which are their responsibility and make
arrangements for their rectification.
Note:
Arrangements may be made either directly by the site engineer or his
depot staff.
10.
CONTRACTOR’S PERFORMANCE
10.1
Quality of the Contractor’s Work
44
10.1.1
The quality of the work done by the Contractor, or his subcontractors,
for both maintenance and additional works shall be to a good standard of
engineering conforming to current techniques and practice unless otherwise
prevented by the need to maintain compatibility with the existing practice. Work
undertaken by the Contractor, or his subcontractors, shall conform to the
requirements of the Health and Safety at Work Act 1974.
10.2
Reports required from the Contractor
i) Quarterly itemised record of equipment, including location, reference numbers
and type, used on the contract both in storage and where installed at outstations.
(see paragraph 5.5)
ii) Monthly report to be provided to the Department showing level of performance
(see paragraph 10.4.)
iii) Monthly update of the RSI schedule, listing sites where work is planned as
well as completed. (see paragraph 10.6.1)
iv) Weekly traffic management schedule (see paragraph 13.4)
v) Notification of action which clears open faults on a daily basis (see paragraph
10.6.5)
vi) Notification when a repair, overhaul or modification has exceeded two weeks
and report at weekly intervals thereafter until the work is complete. (see
paragraph 10.7)
vii) With invoices copies of supporting documentation must be provided before
payment will be authorised (see paragraph 16.1.1)
10.3
Fault Reporting and Recording.
10.3.1 The Contractor shall set up a fault reporting and recording system, which
shall be agreed with the Department within one month of the contract
commencing and be fully operational within two months. One of the main tasks
of the system will be to provide information on the operation and performance of
the maintenance contract.
10.3.2 The fault reporting procedure shall include the following information:
i)
Fault number
ii)
Date fault raised
iii)
Site number/location
iv)
Nature of fault
v)
Date of each visit and all actions taken during the visit
vi)
Date(s) and Duration of any 3rd party delays
vii)
Date(s) of any traffic management(s) request(s) and date(s) when
traffic management(s) occurred
45
10.4
viii)
Dates information about visits were e-mailed to or from the Department
ix)
Date fault was cleared by the engineer and date fault was confirmed
cleared by the Department
Performance Criteria.
10.4.1 When assessing the contractor’s performance the Department will
consider the following areas:
i) Safety of both personnel and equipment
ii) The organisation of remedial work
iii) Fault diagnosis capability and accuracy
iv) Site set up accuracy
v) Refurbishment/repair of site equipment and test equipment
vi) Quality of repair work both electrical and civil
vii) Traffic management organisation and implementation
viii) Project management and reporting standard
ix) Financial control and invoicing accuracy
x) Spares provision and availability
xi) Work done during preventative and routine maintenance site visits
xii) The provision and quantity of "good data".
10.4.2 Allowances will be made for traffic management delays, accidental damage, RSI
visits, Departmental checks and third party works. The Department will agree to a lowering
of the requirements to take into account:
10.4.2.1 The number of days between the date of traffic management and the
date of access to the outstation provided that this is not excessive (traffic
management should take no more than 3 weeks for a non-motorway site and
6 weeks for a motorway. If in some circumstances traffic management takes
longer the Department will require copies of relevant correspondence before
an allowance is made).
10.4.2.2 Third party delays caused by power or telecom companies or
emergency repairs to the road or utilities embedded in the road at an
outstation may be permitted provided evidence is supplied that the relevant
parties have been pursued.
10.4.2.3 Delays requested by the Department for operational reasons.
10.4.3 The Department may agree to a lowering of these requirements for a particular
outstation. The Contractor shall forward records, to the Department, giving a report to
establish the level of performance on a month by month and cumulative basis throughout the
46
term of the contract. This report should be provided monthly in accordance with the
conditions set out in paragraph 10.4.8.
10.4.4
Site visit response times.
The requirement is that an engineer shall attend the outstation within two working
days of the fault being known to the Contractor for 90 percent of the faults and
within three days for 100 percent of the faults. Note that 10.4.3 may impose a
much faster response than the maximum allowed if outstation data is being lost.
10.4.5
Time to repair.
Each fault shall be logged separately, commencing at the time when the
Contractor knows the fault and ending when the outstation fault is cleared and the
Department confirm the date that it is producing “good data” again. The time shall
include any delays while awaiting traffic management but these delays shall also
be recorded separately.
10.4.6
Logging of faults and fault clearances.
The Contractor shall log the times when he is advised or when his staff are aware
of faults together with the nature of the fault(s) both as advised and as
subsequently determined. The Contractor shall also log when his engineer
arrives on site, when the engineer reports that the fault has been cleared and or
whether further action (e.g. traffic management) is required and what that action
is. The Contractor shall record when further action is due, when it actually
commences and finishes.
10.4.7
Site visits and “No fault found” records.
The Contractor shall record, on a site by site basis, the number and time of
separate visits to each outstation from the time when the Contractor was aware of
a reported fault up to and including the time when the outstation was again
producing good data, and also record the number (and time) of “no fault found”
visits made to each outstation site.
10.4.8 Documentation
All documentation, photographs and test results generated by the Contractors
engineers on site, and which is contractually required to be produced, will be
e-mailed to the Department. Those e-mails shall meet the following criteria:
90% should be received within 2 working days of generation; and
100% will be received within 3 working days of generation.
Note: Day 1 will be the first working day after the day of the visit.
10.4.9 The performance records, of paragraph 10.4, shall be forwarded
monthly to arrive at the Department 5 working days before the monthly
meeting (see paragraph 12.1) or not later than 10 days after the end of the
month which ever is the earlier.
10.5 The Departments’ representatives will make, from time to time, sample site
visits to check on the site related items in 10.4.
47
10.6
Additional Information to be supplied to the Department.
10.6.1
Within 2 weeks of the commencement of the contract the Contractor
shall forward, to the Department, for approval, a schedule detailing when the
routine maintenance visits will be carried out.
The schedule shall itemise the work on a weekly basis.
Once the contract is let the Contractor shall forward to the Department a monthly
update of the schedule, showing work completed and new work planned.
10.6.2
The Contractor shall set up an information reporting system, details of
which shall be presented and agreed with the Department within a month of the
commencement of the contract. The system shall be based around ISO9001 or
equivalent.
10.6.3 The Contractor shall provide the Department with the following information
each month in accordance with the conditions of paragraph 10.4.9:
a) A record of the outstations where a RSI visit has been made. The
Contractor shall retain a record of the checks and adjustments made
together with details of calibration values. A pro-forma will be used; the
layout of which will be agreed between the Department and the contractor
prior to the contract commencing. (see appendix F) This pro-forma will
include spaces to list other work done and other work necessary but not to
be done at the time. The pro-forma will be e-mailed to the Department
after the completion of every visit
10.6.4
The contractors depot information system shall also include a method
of recording the modification status of each outstation. The modification state
shall also be recorded on the outside of the equipment affected using a suitable
label supplied by the Contractor. During periods when modifications are actually
being carried out the information shall be updated on a weekly basis.
10.6.5
When a fault has been cleared and the Contractor has restored the
outstation to a fully operational state he shall immediately inform the Department
by fax and request confirmation that “good data” is being received by the
Department before the fault is fully cleared. The details to be supplied as per
10.3.2 and the electronic documentation procedure agreed between the
contractor and the Department.
10.7
Delays
10.7.1
The Contractor shall inform the Department when a repair, overhaul or
modification has exceeded two weeks and report at weekly intervals thereafter
until the work is complete.
11
DOCUMENTATION AND TRAINING
11.1
The Department will provide the Contractor with the following information:
48
11.1.1 At the commencement of the contract the Department will provide the
Contractor with a complete set of technical documentation as provided by the
manufacturers of the electronic equipment.
11.1.2
The Department will also provide the Contractor with documentation covering the installation and
commissioning of outstations. This documentation is provided on the basis of “best advice based upon Departmental
experience”. The contractor must follow these instructions.
11.1.3
The Department will provide 5 copies of its maintenance software package, which will be installed on each
engineer's laptop. The usage of this software is outlined in Appendix E.
11.1.4
Prior to the commencement of the contract the Department may hold a
3-day training course subject to staff availability.
11.1.5
Any training courses will be based in Bristol with some visits, for the
engineers, to outstation sites.
The Contractor shall provide transport for his staff.
No payments shall be paid by either party to the other party for the costs involved
in the training.
11.2
All documentation will be provided in good faith but the accuracy and
adequacy cannot be guaranteed.
Clauses 11.1.1 to 11.1.5 shall not diminish in any way the responsibility of the
Contractor for the maintenance work.
The Contractor shall feed back information as to any documentation errors or
changes required in the documentation.
11.3
The Contractor shall keep all documentation provided the Department
in a clean condition may take working copies of the original Applied Traffic
manuals as well as the Department documentation.
11.4
At the end of the contract all original documentation, which shall be
complete, together with any copies shall be handed back to the Department.
12
MEETINGS
12.1
Progress meetings will be held on a monthly basis. The venue for the
meetings will be negotiated at the commencement of the contract.
Representatives from the Contractor shall be the project manager and additional
staff if deemed necessary by either party.
49
13
TRAFFIC MANAGEMENT
13.1 The Contractor shall be responsible for arranging traffic management including
that on motorways and trunk roads. The Contractor shall ensure that when arranging
traffic management it fully complies with the requirements of the New Roads and Street
Works Act 1991. With larger repairs and where in-road work cannot be completed at a
location in one day, traffic management should normally be arranged on consecutive
days as one session to minimise delays. The costs for the Traffic Management shall
be billable to the Department at cost; the contractor will be able to bill the Department
for an arrangement fee at B of Q rates in Schedule B for each session in a given
location
13.2 The costs quoted for any of the work detailed in this document shall exclude any
associated traffic management costs. The Contractor will however be responsible for all
recurrent traffic management costs if the need for traffic management results from:
13.2.1 Work which could have been done during a previous traffic management,
at the same site, within the previous two months period,
13.2.2 The re-installation/replacement of equipment which proves faulty within two
months of its installation.
13.3 Delays may occur before traffic management can be implemented. Where the
delay is outside the Contractor's control then the outstation will be deemed to be
producing "Good Data" during the period of that delay. The outstation will be (again)
deemed not to be producing "Good Data" starting from and including the day traffic
management or contra-flow is implemented until the maintenance contractor completes
setting up the site. The site will then be deemed to be producing “good data” from the
date the setting up has been completed.
13.4
The contractor will provide the Department with a weekly update of the
traffic management schedule, which shall contain the following details:
i)
Date the traffic management order was placed
ii)
Date of the traffic management
iii)
Time of the traffic management
iv)
Fault number
v)
Site number
vi)
Location
vii)
Duration of the traffic management
viii)
Brief outline of the work to be carried out
ix)
Number and type of sensors required (if applicable)
x)
Name of the engineer assigned to do the work.
50
14
SAFETY MEASURES
14.1 The Contractor shall at all times ensure that all staff employed on site are
aware of and comply with, the Health and Safety at Work Act 1974 and with Chapter
8 of The Traffic Signs Manual.
14.2 The Contractor shall comply with all safety instructions issued to the
Contractor by the Department, the relevant Highway Authority or the police. These
instructions may be verbal or written.
14.2.1 Where compliance with an instruction results in work not being carried
out the Contractor shall notify the Department verbally by the end of
the next working day and in writing within a week.
14.3 The Contractors attention is particularly drawn to the hazards of working on
motorways. The recommendations contained in guidance documents issued by the
Department and its agencies shall be followed at all times.
14.4 The Contractor shall inform the Department of any aspects of the outstation
system that he considers does not conform to the legal requirements of the Health
and Safety at Work Act 1974.
The Contractor shall, if so required by the Department, carry out the necessary
modifications to meet the H&S Act. The cost of such work shall be subject to B of Q
rates and or CVO (contract variation order) rates if not otherwise covered.
14.5 At site 886, Banbury, the Contractors staff are required to park their
vehicles to the south of the cabinet and on the same side of the road.
15
CONTRACTOR’S VEHICLES
15.1 The Contractor shall supply each member of the dedicated team with a
vehicle that possesses all round visibility and shall be capable of carrying the
equipment required to enable the engineer to undertake the work of the contract.
The vehicle shall be finished in a high visibility colour (e.g. white or yellow).
15.2 Each vehicle shall be equipped with sufficient flashing amber beacon(s) (not
wired into the ignition system). The beacon(s) shall be clearly visible from the rear of
the vehicle when the tailgate is fully opened.
51
16
PAYMENT TERMS
16.1
Fixed Price
Payment for the fixed price maintenance shall be on a three monthly basis [Note:
The quarterly invoice will be a quarter of the annual fixed price provided the
RSI’s scheduled for that quarter have been completed. Invoices shall be
submitted for payment not later than two weeks after the period to which the
payment relates.
These invoices shall itemise costs of: 16.1.1
Those outstations where routine maintenance has been carried out
during the three month period.
16.2
Additional Work and Fault Visit Costs
16.2.1
Invoices shall be submitted within two weeks of the work being
completed subject to Departmental acceptance to TSR1, Zone 2/14, Gt. Minster
House, 76 Marsham St., London SW1P 4DR.
16.2.2
All invoices relating to additional work shall provide supporting
evidence of costs and full details of the work carried out including serial numbers
of equipment exchanged, labour, transport, equipment and material costs with
any mark up together with traffic management costs and any identification
numbers issued for this work where applicable. All work should be referenced to
a fault number and the electronic documentation detailing the work undertaken
during visits.
16.2.3
All invoices and supporting documentation shall also be copied to the
Departments engineering support section in Zone 2/15A E, Temple Quay House,
2 The Square, Temple Quay, Bristol BS1 6HA.
16.2.4
If the invoice relates to a contract variation order (CVO) item then the
CVO reference shall also be quoted.
16.2.5 If, when carrying out additional work, the work is spread over different
outstation sites and over a period exceeding three months then the Contractor
may request that payment be made in stages. Each stage invoice shall be for
a period of not less than three months and apply only to the outstation where
the work has been fully completed. The details of the work completed during
the period shall be detailed and supplied with the invoice.
52
Appendix A
SITE
ROAD
No.
272 M9
273 M8
700 U/C
701 C
702 U/C
703 B1377
704 U/C
705 U/C
706 B3215
707 A25
708 C
709 U/C
710 B6020
711 M90
712 A327
714 U/C
715 A389
716 C
717 A49
718 C
719 C
720 A3054
TYPE
M/W
M/W
U/N
C/N
U/N
B/N
U/B
C/B
B/N
P/B
C/B
U/B
B/B
M/W
P/N
U/N
P/B
C/B
T/B
C/N
C/N
P/B
ATDC OUTSTATION SITE LOCATIONS
LOCATION
BANNOCKBURN J8-J9
HARTHILL
LLANEGRYN, GWYNEDD
KEYHAM, LEICESTERSHIRE
HOLLOW LANE, RAMSEY, CAMBS
LONGNIDDRY, LOTHIAN
ST HELENS, MERSEYSIDE
BEESTON, NOTTS
OKEHAMPTON, DEVON
GUILDFORD, SURREY
OAKLEY, BEDFORDSHIRE
BENTLEY
RAINWORTH, NOTTINGHAMSHIRE
KELTY J4-J5
ARBORFIELD GARRISON
MILBORNE PORT, DORSET
BODMIN, CORNWALL
PENSILVA, CORNWALL
HEREFORD
EAST OF CLIBURN, NR PENRITH
LOVER, WILTSHIRE
ISLE OF WIGHT
MAP
GRID
LANES
REF.
57 28196881
4
65
4
124 26063058
2
141 46783063
1
142 52972847
1
66 45197663
2
108 35213982
2
129 45123368
2
191 25980958
2
186 49951508
4
153 50082532
2
111 45594061
2
120 45903580
2
58 31326945
4
175 47631660
2
183 36721194
2
200 20740670
2
201 22900700
1
149 35072394
4
91 36095242
2
184 42101198
2
196 45080904
2
DIR. 1 CYCLE
LANE
W
W
E
N
E
N
S
E
W
S
W
S
S
S
S
E
S
S
N
E
S
1
0
0
0
2
0
1
0
2
1
1
0
0
1
2
1
1
1
1
1
53
SITE
ROAD
No.
721 U/C
722 B7020
723 A75
724 A91
726 U/C
729 A110
730 A5268
731 A6044
732 A1074
733 A651
734 A6050
735 A4
736 A58
737 A610
738 A638
739 A151
740 A56
741 A43
742 A374
743 A73
744 A96
745 A30
821 C
822 C
823 C
TYPE
U/B
B/N
T/N
T/N
U/B
P/B
P/B
P/B
P/B
P/B
P/B
P/B
P/B
P/B
P/B
P/N
P/N
P/B
P/B
P/B
P/B
T/N
C/N
C/N
C/N
ATDC OUTSTATION SITE LOCATIONS
LOCATION
LIGHTWATER, SURREYHEATH
UPPER DORMANT, DUMFRIES
EAST RIGGS, DUMFRIES
MILNATHORT, TAYSIDE
BETTWS
CHINGFORD
CHESTER
PRESTWICH
NORWICH
GOMERSAL
LYDGATE, NR OLDHAM
READING
LEEDS
RIPLEY
BARNBYMOOR, NOTTINGHAMSHIRE
WHAPLODE, LINCOLNSHIRE
THORNTON IN CRAVEN
GEDDINGTON
PLYMOUTH
AIRDRIE
BRODIE
HONITON
KIRKBY, MALZEARD
1 KM SW OF HACKTHORN, LINCS
MAXWORTHY CROSS, CORNWALL
MAP
175
85
85
58
170
177
117
109
134
104
109
175
104
120
120
131
103
141
201
64
27
192
99
121
190
GRID
LANES
REF.
49291620
2
30985765
2
32005677
2
31507068
2
29021863
2
53801945
2
34023666
4
38114024
2
62103093
2
42054270
2
39744037
2
47201729
3
42824324
2
44193500
2
46693842
2
53373247
2
39174493
2
48912830
4
25040555
4
98065715
31541009
42114720
49833816
22720947
2
4
1
1
1
DIR. 1 CYCLE
LANE
E
2
S
W
S
N
1
E
1
N
0
E
0
E
1
N
0
S
0
W
0
E
0
S
0
N
0
E
1
N
0
S
0
N
1
W
W
W
E
S
0
0
1
1
1
54
SITE
ROAD
No.
826 C193
827 C
830 C
833 U/C
841 U/C
846 U/C
847 U/C
850 A259
852 A39
853 A303
856 B430
857 A49
858 A446
859 A61
860 A16
863 A1174
865 A596
868 A465
869 A483
870 A75
871 A952
872 A77
873 A82
874 A12
875 A66
TYPE
C/N
C/N
C/N
U/N
U/N
U/N
U/N
T/N
T/N
T/N
B/N
T/N
T/N
T/N
T/N
P/B
T/N
T/N
T/N
T/N
T/N
T/N
T/N
T/N
T/N
ATDC OUTSTATION SITE LOCATIONS
LOCATION
SAFFRON WALDEN, ESSEX
SALTWOOD, NR HYTHE, KENT
EUSTON, SUFFOLK
HILLSIDE, GRAMPIAN
BLUBBERHOUSES
PENMON, ANGLESEY
TARLAND, GRAMPIAN
BEXHILL, SUSSEX
FAIRY CROSS, NEAR BIDEFORD
SPARKFORD, NEAR WINCANTON,
MIDDLETON STONEY, OXON
WESTON, SHROPSHIRE
NEAR NEC, BIRMINGHAM
DRONFIELD, DERBYSHIRE
SWABY, LINCOLNSHIRE
WOODMANSEY, HUMBERSIDE
LEEGATE, CUMBRIA
NR MERTHYR TYDFIL, MID GLAM
NEAR WELSHPOOL, POWYS
CASTLE DOUGLAS, GALLOWAY
CRUDEN BAY, GRAMPIAN
KILMARNOCK, STRATHCLYDE
SPEAN BRIDGE, HIGHLAND
EAST BERGHOLT
TEMPLE SOWERBY
MAP
154
179
144
45
104
115
37
199
180
183
164
126
139
119
122
107
85
160
126
84
30
70
41
155
91
GRID
LANES
REF.
55112380
1
61481362
2
59242810
1
39187978
1
41524570
1
26213800
1
34728058
1
56971078
2
24041243
2
36431265
4
45342232
2
35603292
2
42012854
4
43503776
5
53873765
2
50594376
2
31995462
2
30062057
2
32433092
2
28005691
2
41008392
2
24136339
4
21777796
2
60712366
4
36205265
2
DIR. 1 CYCLE
LANE
S
1
S
1
N
1
N
0
S
1
N
1
S
0
W
0
W
2
W
0
S
0
N
0
N
1
S
0
S
0
S
1
W
1
W
0
S
1
W
0
W
0
S
0
W
0
E
0
E
0
55
SITE
ROAD
No.
876 A9
884 A338
885 A509
886 A361
889 A4103
890 A520
891 A607
892 A149
893 A553
894 A5
895 A686
897 A482
898 A698
899 A920
900 A890
901 A849
903 A82
907 A686
912 A477
914 A6119
917 A308
923 A617
926 A629
927 A658
933 A478
TYPE
T/N
P/B
P/N
P/B
P/N
P/B
P/N
P/N
P/N
T/N
P/N
P/N
P/N
P/N
P/N
P/N
P/N
T/B
T/B
P/B
P/B
P/N
P/B
P/B
P/B
ATDC OUTSTATION SITE LOCATIONS
LOCATION
TOMATIN
COLLINGBOURNE KINGSTON, WILTS
OLNEY, BUCKINGHAMSHIRE
NEAR BANBURY, OXON
RIDGEWAY CROSS, HEREFORDS
STONE, STAFFORDSHIRE
NEAR MELTON MOWBRAY, LEICS
STALHAM, NORFOLK
BIDSTON, MERSEYSIDE
HINTS
MELMERBY, NEAR PENRITH
NEAR ABERAERON, DYFED
COLDSTREAM, BORDERS
HUNTLY, GRAMPIAN
STRATHCARRON, HIGHLANMD
MULL, STRATHCLYDE
DUMBARTON
PATHHEAD
LLANDDOWROR, NEAR ST CLEARS
BROWNHILL, BLACKBURN
MAIDENHEAD, BERKSHIRE
NEW HOUGHTON
DROPPING WELL, NR ROTHERHAM
OTLEY ROAD, BRADFORD
TENBY, DYFED
MAP
27
174
152
151
149
127
129
134
108
139
91
146
74
29
25
49
64
66
158
103
175
120
111
104
158
GRID
LANES
REF.
78693225
2
42361568
2
48882509
2
43962322
2
37162474
2
39043344
2
47863221
2
63653253
2
32843906
4
15990329
2
35975360
2
24762599
2
38316393
2
35018405
2
19248385
2
16927387
2
41137540
2
33986638
2
22562140
2
36894308
4
48731827
2
44963652
4
43923944
2
41704340
2
21302010
2
DIR. 1 CYCLE
LANE
S
0
S
0
N
1
S
0
W
2
E
0
W
1
W
0
W
1
E
0
N
0
N
1
N
0
W
0
E
0
N
0
N
0
N
0
E
0
E
0
E
2
S
1
S
0
S
1
N
2
56
SITE
ROAD TYPE
No.
940 B3082 B/B/
941 B245
B/B
942 B4216
943 B2028
944 B195
945 B3058
946 B186
947 B4489
948 B810
949 B5259
950 C
951
952 B4512
953 B4034
954 B365
955 B3334
956 B2100
957 B5235
960 U/C
963 U/C
966 U/C
967 U/C
969 C
970 U/C
B/B
B/B
B/B
B/B
B/B
B/B
B/B
B/B
C/B
B/B
B/B
B/B
B/B
B/B
B/B
B/B
U/B
U/B
U/B
U/B
U/B
U/B
ATDC OUTSTATION SITE LOCATIONS
LOCATION
TARRANT KEYNISTON
LONDON RD, TONBRIDGE (Nr Stacey
Rd)
NEWENT, GLOUCESTERSHIRE
WEST COMMON, HAYWARDS HEATH
WELWYN GARDEN CITY
BASHLEY, HAMPSHIRE
SOUTH OCKENDON
MORRISTON, NR SWANSEA
POLMONT, CENTRAL
KIRKHAM, LANCASHIRE
ORMESBY ROAD, MIDDLESBOROUGH
HETTON LE HOLE
THORNTON, LANCS
BLETCHLEY
WALTON ON THAMES
ROWNER
ROTHERFIELD
WEST HOUGHTON
CRAMLINGTON, NORTHUMBERLAND
HINDLEY, WIGAN
SANDY LODGE WAY, NORTHWOOD
HUTTON
COPTHORNE PARK, SHREWSBURY
EAST KILBRIDE, LANARKSHIRE
MAP
GRID
LANES
REF.
195 93290457
2
188 55831475
2
162
198
166
195
177
159
65
102
93
88
102
165
176
196
188
109
88
109
176
182
126
64
37202261
53411252
25251313
42420970
59388054
26501981
29326784
34174320
45185184
35294694
34914178
86213335
10196598
57570246
55691302
65920376
42705761
36334039
50901923
34095862
34763128
26516540
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
DIR. 1 CYCLE
LANE
S
2
W
2
S
W
S
N
N
N
S
W
S
E
N
W
S
E
W
S
W
N
S
E
E
N
1
0
2
2
2
1
0
1
2
2
2
2
2
2
1
2
1
1
2
2
1
0
57
SITE
ROAD TYPE
No.
971 Unc
U/B
976 B6478
977 B4494
978 B1382
979 B6225
982 A850
2521 M1
2523 M1
2531 M1
2533 M1
2542 M4
2544 M4
2552 M4
2554 M4
2571 M5
2573 M5
2581 M5
2583 M5
2591 M6
2593 M6
2611 M6
2613 M6
2632 M6
2634 M6
B/N
B/N
B/N
B/N
P/B
M/W
M/W
M/W
M/W
M/W
M/W
M/W
M/W
M/W
M/W
M/W
M/W
M/W
M/W
M/W
M/W
M/W
M/W
ATDC OUTSTATION SITE LOCATIONS
LOCATION
STANHOPE ROAD, SMETHWICK (nr
Davison Rd)
NEAR WADDINGTON, LANCS
NEAR BRIGHTWALTON, BERKS
PRICKWILLOW, CAMBS
MILNROW, EAST OF ROCHDALE
BROADFORD
NOTTINGHAM BETWEEN J25-26
NOTTINGHAM BETWEEN J25-26
BARNSLEY BETWEEN J36-37
BARNSLEY BETWEEN J36-37
MAIDENHEAD BETWEEN J7-8/9
MAIDENHEAD BETWEEN J7-8/9
MEMBURY BETWEEN J14-15
MEMBURY BETWEEN J14-15
NORTH NIBLEY BETWEEN J13-14
NORTH NIBLEY BETWEEN J13-14
BROMSGROVE BETWEEN J4-5
BROMSGROVE BETWEEN J4-5
PENRITH BETWEEN J40-4
PENRITH BETWEEN J40-4
NEWTON-LE-WILLOWS, J22-23
NEWTON-LE-WILLOWS, J22-23
COVENTRY BETWEEN J2-3
COVENTRY BETWEEN J2-3
MAP
GRID
LANES
REF.
139 40242873
2
103
174
143
109
32
129
129
111
111
175
175
174
174
162
162
139
139
90
90
108
108
140
140
37224457
44401776
55802820
39294139
16508231
45053429
45053428
43414012
43344039
49171797
49171796
43471737
43471736
37211975
37211974
39612747
39612747
35035301
35035300
35913963
35913963
43772829
43762828
2
2
2
2
2
3
3
3
3
3
3
3
4
3
4
3
3
3
3
3
3
3
3
DIR. 1 CYCLE
LANE
N
0
S
N
W
S
S
N
S
N
S
E
W
E
W
N
S
N
S
N
S
N
S
E
W
1
2
2
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
58
SITE
ROAD
No.
2662 M3
2664 M3
2671 M11
2673 M11
2682 M50
2684 M50
2691 A1(M)
2693 A1(M)
2742 M56
2744 M56
2752 M40
2754 M40
2791 M18
2812 M55
2814 M55
2822 M20
2824 M20
2831 M3
2833 M3
2852 M180
2854 M180
2871 M25
2873 M25
2891 M1
2893 M1
TYPE
M/W
M/W
M/W
M/W
M/W
M/W
M/W
M/W
M/W
M/W
M/W
M/W
M/W
M/W
M/W
M/W
M/W
M/W
M/W
M/W
M/W
M/W
M/W
M/W
M/W
ATDC OUTSTATION SITE LOCATIONS
LOCATION
CAMBERLEY BETWEEN J3-4
CAMBERLEY BETWEEN J3-4
THEYDON BOIS J5-6
THEYDON BOIS J5-6
PENDOCK BETWEEN J1-2
PENDOCK BETWEEN J1-2
DURHAM
DURHAM
FRODSHAM J12-14
FRODSHAM J12-14
TETSWORTH BETWEEN J6-7
TETSWORTH BETWEEN J6-7
CANTLEY/COMMON BET J3-4
WOODPLUMPTON E/B BETWEEN J1-3
WOODPLUMPTON W/B BETWEEN J1-3
EAST SMEETH J10-11
EAST SMEETH J10-11
EAST STRATTON BETWEEN J8-9
EAST STRATTON BETWEEN J8-9
BRIGG BETWEEN J4-5
BRIGG BETWEEN J4-5
OCKENDON BETWEEN J29-30
OCKENDON BETWEEN J29-30
LEICESTER BETWEEN J21-22
LEICESTER BETWEEN J21-22
MAP
186
186
167
167
150
150
93
93
117
117
165
165
111
102
102
189
189
185
185
112
112
177
177
140
140
GRID
LANES
REF.
48941605
3
48941605
3
54661990
3
54661990
3
37782330
2
37782330
2
43125386
2
43145386
2
35123779
3
35123779
3
47012003
3
47012003
3
46344026
4
35154338
4
35154338
4
60691392
3
60691392
3
45351400
2
45351400
2
50184098
4
50184097
4
55831851
3
55841851
3
45163063
3
45163063
3
DIR. 1 CYCLE
LANE
E
0
W
0
N
0
S
0
E
0
W
0
N
0
S
0
E
0
W
0
E
0
W
0
N
0
E
0
W
0
E
0
W
0
N
0
S
0
E
0
W
0
N
0
S
0
N
0
S
0
59
SITE
ROAD
No.
2902 M40
2904 M40
2911 A1(M)
2913 A1(M)
2922 M4
2924 M4
2931 M6
TYPE
M/W
M/W
M/W
M/W
M/W
M/W
M/W
2933 M6
M/W
7272 M27
7274 M27
7281 M1
7283 M1
8552 A14
8554 A14
30852 A39
30857 A49
30858 A446
32634 M6
32873 M25
M/W
M/W
M/W
M/W
T/N
T/N
T/N
T/N
T/N
M/W
M/W
ATDC OUTSTATION SITE LOCATIONS
LOCATION
WARWICKSHIRE BET J15-16
WARWICKSHIRE BET J15-16
SAWTRY
SAWTRY
TRE-BRYN BETWEEN J35 & 36, MP266/6
TRE-BRYN BETWEEN J35 & 36, MP266/6
ALLOSTOCK BETWEEN J18 & 19, MP
281/4
ALLOSTOCK BETWEEN J18 & 19, MP
278/9
SOUTHAMPTON BETWEEN J4-4
SOUTHAMPTON BETWEEN J4-4
BOLSOVER BETWEEN J29-30
BOLSOVER BETWEEN J29-30
NEWMARKET
NEWMARKET
FAIRY CROSS, NEAR BIDEFORD
WESTON, SHROPSHIRE
NEAR NEC, BIRMINGHAM
COVENTRY BETWEEN J2-3
OCKENDON BETWEEN J29-30
MAP
151
151
142
142
170
170
118
GRID
LANES
REF.
42302653
3
42302652
4
18277943
4
18347941
4
29421820
3
29421820
3
37413691
3
DIR. 1 CYCLE
LANE
E
0
W
0
N
0
S
0
E
0
W
0
N
0
118
37333715
3
S
0
185
185
120
120
154
154
180
126
139
140
177
44201171
44201171
44463713
44463713
56512661
56512660
24041243
35603292
42012854
43762828
55841851
2
2
3
4
3
4
1
1
1
2
2
E
W
N
S
E
W
W
N
N
W
S
0
0
0
0
0
0
0
0
1
0
0
60
Appendix B
Representational Map of National Core ATDC Locations in Great Britain
61
APPENDIX D
Appendix C
CONTRACTOR’S PROVISION OF SPECIALIST TOOLS
Suggested list of equipment required for each site engineer
a
Dual Channel Digital Storage Oscilloscope
Note: The Oscilloscope at present used by the current maintenance contractor is a 100M
Samples/Second dual channel oscilloscope. A suitable Oscilloscope for the new contract
would be one, which has a specification that either matches or betters.]
b
1000v Megger also with a 100V range
c
Laptop computer fitted with and capable of running Windows 98 or better based
programs, preferably Office 97 and having installed an ISP E-mail programme. Along
with either an internal or external modem capable of communicating at, at least 19200
baud.
d
Lightweight portable printer and paper
e
Appropriate cables and connectors
f
Portable LCR meter
LIST OF SPARES
The Department will free issue the following spares.
Item
Quantity
Mains Electrical Distribution boards
10
US Robotics Sportster Modem
20
Heaters with thermostats
12
Hi-Trac Classifier
|
|
|
|?
|
|
|
|
|
|
|
|
|
Hi-Trac spares:
Power supply units
Backplanes
Front Panels
Primary boards
Secondary boards
Dip boards
Lead sets (for use with Hi-Trac classifier)
12V 38 Amp hour sealed lead acid battery (for use with the Hi-Trac classifier)
Cabinets with plinths
8
Standard replaceable axle/BL sensors 2.5m
“
“
“
“
2.75m
“
“
“
“
3.0m
“
“
“
“
3.5m
“
“
“
“
4.0m
“
“
“
“
4.5m
“
“
“
“
5.0m
20
|
|
|
|
|
|
| Note 1*
|
|
|
|
62
APPENDIX D
2.11 NOTE: 1 The quantity of sensors will be finalised at the commencement of contract but will cover all lengths.
2.12
Continued
Item
Loop Cable 1000m drum
Quantity
2
2.13 Contractor to collect the above spares from TSSD Depot at Kennedy Way, Yate and place them into storage
(see storage provision item 6.6)
63
Appendix E
CHECKS AND WORK TO BE CARRIED OUT AS APPROPRIATE DURING ROUTINE SITE SAFETY VISITS AT
SITES EQUIPED WITH HI-TRAC CLASSIFIERS
An outline version of the pro-forma to be used during RSI visits is included at appendix H. It is intended that
information will be recorded by use of the pro-forma, which will be agreed before the start of the contract. The
completed pro-forma is to be e-mailed to the contractors office prior to the departure of the engineer from the site.
2.14 At the six monthly visits the Contractor shall perform the following tasks.
1. Carry out a visual check of all the equipment on site, the condition of the road and road sensors array taking
particular note whether: 1.1 There is any physical damage to any of the equipment on the site. Make safe if necessary.
1.2 The sensor(s) are secure in the road. If site fitted with old style replaceable sensors ensure guides are secure
in road.
1.3 If fitted with old style replaceable sensors, the replaceable sensors are secure in the housings.
1.4 All back filling is satisfactory and intact.
1.5 The road surface is in good condition i.e. it is not cracking or forming potholes around the array.
1.6 Record the serial numbers of the classifier and modem.
2. Carry out an electrical safety inspection to check:
2.1 Earthing and earth bonding is satisfactory.
2.2 There are no frayed/corroded wires and the insulation is in good condition (particularly check heater cable).
2.3 All termination’s and connectors are secure.
2.4
Check the condition of the telecom socket and contacts and the operation of the telephone line.
2.5
Check the operation of the heater and set the thermostat to 5 C.
3. Check that:
3.1 The cabinet is securely fixed.
3.2 The locks are securing the cabinet effectively.
3.3
That the hinges are free.
4. Connect laptop to Classifier and download 3 weekdays data (including the data for the day of the visit). Using the
Raw data analysis software supplied by the Department, undertake the following checks on the data and note
any shortcomings:
4.1 That the average lengths for cars in all lanes is within tolerance bands indicated by the software.
4.2 Check the average length of cars is near to the expected length given by the software.
4.3 The average speed and overhang is within expected ranges.
4.4 Check and note that the lane distribution is consistent with accurate operation of the site.
4.5 Check the percentage validity code errors for the site are within acceptable limits (for a multilane site the %
should be well below 10%).
4.6 Check that the Twmv percentages are within acceptable limits (for most sites the percentage should be well
below 1%).
4.7 Check the Rigid Good vehicle data appears to be consistent with the accurate operation of the site.
4.8 Check the proportion of N classifications (for most sites the percentage should be well below 0.5% in all
lanes and in total for all lanes).
4.9 Check the Lgv Car Ratios (for most sites the percentage should be between 10 and 15%).
4.10 Check the Class index counts and check that they are consistent with the accurate operation of the site.
5. Enter the real time view mode of operation using the laptop and examine the data produced to check for the
accurate and consistent operation of the site. Also while in this mode try to diagnose the cause of any
shortcomings that were noted whilst undertaking the data checks in paragraph noted in 4. Only when satisfied
that the site is operating correctly or when diagnosis of the operational shortcoming noted in 4 has been made
continue with the following equipment checks.
6. Using 100-Volt Meggar record the insulation readings of the loops in Meg Ohms, Using a LCR meter record the
series resistance in Ohms and Inductance of the loops in Milli Henrys. Record the types of tails fitted to the
inroad loops and note any conditions that may effect their performance.
7. Disconnect each sensor in turn including Cycle sensors and measure and record the following parameters:7.1 Resistance (Meggar 100 Volt range) between the signal and screen in Meg Ohms.
64
7.2 Resistance (Meggar 100 Volt range) between screen and earth in Meg Ohms.
7.3 The capacitance of the sensors should be measured using the LCR meter in Nano Farads.
8. Using a digital Storage Oscilloscope check and record the noise levels (peak to peak) in milli Volts produced by
all the sensors (Axle and Cycles). All of the sensor measurements are recorded/undertaken at the BNC
interface between the sensors and the Hi-Trac.
9. Inspect the vehicle-generated waveforms in the Waveform viewing mode with the Hi-Comm software package
and optimise the waveforms using the Dip Gain boards.
10.Optimise the performance of the site using the software configurable modes of operation to overcome the
performance shortcoming diagnosed by the checks undertaken in paragraph 4.
11. Undertake a 10-minute sensor check and record the Sensor, Loop and vehicles counts for each lane. Record
the sensor waveforms for a light and heavy vehicle in each lane.
12. Observe the traffic for a period of 30 minutes to check if the classifier is operating correctly. If necessary make
additional adjustments both software and hardware to further optimise the performance of the site if further
deficiencies are seen. If any changes are required to be made to the set up of the Hi-Trac classifier after the
start of the 30-minute block listing, the block listing must be re-started. During the block listing note all misclassified vehicles noting time and vehicle number and record reason for mis-classification. The start time finish
time along with the starting vehicle and last vehicle serial numbers should be recorded.
14. After completing all operational checks on the Classifier the engineer will save on his laptop a copy of the
configuration file.
13. Check the voltage and condition of the backup battery the voltage to be measured in Volts.
14. A minimum of three photographs of each outstation should be taken during every RSI visit and should be
provided to Department in digital format. At least one photograph should confirm the road and sensor array
conditions and one facing each direction to show the road condition approaching the site. The pro-forma (see
Appendix H) should be used to note any significant changes at the site along with additional photographs
where required to show specific problems.
15. Lubricate and check the operation of all locks and door hinges.
16. If a fault(s) is discovered on third party equipment, i.e. telecom or the electricity supplier, the Contractor shall
make arrangements with the third party to rectify the fault.
17. If traffic management is required in order to rectify a fault this shall be arranged, and the necessary
rectification work carried out as soon as possible.
18. Before departure from site notify the contractors office of the faults found and obtain fault number for this work.
Complete the site logbook recording the results of all of the tests undertaken. In addition details of all minor
works undertaken, faults that have been found for which further corrective action is required and any other
points noted during the visit should also be recorded. Complete the site visit pro-forma (see Appendix H and I),
including the details of traffic management required to undertake the additional works and lanes that will
require access. Record any general comments on the site and print and leave a copy of the pro-forma in the
cabinet. Immediately prior to departure from site all of the documentation required to be produced/generated
including waveform files, data files, information files and photographs shall be E mailed to the contractors office
to be forwarded on to the Department.
19. Check that the equipment is all connected and correctly configured for remote access; the modem is plugged
into the telecom socket and fasten the doors firmly.
65
Appendix F
REPLACEMENT OF AXLE AND CYCLE SENSORS.
Axle sensors are to be replaced on an as required (i.e. when they become defective or wear out) rather than on a regular
basis.
During the replacement of an Axle sensor(s) the engineer shall undertake the following tasks. The results shall be
recorded in the site logbook and on a site fault pro-forma (see annex H) 1 copy of which shall be retained by the contractor
and a second copy will be sent to the Department. The pro-forma must include all details relating to the Fault and work
previously undertaken on the outstation e.g. Fault Number, Date raised, date of visit and previous visits if relevant, Time
on site/off site, findings, corrective action taken, test results, further corrective action required and any other comments.
1. Record the following parameters prior to the installation of the replacement sensor(s).
1.1 Serial number of new sensor(s).
1.2 Resistance as measured using Meggar 1000 Volt range between feeder cable screen and core. The reading
will be recorded in Meg Ohms and shall not be less than 999M Ohms.
1.3 Resistance as measured using Meggar 1000 Volt range between feeder cable screen and earth. The reading
will be recorded in Meg Ohms.
1.4 Capacitance between feeder cable screen and core. The readings will be recorded in Nano Farads
2. Record the following parameters after the sensor has been installed and the feeder cable has
terminated and entered into a BNC connector within the ATDC cabinet.
be
cut
to
length,
2.1 Resistance as measured using Meggar 100 Volt range between feeder cable screen and core. The reading
will be recorded in Meg Ohms.
2.2 Resistance as measured using Meggar 100 Volt range between feeder cable screen and earth. The reading
will be recorded in Meg Ohms.
2.3 Capacitance between feeder cable screen and core. The readings will be recorded in Nano Farads
3. Inspect the vehicle-generated waveforms in the Waveform viewing mode with the Hi-Comm software package and
optimise the waveforms using the Dip Gain boards. Optimise the performance of the site using the software
configurable modes of operation to overcome the performance shortcoming diagnosed whilst the inspection is being
undertaken.
4. The raw peak outputs should be recorded. In addition a waveform for a light and heavy vehicle in each lane, in which a
sensor has been changed, should be recorded.
5. Only after the performance of the site has been optimised (as outlined in paragraph 3) set the local terminal to block
listing mode. Observe the passage of traffic for a period of 1 hour on motorways and 2 hours on all other types of road and
check if the Classifier is producing accurate vehicle classifications. If necessary make additional adjustments using both
software and hardware to further optimise the performance of the site, if further deficiencies are seen. If any changes
are required to be made to the set up of the Hi-Trac classifier after the start of the specified block listing periods, the
block listing must be re-started and a further block listing period must be completed. During the block listing note all
mis-classified vehicles noting time and vehicle number and record reason for mis-classification. The start time finish
time along with the starting vehicle and last vehicle serial numbers should be recorded.
[Note: A competent site engineer will validate the vehicle classification accuracy. This will be undertaken by the
engineer comparing the ATDC classification produced following the passage of a vehicle, against the
classification that the competent engineer would assign to that vehicle.]
Cycle sensors are to be replaced on an as required (i.e. when they become defective or wear out) rather than on a regular
basis.
During the replacement of cycle sensor(s) the engineer shall undertake the following tasks. The results shall be recorded
in the site log book and on a site fault pro-forma which shall be e-mailed to the contractors office prior to the engineers
departure from site and a copy shall also be printed out and retained on site. The pro-forma must include all details
relating to the fault and work previously undertaken on the site e.g. Fault Number, Date raised, date of visit and previous
66
visits if relevant, Time on site/off site, findings, corrective action taken, test results, further corrective action required and
any other comments.
1. Record the following parameters prior to the installation of the replacement sensor(s).
1.1 Serial number of new sensor(s).
1.2 Resistance as measured using Meggar 1000 Volt range between feeder cable screen and core. The reading
will be recorded in Meg Ohms and shall not be less than 999M Ohms.
.
1.3 Resistance as measured using Meggar 1000 Volt range between feeder cable screen and earth. The reading
will be recorded in Meg Ohms.
1.4 Capacitance between feeder cable screen and core. The readings will be recorded in Nano Farads
2. Record the following parameters after the sensor has been cut to length and for both types of classifiers terminated in a
B.N.C. connector within the ATDC cabinet.
2.1 Resistance as measured using Meggar 100 Volt range between feeder cable screen and core. The reading
will be recorded in Meg Ohms.
2.2 Resistance as measured using Meggar 100 Volt range between feeder cable screen and earth. The reading
will be recorded in Meg Ohms.
2.3 Capacitance between feeder cable screen and core. The readings will be recorded in Nano Farads
3. Inspect the vehicle-generated waveforms in the Waveform viewing mode with the Hi-Comm software package and
optimise the waveforms using the Dip Gain boards. Optimise the performance of the site using the software
configurable modes of operation to overcome the performance shortcoming diagnosed whilst the inspection is being
undertaken.
4. A waveform for a Light and Heavy vehicle in each lane in which a sensor has been changed should also be
recorded along with if possible a waveform for a Pedal Cycle or a Motorcycle.
5. Only after the performance of the site has been optimised (as outlined in paragraph 3) set the local terminal to block
listing mode. Observe the passage of traffic for a period of 1 hour and check if the Classifier is producing accurate vehicle
classifications. If necessary make additional adjustments using both software and hardware to further optimise the
performance of the site, if further deficiencies are seen. If any changes are required to be made to the set up of the HiTrac classifier after the start of the specified block listing period, the block listing must be re-started and a further block
listing period must be completed. During the block listing note all mis-classified vehicles noting time and vehicle
number and record reason for mis-classification. The start time finish time along with the starting vehicle and last
vehicle serial numbers should be recorded.
If the period between the sensors replacement and the previous RSI is more than 3 months then a complete set of routine
checks as detailed in Appendix E, shall be carried out.
67
2.15 Appendix G
Example of electronic documentation process to be followed for a fault raised by TSR
1) TSR contact maintenance contractor and obtain next fault number from the fault reporting system maintained by
the contractor.
2) TSR insert the details of the fault on an electronic document and the file is named in accordance to a file naming
structure agreed between the department and the Contractor. The Document is then sent by TSR to the
maintenance contractor and Highways Agency.
3) Maintenance contractor sends the electronic document to the engineer assigned to attend the fault.
4) Engineer completes the form detailing diagnosis of the problem, the results of all tests undertaken, serial numbers
of any equipment installed and removed along with the details and extent of additional works required along with
the traffic management required to complete these works. Prior to departure from site the engineer “e mails” the
electronic form to the maintenance contractor, TSR and Highways Agency. The engineer also prints off a form and
it is retained on site in the site book.
5) If additional works are required a new electronic form should be generated, using the agreed naming procedure,
by the engineer and paragraph 4) repeated for every visit made to site.
6) Maintenance contractor submits invoice details of information required to be submitted as (WRD paragraph
16.2.2). Copies of the electronic documents accompany the submitted invoice for every visit made.
68
Example of electronic documentation process to be followed for a fault raised by maintenance contractor.
1) Maintenance Contractor obtains next fault number from the fault reporting system they maintain the file is named
in accordance to a file naming structure agreed between the department and the Contractor. Complete an
electronic form detailing nature of the Fault “e mail” the form to TSR and Highways Agency.
2) Maintenance contractor sends the electronic document to the engineer assigned to attend the fault.
3) Engineer completes the form detailing diagnosis of the problem, the results of all tests undertaken, serial numbers
of any equipment installed and removed along with the details and extent of additional works required along with
the traffic management required to complete these works. Prior to departure from site the engineer “e-mails” the
electronic form to the maintenance contractor, TSR and Highways Agency. The engineer also prints off a form and
it is retained on site in the site book.
4) If additional works are required a new electronic form should be generated, using the agreed naming procedure,
by the engineer and paragraph 3) repeated for every visit made to site.
5) Maintenance contractor submits invoice details of information required to be submitted as (WRD paragraph
16.2.2). Copies of the electronic documents accompany the submitted invoice for every visit made.
69
Example of electronic documentation process to be followed for a fault diagnosed by a maintenance engineer
whilst undertaking a RSI or attending an unrelated fault.
1) Maintenance engineer contacts depot and informs them verbally of the nature of the fault and obtains the next
fault number from the fault reporting system the maintenance contractor maintains. The engineer completes the
normal RSI electronic form but also completes the fault paragraph on the form with the fault details. The electronic
RSI form is named in accordance with the file naming structure agreed between the department and the
Contractor. The completed form along with the WRD defined RSI documentation is “e-mailed” to the contractors
depot, TSR and Highways Agency prior to departure from site. The engineer also prints off a form and it is
retained on site in the site book.
2) During follow up visits the engineer completes a form detailing all works undertaken, the results of all tests, serial
numbers of any equipment installed and removed along with the details and extent of additional works required
along with the traffic management required to complete these works. Prior to departure from site the engineer “emails” the electronic form to the maintenance contractor, TSR and Highways Agency. The engineer also prints off
a form and it is retained on site in the site book.
3) If additional works are required a new electronic form should be generated, using the agreed naming procedure,
by the engineer and paragraph 2) repeated for every visit made to site.
4) Maintenance contractor submits invoice details of information required to be submitted as (WRD paragraph
16.2.2). Copies of the electronic documents accompany the submitted invoice for every visit made including the
RSI sheet.
70
2.16 Appendix H
R.S.I. REPORT SHEET
ATDC CONTRACT
SITE NUMBER:
Location:
VISIT DATE:
Immediately Below only to be completed if a Fault is raised during the R.S.I. Visit
DATE FAULT
RAISED:
FAULT NUMBER:
DETAILS OF
REPORTED FAULT:
SITE CONDITION
CHECK LIST
FAULT RAISED BY:
COMMENTS
WORK UNDERTAKEN/REQUIRED
Physical site damage
Backfilling/Epoxy
Road surface
Cabinet doors locks
And hinges
Earthing and bonding
Duct seals
Wiring including
Heater
Terminations
Telephone line
including
Phone socket
71
ELECTRONIC
EQUIPMENT
CHECK LIST
COMMENTS/SERIAL NUMBER
WORK UNDERTAKEN/REQUIRED
Classifier
Modem
Loop Test Results
LOOP
Impedance
To earth
Series
Resistance
Inductance
Cable
Type
Remarks
Lane 1
Lane 2
Lane 3
Lane 4
72
Sensor Test Results
(As tested at the BNC Connector at the back of the Equipment)
Impedance
Axle
Sensor
Sig/Scr
Scr/E
Cap
NF
Raw Noise
MV
Raw Peak Signal
+ve
-ve
1
2
3
4
5
6
7
8
Sensor Type (B.L./T2000)
Lane 1
Lane 2
Lane 3
Lane 4
Sensor Length (Metres)
Lane 1
Lane 2
Lane 3
Lane 4
Cycle Sensor Test Results
(As tested at the BNC Connector at the back of the Equipment)
Cycle
Sensor
Capacitance
(Nf)
Noise
(mV)
Impedance
Sig/Scr
Impedance
Scr/E
Signal
(mV)
1A
1B
2A
2B
73
Sensor Interval Test Results.
(For 10 Minute interval Test)
Lane Number
1
Sensor 1 Count
Sensor 2 Count
Loop Count
Vehicle Count
2
3
4
5
6
7
8
Sensor Waveforms
Lane 1
Light Vehicle Description
Heavy Vehicle Description
1
2
3
4
5
6
7
8
Activities Check List
Activity
Remarks/Comments
Starting vehicle Record
Number
Finishing vehicle Record
Number
Take and monitor ½ hour V.B.V. block listing.
Check heater Thermostat set to 5 C
Check Voltage and condition of backup battery
Lubricate Locks/Hinges
Check and
necessary
clean
B.T.
socket
connections
if
All Equipment switched on and Connected
B.T.Line/Modem Connected
74
Classifier Check List
Comments
Speed Limit at classifier?
Is there a different speed limit sign in sight?
If yes, record limit and direction from classifier
Is street lighting present at the classifier?
Is car parking likely?
Is a bus stop nearby?
Is a school nearby?
Is a works entrance nearby?
Are there shops nearby?
Is there a pedestrian crossing nearby?
Are there traffic lights nearby?
Approximate distance from nearest road
junction
Type of road surface
Loading/Un-loading nearby?
Other factors that may influence site data
collection.
Details of all minor work Undertaken During R.S.I.
Details of All Outstanding Work
75
Details of Traffic Management Requirements and Lanes that need to be Accessed
General Comments
Pre-Site Departure Activities
Comments
Print out copy of R.S.I. for retention at site
Complete Log book
“E” mail R.S.I. sheet, Sensor Waveforms, Vehicle
Record File and any photographs to Office
Engineers Details
Name of Engineer Attending R.S.I.
Time of Arrival :-
Time of Departure:-
76
2.17 Appendix I
SITE FAULT VISIT REPORT SHEET
ATDC CONTRACT
FAULT NUMBER:
DATE FAULT RAISED:
DETAILS OF REPORTED FAULT:
VISIT NUMBER FOR THE ABOVE FAULT:
SITE NUMBER:
LOCATION:
VISIT DATE:
SITE CHECK LIST
FAULT RAISED BY:
COMMENTS
WORK UNDERTAKEN/REQUIRED
Physical site damage
Backfilling/Epoxy
Road surface
Cabinet doors locks
And hinges
Earthing and bonding
Duct seals
Wiring
Terminations
Telephone line
including
Phone socket
77
ELECTRONIC
EQUIPMENT
CHECK LIST
COMMENTS
WORK UNDERTAKEN/REQUIRED
Classifier
Modem
Loop Test Results
LOOP
Impedance
To earth
Series
Resistance
Inductance
Cable
Type
Remarks
Lane 1
Lane 2
Lane 3
Lane 4
Sensor Tests Prior to Installation
Sensor Serial
Number
Date of Manufacture
Impedance
Core/Screen
Cap
NF
Location Installed
78
Sensor Test Results
(As tested at the BNC Connector at the back of the Equipment)
Impedance
Axle
Sensor
Sig/Scr
Scr/E
Cap
NF
Raw Noise
MV
Raw Peak Signal
+ve
-ve
1
2
3
4
5
6
7
8
Sensor Type (B.L./T2000)
Lane 1
Lane 2
Lane 3
Lane 4
Sensor Length (Metres)
Lane 1
Lane 2
Lane 3
Lane 4
Cycle Sensor Test Results
(As tested at the BNC Connector at the back of the Equipment)
Cycle
Sensor
Capacitance
(Nf)
Noise
(mV)
Impedance
Sig/Scr
Impedance
Scr/E
Signal
(mV)
1A
1B
2A
2B
79
Sensor Interval Test Results.
Lane Number
Sensor 1 Count
Sensor 2 Count
Loop Count
Vehicle Count
1
2
3
4
5
6
7
8
Sensor Waveforms
Lane 1
Light Vehicle Description
Heavy Vehicle Description
1
2
3
4
5
6
7
8
Details of All work Undertaken During Visit.
Starting Vehicle
Record Number
1 Hour
Check
Finishing Vehicle
Record Number
Details (Commencement and finish times) of
Operational checks undertaken (if applicable).
Starting Vehicle
Record Number
2 Hour
Check
Finishing Vehicle
Record Number
Details of All Outstanding Work
80
Details of Traffic Management Requirements and Lanes that need to be Accessed
General Comments
Pre-Site Departure Activities
Comments
Complete Log book
“E” mail Fault sheet, any Sensor Waveforms,
Vehicle Record Files, photographs or site generated
information to Office
Engineers Details
Name of Engineer Attending Fault:Time of Arrival :-
Time of Departure:-
81
Schedule A
2.18 BILL OF QUANTITIES Sheet 1
ITEM 1. . FIXED PRICE COSTS
Item
Ref.
Description
Cost £
1.1
Sects 6.1
and 6.4
Annual cost to undertake the routine and preventative maintenance
of all the sites
.....................
1.2
Sects 7.1
and 7.2
Change in annual maintenance costs (items 1.1, 1.2, or 1.3 as
appropriate) for the addition or reduction of a site with 1 lane.
+
.....................
1.3
Sects 7.1
and 7.2
Change in annual maintenance costs (items 1.1, 1.2, or 1.3 as
appropriate) for the addition or reduction of a site with 2 lanes.
+
.....................
1.4
Sects 7.1
and 7.2
Change in annual maintenance costs (items 1.1, 1.2, or 1.3 as
appropriate) for the addition or reduction of a site with 3 lanes.
+
.....................
1.5
Sects 7.1
and 7.2
Change in annual maintenance costs (items 1.1, 1.2, or 1.3 as
appropriate) for the addition or reduction of a site with 4 lanes.
+
.....................
1.6
Sects 7.1
and 7.2
Change in annual maintenance costs (items 1.1, 1.2, or 1.3 as
appropriate) for the addition or reduction of a lane(s) of in-road
equipment at a site, (costs per lane per site).
+
.....................
1.7
Sects 7.1
and 7.2
Change in annual maintenance costs (items 1.1, 1.2, or 1.3 as
appropriate) for the addition or reduction of 2 cycle sensors at a
site, (costs per lane per site).
+
.....................
Note 1.
All costs to be exclusive of VAT.
82
NOTES FOR USE WITH SCHEDULE B
All costs to be exclusive of VAT.
Definition of Terms
Definition of Terms
“Removal” involves the following:2.19 All the civil, mechanical and electrical works required to complete the taking away of all the relevant
equipment, the fittings and fixings from the place where they were installed.
2.20 Transportation of salvaged equipment from site to contractor’s depot.
2.21 The proper disposal of unusable and or unwanted equipment.
2.22 Storage of salvaged equipment.
2.23 Making good the condition of the site (see paragraph 5.15).
2.24 “Install” involves the following:2.25 Transportation of equipment and materials to site.
2.26 All the civil, mechanical and electrical works required to complete the site works and to position, interconnect
and fix all the relevant equipment.
2.27 Removal of debris from site and making good the condition of the site.
“Commission” involves the following :
2.28 Confirming that the equipment is physically installed in accordance with the appropriate drawings and
requirement documents and that the correct materials have been used.
2.29 Confirming that the correct type and quantity of equipment has been installed and that the physical and
electrical interconnections between all the installed equipment are correctly made.
2.30 Carrying out all adjustments and or rectification work which may be required.
2.31 Setting the equipment to operate correctly and functionally testing the installed equipment which shall include
checking and recording the performance against defined levels.” Commission” includes all necessary
engineering support if Site Set Up is required, see Appendix I, paragraph 1.
83
Schedule B
BILLS OF QUANTITIES
Sheet 1
ITEMS 2. TO 5. Inclusive
ADDITIONAL COSTS ( see Paragraph 7 )
WORKS REQUIRED
Item
NOTES 1.
2.
Description
Qty
Removal
Costs £
Install Costs
£
Commission
Costs £
2
Complete Outstation Site without
Cycle Sensors and which has:-
2.1)
2.2)
2.3)
2.4)
1 Lane
2 Lanes
3 Lanes
4 Lanes
1
1
1
1
..............
..............
..............
..............
..............
..............
..............
..............
...............
..............
..............
..............
3
Roadside Equipment
3.1)
3.2)
3.3)
3.4)
3.5)
All equipment inc. 609 cabinet(s)
All equipment no 609 cabinet(s)
All equipment no classifier cabinet
Classifier Cabinet (GKH010)
609 Cabinet
1
1
1
1
1
..............
..............
..............
..............
...............
..............
..............
..............
..............
..............
..............
..............
..............
xxxxxxxxxx
xxxxxxxxxx
4
All in-road Equipment at an
outstation site (with cycle sensors) Per site
4.1)
4.2)
4.3)
4.4)
1 Lane Equipment *
2 Lanes Equipment *
3 Lanes Equipment
4 Lanes Equipment
1
1
1
1
..............
..............
..............
..............
..............
..............
..............
..............
..............
..............
..............
..............
5
All in-road Equipment at an
outstation site (no cycle sensors) Per site
5.1)
5.2)
5.3)
5.4)
1 Lane Equipment *
2 Lanes Equipment *
3 Lanes Equipment
4 Lanes Equipment
1
1
1
1
...............
..............
..............
...............
..............
..............
..............
..............
..............
..............
..............
..............
The lane(s) referred to in items 2 ,4 and 5 shall be the number of running lanes.
The costs for items 2 to 5 inclusive shall include the cost of labour. Also see note 3 on sheet 2 of Schedule B.
3.

Install costs include provision of all materials other than those specified in items 9.1 to 9.31 inclusive of Schedule C. The relevant materials within items 9.1
to 9.31 inclusive shall be chargeable separately where appropriate.
These shall also apply where lanes are added to an existing site.
84
Schedule B
BILLS OF QUANTITIES
ITEMS 6. AND 7.
ADDITIONAL COSTS
WORKS
Item
REQUIRED
Description
Qty
Removal
Costs £
Install
Costs £
Commission
Costs £
..............
..............
..............
..............
..............
..............
..............
..............
..............
..............
..............
..............
..............
..............
..............
..............
..............
..............
...............….
....................
xxxxxxxxxx
xxxxxxxxxx
....................
....................
...................
…................
…................
6
Individual in-road
Equipment Items,
Per site
6.1) Permanent BL sensor
6.2) Additional permanent
BL sensor
6.3) BL (cycle length)
Sensor
6.4) Additional BL (cycle
length) Sensor(s)
6.5) Double D Loop
1 Per
Sensor
1 Per
Sensor
1 Per
Sensor
1
7
Trenching Per
Metre
7.1) Through Earth
7.2) Through
Asphalt/paving/concrete
Per
Metre
NOTES 1.
2.
3.
..............
..............
The costs for items 6 and 7 inclusive shall include the cost of labour.
Install costs include provision of all materials other than those specified in items
9.1 to 9.31 inclusive of Schedule C. The relevant materials within items 9.1 to 9.31
inclusive shall be chargeable separately where appropriate.
Items 2,3,4 and 5 all include the costs for items 6.1 to 6.9 where relevant.
85
Schedule C
BILLS OF QUANTITIES
Sheet 1
ADDITIONAL COSTS
ITEM 9. Materials
Unit cost £
Item
Description
1 to 4
5 to 24
25+
9.1
9.2
9.3
9.4
9.5
9.6
9.7
9.8
9.9
9.10
9.11
9.12
9.13
9.14
9.15
9.16
9.17
Applied Weighing Classifier
Modem, US robotics Sportster V34
Wiring loom for Classifier Cabinet
12V 38 Amp hour Sealed Lead Acid Battery
Cabinet Heater (60 watts)
Mains Distribution Panel
Classifier Cabinet c/w plinth
Earth spike with clamp
BL axle Sensor 1.50m (used as cycle sensors)
BL axle Sensor 2.00m (used as cycle sensors
BL axle Sensor 2.50m
BL axle Sensor 2.75m
BL axle Sensor 3.00m
BL axle Sensor 3.25m
BL axle Sensor 4.00m
BL axle Sensor 4.50m
BL axle Sensor 5.00m
...............
...............
...............
...............
...............
...............
...............
...............
...............
...............
........ ......
...............
...............
...............
...............
...............
...............
...............
...............
...............
...............
...............
...............
...............
...............
...............
...............
............ ..
...............
...............
...............
...............
...............
...............
xxxxxx
xxxxxx
xxxxxx
xxxxxx
xxxxxx
xxxxxx
xxxxxx
xxxxxx
xxxxxx
xxxxxx
xxxxxx
xxxxxx
xxxxxx
xxxxxx
xxxxxx
xxxxxx
xxxxxx
Qty
9.18
9.19
9.20
Loop Cable per 100m reel
16mm oval EGA tube per 3 metre length
10mm copper tube per 40m reel
1
1
1
Cost £
...........
...........
...........
2.32 NOTE Item 9.6 Mains distribution panel comprises a single way consumer unit (for electricity board
termination), an earth leakage circuit breaker, a single and a twin 13 amp socket, a thermostat and a RC
suppresser all wired and mounted on a wooden board, with 0.5 metre of heat resistant wire from the
thermostat for connection to a heater.
2.33 ]
86
Schedule D
2.34 ADDITIONAL COSTS Sheet 1
ITEM 10. Equipment Mark Up
Percentage mark up on equipment, not listed under item 9 and excluding test equipment, purchased specifically for
the contract work, where the equipment is not provided by the Department.
............%
87
Schedule E
Sheet 1
2.35 ITEM 11. Engineer’s Rates
2.36 Additional Costs
Item
Ref.
11.1
5.7
Project
Engineer
Dedicated
Site Engineer
Trained
Site Engineer
Engineer’s Hourly Rate
........
........
........
11.2
Engineer’s Daily Rate
........
........
........
11.3
Engineer’s Weekly Rate
........
........
........
11.4
Travel Rate per mile
.................
88
Schedule E
Sheet 2
2.37 ITEM 12. Fault Attendance and Traffic Management
Unscheduled Maintenance/Additional Costs
Item
Ref
2.1
7.7
2.2
2.3
7.7
13.1
Description
Fixed payment for attendance at site within 3
working days of notification of a fault and
subsequent fault visits to correct same fault.
Reduced fixed payment for attendance at site more
than 3 working days after notification of a fault and
subsequent fault visits to correct same fault.
Cost in Pounds
.................................
...
.................................
..
Arrangement fee for Traffic Management required to
carry out any in-road work at a single location.
.................................
...
89
Schedule F
Sheet 1
STAFF ALLOCATION AND DEPOT DISTRIBUTION
FOR THE MAINTENANCE OF THE AUTOMATIC TRAFFIC DATA
COLLECTION SYSTEM
Total
System
Allocation
No of Depots Proposed
Location of Depots
Location of Project Manager
Total No of Dedicated Field Engineers
Total No of Trained Engineers
Annual % of total time each trained staff member will spend on
this contract
[Note: We have specified the posts of Project Manager and Project Engineer as
separate titles to allow potential contractors to employ a project manager who is not
necessarily an engineer. However there is no reason why these posts cannot be
allocated to one suitably qualified individual. The post of project engineer must be a
dedicated post
CVs, detailing experience and qualifications, to be supplied on separate sheets for:The Project Engineer
Each Dedicated Field Engineer
Each Trained Engineer
Annex 8.4
International Road Traffic Statistics Review
Dr John Disney
Nottingham Business School
Nottingham Trent University
Aims
This paper investigates Road Traffic Statistics worldwide so that the RTS QR
is able to share international best practice. A subsidiary aim is to check that
there is no impending directive / guidance from the European Commission
(EC) regarding RTS which may affect GB in the immediate future.
Objectives


To investigate and review Road Traffic Statistics published
internationally with particular reference to their collection methodology
To investigate whether revisions or amendments are being proposed to
these statistics
Methodology
The research was undertaken by internet searching and e-mail queries to
appropriate personnel in the EC, Office for National Statistics (ONS), US
Department of Transportation Federal Highway Administration and elsewhere
between November 2004 and February 2005. Some useful contacts were
made which will enable future developments to be monitored.
Global Findings
Whilst the EC publishes copious numbers of transport related publications
from the 2001 EC White Paper on European transport policy for 2010: time
to decide to European Energy and Transport Trends to 2030 which all
depend heavily upon statistics, these publications do not appear to be
addressing the collection or calculation of these statistics to the extent that the
RTS QR has undertaken.
I searched the Energy and Transport in Europe Digest from April 2002
onwards and ordered the DVD released on Transport Statistics in January
2004 (Everything on transport statistics – DVD). This is the only reference
to “Road Statistics” although there were references to both Rail and Air
Statistics. The DVD is a repository of all public documents and data related to
transport statistics in Europe and main partner countries with 20 million
statistical data items and over 850 documents. However despite repeated email messages using the link provided I never received this DVD. It is
disappointing to note that some EC publications appear to be so difficult to
obtain.
I also searched the Eurostat website and consulted with the ONS and the
Eurostat Support Service concerning the methodology underpinning Road
Traffic and Road Length Statistics collected across the EC. I only located
rules on statistical compilation and data quality for road freight statistics; the
response from Eurostat is that
“the only data collection we have is a voluntary data collection on road
lengths and on traffic at a very aggregated level. Since it is voluntary, we have
some problems with completeness as well as definitions."
I have studied the Summary Minutes of the 2004 meeting of the Coordinating
Committee for Statistics on Transport (CCST) and am satisfied that there are
no CSST developments likely to impinge on our review. Indeed, there is no
Directive or Regulation requiring Member States to provide Road Traffic or
Length statistics to Eurostat. Statisticians from DfT are represented on
CCST.
I have discovered that the UNECE (United Nations Economic Commission for
Europe) has a programme of work in progress with the ultimate aim of
“Improved international comparability of transport statistics”. Although it is
understood that DfT makes only a limited contribution to this work, the
Department keeps itself informed of developments.
I recommend that the Review uses the Glossary for transport statistics
(Third Edition, 2003) definitions of infrastructure, vehicle categories, traffic
and transport measurement. This document was prepared by the UNECE's
Intersecretariat Working Group on Transport Statistics.
I also researched the European Union Road Federation (ERF) which is a
Road User’s Group and is the European branch of the International Road
Federation (IRF). The ERF publishes European Road Statistics 2004 and
the IRF publishes World Road Statistics but both publications rely upon
secondary data sources and give no information on the methodology used.
Case Studies
I investigated three countries in detail in December 2004 / January 2005: US,
Germany and Netherlands.
United States
The US Department for Transportation Federal Highway Administration
Traffic Monitoring Guide May 2001.
http://www.fhwa.dot.gov/ohim/tmguide/
is recommended; in particular the Executive Summary is a good overview.
The US Federal Highway Administration Traffic (FHWA) recommended data
collection plan is a balance between permanent continuous & temporary short
term (48 hours recommended duration) counts. Manual counts are only used
as a last resort.
Vehicles are classified into thirteen types although short term counts are often
aggregated as there is insufficient data in some categories. Overall the US
seems to be most greatly concerned with truck movements and has instituted
“Weight in Motion” surveys. US Traffic Statistics are issued monthly.
The State of the Practice for Traffic Data Quality White Paper Dec 2002
http://tmcpfs.ops.fhwa.dot.gov/cfprojects/uploaded_files/StateOfThePractice.d
oc
looks forward to how Intelligent Transport Systems (ITS) can be utilised
especially in an urban setting.
Overall I am impressed with the work done in the US and recommend that
DfT keeps a close eye on developments in the US. In particular there is a
wealth of statistical detail regarding confidence intervals and required sample
sizes and how to convert short term data into long term statistics.
Germany
The Federal Republic of Germany is very active in collecting data on traffic
levels, probably prompted by the unification of East and West Germany.
They undertake a nationwide road traffic count on a five year basis to monitor
traffic development and determine traffic volumes on federal trunk roads,
autobahns and other “main through roads”. This was last undertaken in 2000
so was presumably being repeated in 2005 (not confirmed). The methodology
is reported only in German but my understanding is that this is essentially a
survey using manual counts on 43 specified days during the year split across
“Normal Working Days” in May, June and September together with some
Fridays and Sundays in those months and some weekdays in the so called
holiday period between late July and August.
Counts are undertaken on a clockface hour basis between 0700 and 0900
and 1500 and 1900 with vehicles classified into seven types: bicycle;
motorcycle; car and minibus up to 9 seats; bus (10+ seats); van / lorry up to
3.5t; rigid lorry over 3.5t; non-rigid lorries. Rural roads are excluded from this
study. The report is supplemented by extensive maps showing the data
collection points and the estimated traffic flows on “main through routes”.
They also have continuous monitoring by 1170 permanent automatic traffic
counters supplemented as necessary by manual counts. These ATC’s collect
data 24/7 in clockface hours; they do not consider it necessary to use shorter
time intervals. Vehicles are classified into eight distinct categories (note
discrepancy with manual counts) plus “A N Other”. They do not consider that
their current technology is capable of coping with further classifications nor
that this is necessary. An annual evaluation of these ATC’s is undertaken;
data for 2002 was published in August 2004 and was thus the most recent
evaluation at the time of study. Whilst the term “rural federal roads” is used in
these statistics, these account for over 50% of the total vehicle km recorded
so the use of “rural” is different to GB.
It is noteworthy that a charge is made to download information on Traffic
Statistics from the German Government website.
Netherlands
In 1992 a new transport monitoring system was implemented in the
Netherlands as a part of the 2nd National Transport Structure Plan 1990 –
2010. This system is named “To Measure = To Know”. Congestion (which is
not the same as traffic) has been adopted as a key indicator; until 1998 this
was monitored using manual counts taken by the police. Since 1998
electronic detection has been built into the road surface for monitoring
purposes.
To measure this a dynamic measuring network consisting of a high density of
counters in congested areas measures traffic flow (categorized) and speed in
15 minute intervals. There is a data quality problem and also a problem with
the vast amount of data generated. This is a very complex data collection
process and there are difficulties in identifying and resolving causes of
problems.
Data is also collected continuously for monthly statistics (published six weeks
later). Measuring on a rotational basis is accepted as having some cost
savings but the risks and uncertainty which this can introduce are not
considered worthwhile. A commercial contractor is employed to collect the
data but it is validated by the Ministry of Transport.
Data is collected at widely dispersed points throughout the Netherlands and
categorised traffic flow and speeds are recorded at hourly intervals although
the statistics are presented using eleven time periods per day. Problems
encountered are mainly organisational but the need to repair damaged loops
leads to missing data. There are also concerns about the time taken to
process the data and publish the statistics.
Regional Directorates undertake incidental measurements for local purposes.
Conclusions
This research confirmed that the RTS QR is not conflicting with any current or
proposed European policies. The DfT is well ahead of the EC regarding the
collection and publication of Road Traffic Statistics and indeed should not be
afraid to suggest that the EC adopts its policies.
Some EC members such as Germany and the Netherlands are more active
and proactive than others, hence the reason for selecting them as case
studies. In each case there are specific lessons to be learnt from them but no
reason for DfT to make any radical changes.
It is however recommended that the work of the US FHWA is studied and best
practice derived from it, including both data collection techniques and
statistical methodology.
Frequency of publication of statistics varies; of more concern appears to be
the time lag between data collection and publication of statistics.
In all cases studied, it is evident that ad hoc statistics are produced at a “micro
level” often specifically for a particular reason. Furthermore it is apparent that
the overarching authority in each case finds it very difficult to co-ordinate and
optimise these efforts. This does not mean that DfT should avoid this task but
should serve as a warning of the barriers likely to be encountered.
2.38
Annex 8.5
2.39
2.40
DIFFERENCES
BETWEEN ESTIMATES
FROM THE
CONTINUING SURVEY OF ROAD GOODS TRANSPORT AND THE
ROAD TRAFFIC STATISTICS (RTS) FOR HEAVY GOODS VEHICLE
(HGV) KILOMETRES
1.
Introduction
1.1
The review of the CSRGT compared some of the vehicle-kilometre
results with those from the RTS. This note shows the comparisons and
considers the implications for the current review of the RTS system.
1.2
Both the CSRGT and the RTS produce estimates of vehicle-km of HGV
(lorries etc over 3.5 tonnes gross vehicle weight (GVW)). However the
coverage of the two surveys is slightly different, as shown below.
RTS
Based on goods vehicles over 3.5
tonnes gvw
Includes foreign vehicles
Includes cranes, ambulances, fire
engines and agricultural vehicles
2.
CSRGT
Based on GB-registered goods
vehicles over 3.5 tonnes gvw
Excludes foreign vehicles
Excludes cranes, ambulances, fire
engines and agricultural vehicles
Comparisons by vehicle-type
2.1
The total vehicle-km estimated by CSRGT between 1999 and 2001 is
some 22% below that estimated by RTS, as shown below.
Table Z
Comparison of vehicle kms recorded on CSRGT and RTS
Vehicle-type (as
recorded by
CSRGT)
1999
2-axle rigids (R2)
3-axle rigids (R3)
4 axle rigids (R4)
3 & 4 axle artics
(A34)
5 axle artics (A5)
6 axle artics (A6)
Total
CSRGT
RTS1
Difference
Billions kms
% difference
9.3
1.5
1.2
2.2
11.9
1.7
1.5
3.1
-2.6
-0.2
-0.3
-0.9
-22%
-12%
-20%
-28%
5.3
3.6
23.1
7.4
3.3
28.8
-2.1
0.3
-5.7
-29%
8%
-20%
2000
2001
2-axle rigids (R2)
3-axle rigids (R3)
4 axle rigids (R4)
3 & 4 axle artics
(A34)
5 axle artics (A5)
6 axle artics (A6)
Total
2-axle rigids (R2)
3-axle rigids (R3)
4 axle rigids (R4)
3 & 4 axle artics
(A34)
5 axle artics (A5)
6 axle artics (A6)
Total
9.1
1.5
1.2
1.9
11.7
1.6
1.5
2.9
-2.6
-0.1
-0.3
-1.0
-22%
-4%
-21%
-33%
4.4
4.8
23.0
7.2
4.4
29.2
-2.8
0.4
-6.2
-39%
9%
-21%
8.5
1.7
1.1
1.7
11.7
1.6
1.5
2.9
-3.2
0.1
-0.4
-1.2
-28%
5%
-25%
-41%
3.8
5.4
22.2
7.2
4.3
29.2
-3.4
1.1
-7.0
-47%
26%
-24%
2.2
The RTS have been consistently higher than CSRGT veh kms for
many years. Whilst we would expect this, given the RTS's wider
coverage (foreign vehicles, and the miscellaneous vehicles), this
difference is more than could be accounted for by the additional
coverage, and the CSRGT Review concluded that CSRGT was
undercounting overall HGV activity. This finding was based on
comparisons with tachograph readings and other comparisons.
2.3
Detailed comparisons were made by vehicle-type and the conclusions
were as follows:

R2s. The actual difference in this group was probably greater than the
3.2 bn vkm recorded in 2001. This is because the RTS records activity
by R2s pulling drawbar trailers as activity by articulated lorries.
Drawbar trailers do not account for much activity by GB vehicles,
although they account for a larger proportion of foreign lorry activity
(which is not captured by the CSRGT) within GB.

It was also likely that the RTS was mistakenly recording some activity
by light vans as R2s because of limitations with the automatic counting
equipment.

R3. Here there tends to be reasonable agreement between the two
sources, although any drawbar trailer pulled by a R3 will be recorded
as artic. by RTS which means that the real differences may be greater
than shown;

R4. The RTS and the CSRGT should record activity the same way, so
that the significant shortfall must be due to undercounting by CSRGT;

A3/4. The substantial differences between the CSRGT and RTS
estimates if thought to be partly due to two-axle rigids with drawbar
trailers classified as three/four axle artics by RTS. Furthermore, there
is a high incidence of foreign-registered two-axle rigids with two-axle
trailers that are likely to be recorded as four-axle artics by RTS.

A5s. The large difference is partly accounted for by foreign lorries
captured in the RTS, plus A6 lorries with an axle retracted, which are
recorded as A5 in the RTS;

A6s. This is the only category that routinely shows more veh-km in
CSRGT than in the RTS. This is probably because RTS records A6
lorries with an axle retracted as A5.
2.4
In general the main conclusions affecting RTS were that some further
investigations were needed. From the above comparisons, the main
areas are

To investigate possible cases of misclassification, in particular between
light vans and R2s.

To consider the rules for classification of lorries towing drawbar trailers

To consider the rules for classifying miscellaneous vehicles.
3.
Counting trials
3.1
Provided that these concerns are viewed as potentially undermining
the integrity of the RTS figures on HGVs, it is recommended that trials
be undertaken over several hours at two or three ATC sites,
concentrating on counts of van and lorry traffic (incl. Miscellaneous
vehicles). These trials would compare (a) manual counts by a highly
trained team of enumerators or possibly ANPR, (b) manual counts by a
normal team of manual enumerators, and (c) counts form standard
classifying ATC equipment.
3.2
The objectives would be

to assess the extent of any misclassification in normal MCCs and
ATCs, in particular between light van and R2 counts;

to assess the amount of traffic accounted for by rigid lorries towing
drawbar trailers;

to assess the amount of traffic accounted for by miscellaneous vehicles
not covered by CSRGT.
3.3
As a consequence we could then assess more accurately the extent of
the misclassification between CSRGT and the RTS. In addition we
should be able to

improve the training of enumerators, and the programming of automatic
classifiers, to obtain more accurate counts of light vans and R2s; and

decide whether the incidence of drawbar trailer activity, and activity by
miscellaneous vehicles is sufficient to justify recording these vehicles in
separate categories.
4. Summary of Recommendations
4.1
Trials at a few ATC sites could establish whether mis-counting or misclassification is occurring in either manual counts or ATCs, and help to
improve instructions for surveyors, and their training.
Annex 8.6
Built Up/ Non Built Up Road Traffic and Road Length statistics:
There was an unexpectedly high demand from Users for the Road Traffic (RT)
and Road Length (RL) statistics to be split by the Built Up/Non-Built Up
(BU/NBU) definitions.
These geographic areas are defined by local speed limits. BU represents
roads that have a speed limit up to and including 40mph. NBU represents
roads that have a speed limit of over 40mph. In the past these areas were the
closest available representation for what we would now call Urban and Rural
areas. However Local Authorities generally set speed limits for road safety
reasons and the BU/NBU areas are no longer an adequate approximation of
Urban and Rural areas. For example, when an LA sets a 40mph speed limit
on a rural road due to safety concerns, in terms of the BU/NBU definition this
rural road now represents a built up area. A fuller explanation is given in the
Glossaries of the Quarterly and Annual Road Traffic Statistics Bulletins.
Added to this DfT does not have a comprehensive picture of the speed limits
in Great Britain. Local Authorities set and change speed limits as necessary
and no statutory obligation exists to inform a central point of these changes. It
has proved extremely difficult to collect these data and as yet no solution has
been brokered between Local Authorities, Ordinance Survey and DfT. Thus
SR2 estimates of BU/NBU are based on incomplete data. Consequently this
is an area needing improvement.
Furthermore, when new links are added to the roads network they are
assigned a BU/NBU status (solely for the purpose of providing these data to
Local Government Finance division and SR3) which is decided on the basis of
that link's Rural or Urban classification. (Rural = NBU, Urban = BU). There are
a few exceptions to this rule where judgment plays a role, but in no cases are
speed limits used to assign the BU/NBU status to the new link. This is
because SR2 has no access to speed limit information. Although this is SR2's
best estimate this clearly provides inconsistencies in the BU/NBU data.
To remedy this DfT moved to a different way of defining Urban and Rural
areas in 2002. The new definitions are based on Urban and Rural polygons.
The urban area polygons used SR2 are based on the Office for National
Statistics (ONS) 2001 urban area polygons, an estimated population based on
the 1991 census, and Land Use Change Statistics. When the polygons were
first used in 2002, the latest ONS urban area polygons were not available.
For consistency SR2 has continued using the approximated polygons even
though the ONS polygons were published in 2004. It has been recommended
that SR2 move to the latest version of the ONS Urban/Rural polygons. This
should happen at the same time as other revisions to methodology to
minimise disruption.
SR2 no longer publish the RT and RL statistics split by BU/NBU, the
published statistics are split by the new Urban/ Rural definitions. The only
data that we supply split by BU/NBU is to the Local Government Finance
(LGF) branch in the Office of the Department for Communities and Local
Government (DCLG) and Road Maintenance Statistics branch (SR3), DfT.
LGF use these, and other data, to determine what amount of funding Local
Authorities' receive.
SR3 use road lengths data split by built-up and non built-up areas in order to
continue providing the National Road Maintenance Condition Survey
(NRMCS) data on a comparable basis with earlier years. The roads defects
index based on the visual survey has been running on more or less the same
basis since 1977, and is currently disaggregated by road class and whether or
not the roads are in built-up areas.
Speed limits have been considered useful for splitting the NRMCS analysis as
the required standards of road condition will vary with the speed of the road.
However, as the current survey is due to be discontinued in the next two
years and replaced by a new data source, this will be an opportune time for
SR3 to consider whether there is now a more appropriate means of
disaggregating the data. This will be considered as part of their forthcoming
user consultation.
From further investigation, it appears that demand for BU/NBU road lengths is
much less than initially suggested by the User Consultation exercise.
It is acknowledged that there is a well-recognised demand for traffic data split
by speed-limit, especially for comparing with accident data. However, at the
moment the BU/NBU split is the only basis for such analyses.
Comprehensive speed-limit data may be available through Ordnance Survey's
Integrated Transport Network in the future.
Proposed Way Forward: In view of the poor accuracy of the BU/NBU split, it
is suggested that SR2 ceases to supply these data to users. However, before
a final decision is made, further consultation should take place with LGF, SR3
and other users, taking particular account of the road safety interests in traffic
levels in the different speed-limit bands.
Annex 10.1
A COMPARISON OF THE COSTS OF MANUAL AND AUTOMATIC
TRAFFIC COUNTS
Dr John Disney
Nottingham Business School
Nottingham Trent University
Introduction
The purpose of this paper is to compare the cost of performing manual (MTC)
and automatic (ATC) traffic counts. The calculations in this paper are based
upon data supplied in March 2005 by DfT and Leeds City Council. This
comparison has been performed to assist the RTS QR in reaching a
conclusion as to the medium term use of MTC’s and ATC’s (from 2007 at the
earliest), given that the DfT is committed to the current level of MTC’s for
2005 and 2006. This implies that additional ATC’s will not be possible in 2005
or 2006 as the overall budget is constrained.
Manual Traffic Counts
This is the familiar, “tried and tested” method of data collection but there is a
danger that familiarity is breeding complacency and it may well be time for
MTC’s to follow Betamax VCR’s and Routemaster buses into the annals of
history. MTCs are relatively expensive to undertake with the average cost of
the 8447 counts in England being £308 for a 12 hour count. On the quiter
roads (Categories 5 Fairly Quiet Road and 6 Quiet Road), which comprise
74% of the total counts in England, this is a potentially high cost per vehicle
surveyed, whilst there is the additional problem that enumerators in such
circumstances are prone to distractions and loss of concentration. Even if the
enumerators’ readings are accurate, there is the possibility of data recording
errors when hand written data (often itself written in poor weather conditions)
is inputted into Excel spreadsheets before submission to the Department.
Hand held electronic counters would only prove economically viable to
contractors undertaking a large number of MTC’s.
Furthermore there is an assumption that the counts taken in “neutral weeks”
are genuinely representative. With lifestyle changes, including the major
effects on working patterns generated by ICT developments, and widely
varying school holiday patterns (with proposals for further changes over the
coming years), this assumption is open to challenge and is becoming more
difficult to defend without further research.
There is also the likelihood that MTC costs will continue to increase faster
than the RPI as they are based upon labour costs around the “Minimum Wage
Threshold” together with private car user travel expenses. Both of these are
likely to continue to increase above the rate of inflation given the competition
for labour at these rates. Some tenderers, especially local authorities keen to
retain traffic surveys “in house”, are actually tendering at loss leading rates
especially where enumerators are required to travel even moderate distances,
such as 10 miles each way from home to the survey site.
This situation is clearly unsustainable.
Automatic Traffic Counts
To date, the major problems with these have been accuracy of classification,
capital costs of installation and on-going maintenance costs. However new
developments such as Metrocount (in the short term), ANPR and GIS /
Satellite Monitoring (longer term) are likely to overcome the majority of these
problems, especially on minor roads, providing accuracy (after fine tuning of
the algorithms) to match that of current MTC’s.
Such ATC’s are inconspicuous (therefore less prone to vandalism / sabotage)
and easy to install on minor roads in one day (often without the need for a
road closure). They only need to be visited approximately every six weeks to
change the battery and download the data. The 190 current ATC sites
provide the continuous data required by the Department for calculation of
expansion factors and monitoring of overall traffic volumes. Therefore, since
continuous monitoring of most minor roads is unnecessary, such ATC’s could
actually be rotated around different sites at 1 to 4 week intervals.
The advantages of ATC’s are numerous, not least of which is the ability to
collect data over a much longer period of time. However, this has a major
disadvantage in that the volume of data collected would be huge – the DfT
Horizons Research Programme is currently commissioning research into
handling such large data sets.
ATCs do not necessarily substantially reduce the total employee hours
involved in Road Traffic Statistics but transfer these from roadside data
collection (a tedious outdoor occupation usually undertaken by part time
employees) and mundane data inputting to data processing and analysis. This
is in line with changes experienced by many occupations and professions as
ICT advances have a greater impact upon the world of work.
Recommendations
1. Undertake trials of new generation ATCs to test accuracy against MTC’s
and actual traffic flows (possibly by use of video cameras) on minor roads.
2. Undertake research to compare estimated traffic flows from 12 hour MTC’s
with actual flows monitored by ATC’s for minor roads and investigate any
differences found.
3. Undertake research to determine the optimum duration and timing of ATC’s
on minor roads. This can then be used to determine rotational patterns
between sites and the number of ATC’s required to service the desired
number of monitoring sites on minor roads.
4. Investigate the possibility of extending such ATC’s to Category 4 (Ordinary
Roads) and possibly Category 3 (Quite busy roads). This would cover up to
90% of current MTC’s undertaken each year.
5. Research the longer term feasibility of both ANPR and GIS / Satellite
Monitoring. Current evidence suggests that GIS / Satellite Monitoring is likely
to prove to be the most feasible method especially if it is also used to facilitate
any road charging schemes which the Department may introduce.
Conclusions
The cost of the current 3919 minor road ATC’s in England is approx £960k,
leading to a conservative estimated cost of £5m for the next five years (200711) although inflationary pressures described above could see this approach
£6m. This provides just 12 hours of data per site per annum. By comparison
£4m would provide one week’s data per annum at 1000 minor road sites,
allowing for 10% inflation each year in maintenance / rotation costs.
This is a strong argument to move away from MTC’s to new generation ATC’s
in the short term.
Dr John Disney
April 17 2005