QA/QC plan for air quality data
collecting under the
Exchange of Information decision
Discussion paper
November 2004
Ed Buijsman
Frank de Leeuw
Patrick van Hooydonk
Peter de Smet
Wim Mol
The European Topic Centre on Air and Climate Change (ETC/ACC)
is a consortium of European institutes under contract of the European Environmental Agency
RIVM UBA-B UBA-V IIASA NILU AEAT AUTh CHMI DNMI NTUA ÖKO SHMU TNO
QA/QC in AIRBASE
On the front page examples of monitoring stations included in AIRBASE.
Upper left: Urban background station Aberdeen (EoI code GB0729A).
Picture taken from http://www.stanger.co.uk/siteinfo/MonitoringSite.asp?ID=41
Upper right: Street station 45R501-Charleroi (EoI code BE0235A).
Picture taken from: http://www.irceline.be/~celinair/dutch/homenrl_java.html
Bottom right: Rural background station Schauinsland (EoI code DE0003R).
Picture taken from:http://www.nilu.no/NILUwebhtmlsider/ilab/EMEP/germany/default.htm,
see for further information on this station also:
http://www.env-it.de/stationen/dispatcher?event=STATION_SELECT&StationCode= DEUB004t
Bottom left: Rural background station De Zilk (EoI code NL0250A).
Author affiliation
Ed Buijsman, Frank de Leeuw, Patrick van Hooydonk, Peter de Smet and Wim Mol are all at the
National Institute for Public Health and the Environment, Bilthoven, The Netherlands.
Disclaimer. This discussion paper has not been subjected to EEA Member State review.
It does not represent the formal views of the European Environment Agency.
© ETC/ACC, 2004.
European Topic Centre on Air and Climate Change
PO Box 1
3720 BA Bilthoven
The Netherlands
Phone +31 30 2743550
Fax +31 30 2744433
Email etcacc@rivm.nl
Website http://air-climate.eionet.eu.int/
page 2 of 15
QA/QC in AIRBASE
page 3 of 15
Contents
1.
Introduction
4
2.
The data exchange process
5
3.
4.
5.
2.1 Information streams
5
2.2 Responsibility of partners involved
5
2.3 The annual data transmission cycle
6
Current procedure for verification of raw data and meta information
8
3.1 Meta information
8
3.2 Air quality data
8
3.3 Ad-hoc actions
8
Actions to improve the quality of AirBase
9
4.1 Correction factors for PM10 measurement results
9
4.2 Filling the gaps
9
4.3 Spatial coverage
9
4.4 “Phantom” stations
9
4.5 Classification of stations
10
4.6 Quality flags
10
4.7 In depth review of (meta) data quality
11
Planning
12
References
13
Annex A Exchange of Information
14
QA/QC in AIRBASE
page 4 of 15
1. INTRODUCTION
Within the European Union a procedure to exchange information on air quality has
been laid down in Council Decision 97/101/EC (EU, 1997) of 27 January 1997
establishing a reciprocal exchange of information and data from networks and
individual stations measuring ambient air pollution within the Member States (the
“Exchange of Information decision” or shortly EoI decision). Technical details such as
pollutants covered, information on stations and networks to be communicated are
specified in the annexes of the Decision. These annexes were amended by
Commission Decision 2001/752/EC (EU, 2001). Formally the EoI applies only to the
EU Member States, however, upon request of the European Environmental Agency, a
large number of non-EU countries provides, on a voluntary basis, information on air
quality following the requirements of the EoI Decision. In this way, information on a
pan-European scale has become available.
According to the EoI Decision, the European Commission will, each year, prepare a
technical report on meta information and air quality data exchanged, and make the
information available to Member States in a database. The decision also states that
the Commission will call upon the European Environment Agency (EEA) with regard
to the operation and practical implementation of the information system. The
European Topic Centre on Air Quality and Climate Change (ETC/ACC), under
contract to EEA, manages the database system, AIRBASE. The information submitted
under the EoI is stored in AIRBASE and made available to the public on the Internet
via the ETC/ACC website1. The technical report mentioned above is prepared
annually by the ETC/ACC and made available at the website2 (Buijsman et al., 2003)
This paper describes the status of the air quality data flow connected to the EoI
Decision and presents a QA/QC plan for improving the quality of the information
stored in AIRBASE.
1
2
http://air-climate.eionet.eu.int/databases/airbase.html
http://air-climate.eionet.eu.int/reports
QA/QC in AIRBASE
page 5 of 15
2. THE DATA EXCHANGE PROCESS
2.1 Information streams
Before 1 October, EU Member States, member countries of the EEA and, on a
voluntary basis, other European countries, have to submit information on air quality
to the Commission (DGEnv) and/or the EEA. The ETC/ACC facilitates in this process
by providing every year a DEM (Data Exchange Module) in which the requested meta
information and raw air quality data can easily be handled.
The Exchange of Information Decision requires a large set of meta information to be
delivered to the Commission (EU, 1997; 2001). Part of this information is mandatory
and the other items are to be delivered to the Commission ‘to the extent possible’ and
‘as much information as feasible should be supplied’. Table 1 gives a brief summary of
the requested meta-information; Annex A gives a more detailed overview of the
mandatory and non-mandatory items.
All information submitted under the EoI is stored in AIRBASE and made available to
the public on the internet via the ETC/ACC website.
Table 1 Summary of information on measurement stations, measurement techniques and data quality
to be delivered under the Exchange of Information (EoI) (EU, 1997)
Type of information
Examples
Networks
Organisation
Name of the network, geographical coverage, responsible authorities,
organisational information
Stations
General
Name and location of measurement site, type of station
Local surroundings
Type of surroundings
Emission sources
Sources which might influence local air quality
Traffic characterisation
For traffic stations only: traffic density
Measurement configurations
Measurement technique
Analytical method, sampling characteristics, time resolution, calibration
Validation procedure
Procedures, criteria, data quality, data coverage
2.2 Responsibility of partners involved
The responsibilities of the EU Member States – with the extension of the EoI to a
pan-European level this refers more generally to the national data supplier – are
given in the EoI. In relation to the annual data transmission process, the member
country is responsible for a timely delivery of the requested information in one of the
agreed data formats and for the quality of raw data and meta information. All
transmitted data are deemed to be valid except when explicitly indicated with an
error code.
In relation to the data transmission process, the Commission is responsible for:
QA/QC in AIRBASE
page 6 of 15
 the specification of the technical procedure for the transfer of data;
 loading the transmitted data in a data base and making the updated results available
to the Member States;
 the transfer of selected data - in agreement with the Member States - to
international bodies.
Following the EoI the Commission has called upon the European Environmental
Agency for assistance with regards to the operation and practical implementation of
the information system, that is, the first two bullets mentioned above. For this
purpose the ETC/ACC has developed a software tool known as the Data Exchange
Module (DEM) for information exchange and the central data base AIRBASE which
includes a web-based retrieval system.
Other responsibility with regards to reporting and reviewing of the EoI fall outside
the scope of this paper and will not be discussed.
Including the transferred responsibilities with regards to the operation and practical
implementation of the information system, the EEA and its Topic Centre are
responsible for:
 the specification of the technical procedure for the transfer of data;
 loading the transmitted data in a data base and making the updated results available
to the Member States;
 making the information accessible to the public.
2.3 The annual data transmission cycle
The software data exchange tool DEM will be pre-filled with all meta-information on
known EoI, EuroAirnet or EMEP station and distributed before 15 July to the
participating countries. To support the contributing countries, the ETC/ACC operates
a helpdesk for questions directly and specifically related to the current EoI data
delivery and the usage of DEM.
Before 1 October the participating countries submit meta-information and raw air
quality data in one of the agreed formats (see EU, 2001) to the Commission and/or to
EEA. Receipt of the data is acknowledged by the ETC.
Before 1 April data which has been received in one of the agreed formats, will be
uploaded in AIRBASE. When uploading the received data into AIRBASE a test is
performed to avoid overwriting of data already stored in AIRBASE. In this case, the
data supplier is informed and is asked to confirm the changes.
The ETC/ACC is responsible for a correct transfer of data into AIRBASE. Updated
copies of AIRBASE will regularly be made available on the web.
AIRBASE contains both raw data and aggregated and statistical data. To ensure
consistency between the raw (hourly) data and aggregated data (daily means,
statistical parameters) all aggregations and the calculation of a standard set of annual
statistical parameters will be done under the responsibility of the ETC-. Annex B
gives an overview of the statistics calculated by ETC/ACC and the procedures
followed. If a Member States has submitted aggregated data, these will be compared
with the ETC/ACC results. In case deviations between statistical parameters directly
submitted by the data supplier and those calculated by ETC/ACC are larger than
expected regarding the numerical approaches (that is, more than a few %), the data
supplier will be contacted. Note that in AIRBASE only the data calculated by ETC/ACC
will be stored.
The ETC/ACC will not make any modification in raw data and/or meta information
without written permission of the owner of the data (that is, the data supplier). In case
ambiguous information has been received, the data supplier will be contacted and
asked to check and – if needed - to correct the information. As long as no reply of the
QA/QC in AIRBASE
page 7 of 15
data supplier has been received, ambiguous information is flagged and made
inaccessible by external users of AIRBASE. In case the data supplier of an EU Member
States decide that raw data has to be deleted from AIRBASE, DGEnv is informed.
QA/QC in AIRBASE
page 8 of 15
3. CURRENT PROCEDURE FOR VERIFICATION OF RAW
DATA AND META INFORMATION
3.1 Meta information
When AIRBASE information is used for air quality assessments, problems may arise if
the (meta) information delivered is incomplete. Although the EoI distinguishes only
mandatory and non-mandatory information (see Annex A), from an interpretative
point of view three different cases can be defined:
1. According to the EoI data suppliers are obliged to provide information on a
number of items (see Table A.1). The ETC/ACC evaluates existing and newly
submitted information on these mandatory items. In case an item is missing (or is
possibly in error) the data supplier will be demanded for a correction.
2. Lack of certain types of meta information makes the appropriate use of
measurement results questionable or diminishes the value of measurement
results. From a user perspective, therefore, ‘essential meta information’ can be
defined (Buijsman et al., 2003). In other words, what kind of meta information is
needed by a user for adequate application and analysis of data considered. Meta
information types considered as ‘essential’ are:
 Station characteristics in terms of type of station and area.
 Geographical co-ordinates and altitude.
 Station name.
If one of these items is missing or is possibly in error, the data supplier will be
urgently requested for updating. Note that this “essential” set is largely covered by
the mandatory items; only the type of station is additionally included.
3. The EoI requests on a voluntary basis, detailed information on station and
measuring methods (Table A.2). With the exception of PM10 (and PM2.5)
measuring methods (see section 4.1), the ETC/ACC will not perform a systematic
review on completeness and correctness of these items. However, in case a
(possible) error is observed, the data supplier will be contacted.
3.2 Air quality data
According to the EoI all transmitted data is deemed to be valid. Member states have
to ensure that a quality assurance procedure is in place which meets in general the
objectives of the EoI and the AQ daughter directives. Therefore, the ETC/ACC will
not implement any validation procedure. However, the DEM provides tools for
screening the (meta) data (such as mapping of stations, visualisation of time series,
detection of outliers). Before loading new data in AIRBASE, the ETC/ACC will
systematically screen the AQ data for outliers.
3.3 Ad-hoc actions
It is nearly unavoidable that, while analysing the data in AIRBASE either by external
users or by the ETC, sometimes questionable data is found. When this is the case, the
ETC/ACC will inform the data supplier and request for clarification and/or
correction. If the data supplier decides that data has to be deleted, DGEnv is
informed. In case a time series is replaced by corrected data, all statistical values and
aggregated data will be re-calculated. The ETC/ACC strongly encourage all AIRBASEusers to contact the ETC/ACC in case questionable data is detected.
QA/QC in AIRBASE
page 9 of 15
4. ACTIONS TO IMPROVE THE QUALITY OF AIRBASE
Various actions to improve the quality of the information in AIRBASE can be defined.
In contrast to the on-going activities described in section 3, these concern mainly
non-recurring activities. Depending on available resources a number of such actions
could be undertaken. These will be discussed below.
4.1 Correction factors for PM10 measurement results
Submission of information on the measurement method is a non-mandatory item
(see Annex A, Table A.2). However in case of PM10, intercomparison exercises of
different PM10 monitoring methods showed large deviations from the European
Reference Method (as described in CEN standard EN 12341). It was recognised that
results from non-reference methods have to be corrected and enhanced efforts to
harmonise the PM-measurements in the EU are needed. (EU, 2003). Information on
PM10 and PM2.5 measuring methods will be screened on completeness and
correctness. So far no provisions have been made in the EoI to submit information on
correction factors which have been applied by countries to the results from
automated instruments. ETC/ACC will make necessary changes in DEM and AIRBASE
to accommodate the time- and location depending correction factors.
This action is in progress.
4.2 Filling the gaps
In AIRBASE substantial data gaps exist for the period up to 1998 which hampers a
European –wide assessment of systematic changes in air quality over the last 10-20
year. In line with Article 5(4) Member States have to take their responsibility and
should submit historical data for the period 1989-1998 where available. Emphasis
will be laid on the components of the daughter directives. The ETC/ACC will remind
the countries to this obligation.
4.3 Spatial coverage
So far, it is not clear whether the spatial coverage of stations in AIRBASE is complete.
From a legal point of view, the completeness is defined in article 3 of the EoI which
states that data for all stations used from the implementation of the air quality
directive have to be transmitted. As the first reports under the directives
1999/30/EC, 2000/69/EC and 2002/3/EC became recently available, only now a
check on completeness can be made. From an user perspective, current spatial
coverage is not sufficient for a reliable mapping of air quality on a pan-European
scale. Especially in the eastern parts of Europe a higher station density would be
needed. Actions for a further integration of information from EuroAirNet stations
and/or AQ-networks which are more oriented towards public health should be
undertaken.
4.4 “Phantom” stations
In AIRBASE a number of “phantom” stations, i.e. stations for which (in general
incomplete) meta information is available but for which no AQ data has been
submitted, are stored. A list of phantom stations has been sent to the data suppliers
with a request to indicate whether AQ data for this station is available and will be
submitted to AIRBASE or the station can be deleted from AIRBASE.
QA/QC in AIRBASE
page 10 of 15
4.5 Classification of stations
For assessment of urban air quality and population exposure it is essential to know in
which city or agglomeration all stations classified as urban or suburban are located.
For urban stations the ETC/ACC will review the information and contact the data
supplier in case of missing or questionable data. AIRBASE contains data from a large
number of urban monitoring sites 3. By definition, these stations have a low (urban
background) or very low (traffic) spatial representativeness (Larssen et al., 1999). On
the other hand, there is a strong need to gain insight in urban air quality and its
development over time. Given the low spatial representativeness, it is not
straightforward to make general statements on urban air quality based on the results
of urban air quality monitoring stations. The EEA presents annual indicators on
urban air quality. Nevertheless, there have been some discussions on the validity of
the procedures used to produce these indicators. Improvements are certainly
conceivable. First of all, there should be a better understanding of the spatial
representativeness of urban air quality monitoring stations. AIRBASE itself contains
(probably) the information to investigate this.
For certain assessments, information on the type of station is essential for selecting
the appropriate set of stations. However, the criteria used to define the different type
of stations leaves - necessarily - room for different interpretations. There is also the
possibility that certain stations will be influenced by near-by sources. The type of
station should then be defined on the basis of the predominant sources influencing
the station. Recent research has shown the presence of possibilities to define the type
of station using the measurement data themselves (Snel, 2004). Although further
research is certainly needed the method seems promising. The method requires, for
example, nitrogen monoxide (NO) data as input. Unfortunately, only few countries
deliver these data. Most networks perform measurements on an hourly basis. Current
technology allows that in these cases, measurement results for nitrogen dioxide
(NO2), as well as for nitrogen monoxide and nitrogen oxides (NOx), can easily be
made available. From the point of view of quality control the suggestion has been
made to include nitrogen oxide in the regular data delivery.
Apart from the meta information items mentioned under § 3.1, it is advisable to
expand this with the name of the city in which the station is located.
4.6 Quality flags
Measurement data in AIRBASE originate form many different monitoring networks
throughout Europe and are subject to different QA/QC regimes. Larssen et al. (1999)
proposed a set of QA/QC criteria for classification and subsequent selection of air
quality monitoring stations to be included in EUROAIRNET. These QA/QC criteria can
be used in AIRBASE as well to flag data. By doing so, users can select data from
AIRBASE according to their own quality criteria or intended quality of their product.
We propose a slightly modified QA/QC criteria scheme (Larssen et al. (1999)) in
which we consider the accreditation as the most important quality indicator:
1.
Network with stations that are part of a national or local air quality
monitoring network. A complete and accredited QA/QC plan implemented in the
network is the key feature of this quality level.
2.
Network with stations that are part of a national or a local air quality
monitoring network. The work has a complete QA/QC plan implemented.
3.
Individually operated networks or stations (or even a national network)
implementing a minimum QA/QC plan.
In 2002, the number of urban stations ranged from 28 for PM2.5 to 1206 for nitrogen dioxide
(Buijsman et al., 2004).
3
QA/QC in AIRBASE
4.
page 11 of 15
Includes networks or stations without a documented QA/QC plan.
Data suppliers will be asked to classify their networks according to the scheme above.
This information will be included in AIRBASE and also in the next report on meta
information.
4.7 In depth review of (meta) data quality
Despite all procedures, it can not be excluded that AIRBASE contains faulty
information. This holds especially for older data, as these data have been subject to
less stringent quality control procedures. Therefore, it is proposed to perform an in
depth review of the quality of (meta) data in AIRBASE. The most convenient way is to
do this on a country level. A preliminary review of data from Belgium and The
Netherlands revealed some minor discrepancies or inconsistencies.
As another example, experience has shown that station coordinates sometimes are
incorrect, thereby hampering the proper use of air quality data of these stations.
Action will be taken to check the coordinates of all stations present in AIRBASE.
In addition to the quality control procedures for submitted data that has been
described in § 3.2, more extensive procedures will be implemented. These procedures
will be based on the results of research by the ETC/ACC partner CHMI.
QA/QC in AIRBASE
5. PLANNING
Potential activities discussed in the previous section ar summarzied in table 2.
Prioritizing of these actions is currently under discussion.
Table 2 Summary of actions and their status
Action
Discussion paper on correction factors for PM10 measurement results in AIRBASE
Filling gaps in time series
Reserach on spatial coverage
▫ Analysis on AB data
▫ Organization of extension of networks
Exclude ‘phantom stations’ from AIRBASE
Review of system for classification of stations
▫ Definition
▫ Additional data
Development of quality flagging system
In depth review of (meta) data quality
Implementation of CHMI data quality control procedures
page 12 of 15
QA/QC in AIRBASE
page 13 of 15
References
Buijsman E., van Hooydonk P.R., Mol W., Cernovsky L., de Leeuw F.A.A.M. 2004
European Exchange of air quality monitoring meta information in 2002.
ETC/ACC Technical paper 2004/1. ETC/ACC, Bilthoven, the Netherlands.
EU. 1997. Council Decision of 27 January 1997 establishing a reciprocal exchange of
information and data from networks and individual stations measuring
ambient air pollution within the Member States (97/101/EC). Official Journal
L 035, 05/02/1997, pp. 14-22.
EU. 1999. Council Directive 1999/30/EC of 22 April 1999 relating to limit values for
sulphur dioxide, nitrogen dioxide and oxides of nitrogen, particulate matter
and lead in ambient air. Official Journal L 163, 29/06/1999, 41-60.
EU. 2000. Directive 2000/69/EC of the European Parliament and of the Council of
16 November 2000 relating to limit values for benzene and carbon monoxide
in ambient air. Official Journal L 313, 13/12/2000, pp. 12-21.
EU. 2001. Commission Decision of 17 October 2001 amending the Annexes to
Council Decision 97/101/EC establishing a reciprocal exchange of information
and data from networks and individual stations measuring ambient air
pollution within the Member States (2001/752/EC). Official Journal L 282,
26/10/2001, pp. 69–76.
EU. 2002 Commission Decision of 12 February 2002 relating to ozone in ambient air
(2002/3/EC). Official Journal L 67, 09/03/2002, pp 14-30.
EU. 2003. Second position paper on particulate matter, draft for discussion. CaFE
working group on particulate matter, 20 August 2003.
Garber W, Colosio J, Grittner S, Larssen S, Rasse D, Schneider J, Houssiau M. 2002.
Guidance on the Annexes to Decision 97/101/EC on Exchange of Information
as revised by Decision 2001/752/EC. European Commission, DG
Environment, Brussels.
Larssen S, Sluyter R, Helmis C. 1999. Criteria for EUROAIRNET. The EEA Air
Quality and Information Network. Technical Report No. 12. European
Environment Agency, Copenhagen.
Leeuw F de, Buijsman E, Hooydonk P van, Smet P de, Mol W. 2003. QA/QC plan for
air quality data collecting under the Exchange of Information decision. First
draft. European Topic Centre on Air and Climate Change, Bilthoven, The
Netherlands.
Snel S. 2004. Improvement of classifications European monitoring stations for
AirBase. Internal report National Institute for Public Health and the
Environment, Bilthoven, The Netherlands.
QA/QC in AIRBASE
page 14 of 15
ANNEX A EXCHANGE OF INFORMATION
The Exchange of Information Decision requires a large set of meta information to be delivered
to the Commission (EU, 1997, 2001). Part of this information is mandatory (Table A.1) and
the other items (Table A.2) are to be delivered to the Commission ‘to the extent possible’ and
‘as much information as feasible should be supplied”.
Table A.1 Overview of mandatory meta information to be delivered under the Exchange of
Information (EoI)
Item a
Description
I.1.
Name of the network
I.4.1.
Name of the body responsible for network management
I.4.2.
Name of person responsible
I.4.3.
Address
I.4.4.
Telephone and fax numbers
I.5.
Time reference basis
II.1.1.
Name of the station
II.1.4.
Station code given under the present decision and to be provided by the Commission
II.1.8.
Geographical co-ordinates
II.1.10.
Pollutants measured
II.1.11.
Meteorological parameters measured
II.2.1.
Type of area
(a) Numbers according to Annex II of the EoI (EU, 2001)
QA/QC in AIRBASE
page 15 of 15
Table A.2. Overview of non-mandatory meta information to be delivered under the Exchange of
Information (EoI)
Item a
Description
I.2.
Abbreviation (of the network)
I.3.
Type of networks
I.4.5.
E-mail (of the body responsible for the network)
I.4.6.
Website address
II.1.2.
Name of the town/city of location (of the station)
II.1.3.
National and/or local reference number or code
II.1.5.
Name of technical body responsible for the station
II.1.6.
Bodies or programmes to which data are reported
II.1.7.
Monitoring objectives
II.1.9.
NUTS level IV
II.1.12
Other relevant information
II.2.2.
Type of station in relation to dominant emission sources
II.2.3.
Additional information about the station
III.1.1.
Name (of measurement equipment)
III.1.2.
Analytical principle or measurement method
III.2.1.
Location of sampling point
III.2.2
Height of sampling point
III.2.3
Result-integrating time
III.2.4
Sampling time
(a) Numbers according to the Annex II of the Exchange of Information (EU, 2001).