Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
1
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
Structure
1 Objective and Scope
2 Format and Philosophy of the Workshop
3 Some Discussion Points
4 Schedule
5 Abstracts
6 Posters
7 Epilog
2
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
3
1 Objective and Scope
Runoff generation is considered as one of the most important processes in catchment
hydrology. The correct separation of the effective precipitation into different runoff
components is crucial for the modelling of transport processes of solutes and
contaminants. The water quality depends on the varying contributions from surface
runoff, soil water and groundwater. In recent years tracer methods combined with
hydrometric measurements have proved to be suitable to identify runoff generation
mechanism.
Many processes are still not well understood in the hydrological micro scale. However, a
challenge in future will be an adequate modelling of the dominating runoff generation
processes in the hydrological meso- and macro scale, in order to meet the needs of
future water resources management. Therefore an improved process oriented modelling
is needed. The model uncertainties have to be minimized by developing new and
independent calibration and validation techniques (e.g. multi-response validation). The
objective of the workshop is to bridge the gap between field based experimental
research and hydrological basin modelling.
2 Format and Philosophy of the Workshop
The four day workshop program includes the oral presentation of papers, a half-day
excursion and discussions in smaller groups. Oral presentations are scheduled for 20
minutes (including 5 minutes discussion), poster-paper presentations are scheduled for
5 minutes. In addition, a poster session will be organised. The most important part of the
workshop will be the discussion that take place not only in the meeting room but also in
the breaks, around the poster and within the discussion rounds in smaller groups.
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
4
3 Some Discussion Points
While preparing the workshop the following research questions arose (please, add some
more of your own):
1. How can we isolate different runoff response mechanisms?
2. How can preferential flow behaviour be parameterized at the catchment scale?
3. What is the role of antecedent wetness on stormflow response?
4. Can we reconcile the different view on the contribution of event water to the direct
hydrograph of tracer and hydrometric studies?
5. How can we include soil/rock volumes for explicit water and tracer mass balance?
6. How can we use further hydrometric measurements (snow cover, soil moisture,
groundwater levels etc.) for model calibration / validation?
7. What is the minimal set of measurements necessary to characterize a hydrological
system?
8. Are isotope and hydrochemical tracers still useful when the spatial and temporal
variability of the end members is considered (and resulting uncertainty)?
9. How do we construct model reservoirs that capture dominant flow response but with
a minimal number of tuneable parameters?
10. How can we use tracers for model calibration / validation?
11. How do we rank importance of different hydrologic response units in terms of flow
and chemistry?
12. Can a catchment be discretized into a simple set of response units?
13. How can a better understanding of runoff generation mechanism in smaller scales be
transferred to an improved river basin management and policy?
14. How do we scale-up our process knowledge from point scale to headwater and then
meso-scale catchments?
15. If raster cells, which cell size is suitable for process adequate modelling?
16. What is the big difference between the well investigated micro-scale headwaters and
the meso-scale catchments?
17. How do we classifiy different headwater catchments?
18. Should we conduct joint field campaigns at a single site (like the atmospheric
scientists) to capture all scales of behaviour?
19. The discussions during the workshop, especially the small group discussions will
address some of these questions.
20. …..
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
5
4 Schedule
4.1. Overview
The workshop will be held on 9-12 October 2000, at the Institute of Hydrology, University
of Freiburg. The sessions will take place at the lecture hall "Fahnenbergplatz”, which is in
the basement of the building.
Timetable:
Time
MORNING
Monday
Tuesday
Wednesday
Thursday
Registration
8.30-10.00
Session 2_I
9.00-11.00
Session 3_II
9.00-10.20
Session 4_I
9.00-10.40
Coffee Break
Coffee Break
Coffee Break
Coffee Break
Session 1_I
11.00-12.40
Poster Session
11.30-13.00
Session 3_III
11.00-13.00
Lunch
Lunch
Lunch
Small Group
Discussion
11.10-13-00
Lunch
Session 1_II
14.00-15.35
Session 3_I
14.00-15.30
Excursion to the
Dreisam Research
Basin
14.00-18.00
Coffee Break
Coffee Break
Session 1_III
16.00-18.00
Small Group
Discussion
16.00-18.00
Opening Session
10.00-11.00
AFTERNOON
EVENING
Reception
18.00-late
Session 4_II
14.00-14.45
Finale Discussion
Evaluation
15.00-15.30
Conference Dinner
19.00-late
A welcome reception with snacks and beverages from the region takes place on Monday
evening at the Institute of Hydrology.
On Wednesday afternoon a half day excursion will lead to the nearby Dreisam research
basin, where several experimental test sites will be visited.
A short trip through the countrysite (foothills of the Black Forest) will bring us to the
location where the Conference Dinner will be served. The dinner and the wine tasting will
take place in a classical wine cellar in a small village close to Freiburg.
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
6
4.2 Detailed Schedule
DAY 1 (Monday, October 9, 2000)
08:30-10:00
Registration
10:00-10:30
Opening by the Chairman Prof. Dr. Chris Leibundgut
Welcome from the University of Freiburg by the Rector Prof. Dr. Wolfgang Jäger
Opening from the German IHP/OHP by Prof. Dr. Hofius
10:30-11:00
Coffee Break
Session 1:
Physical Processes and Mechanisms of Stream Flow
Generation at the Catchment Scale
Session Chair:
Chris Leibundgut
11:00-11:20
Introduction to the workshop
Chris Leibundgut, Stefan Uhlenbrook
11:20-11:40
A decision scheme to identify dominant flow processes at the plot-scale for the
evaluation of contributing areas at the catchment-scale
Simon Scherrer, Felix Naef
11:40-12:00
Spatial variability and scale dependent runoff generation
Bernd Diekkrüger, Helge Bormann, Christian Renschler
12:00-12:20
Soil moisture patterns and runoff generation in a small Dartmoor catchment,
Southwest England
Erik Meyles, Andrew Williams, John Dowd, Les Ternan
12:20-12:40
Discussion of the first papers, at the beginning, the chairman is asked to summarize
the essentials
12:40-14:00
Lunch
(Places are reserved in the Stadthotel Kolping, Karlsstr. 7, (5 minutes walk); choice
between one vegetarian and one normal meal; costs are 15,- DM and 18,- DM,
respectively)
Session Chair:
Alfred Becker
14:00-14:20
Space-time patterns of runoff generation in the Löhnersbach catchment
Robert Kirnbauer, Günter Blöschl, Peter Haas, Gabriele Müller, Bruno Merz
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
14:20-14:40
7
Neither macropores nor interflow: generation of spiky discharge peaks in small
forested catchments and its implications on stream water chemistry
Gunnar Lischeid
14:40-15:00
Spatial distribution and temporal behaviour of dry weather flow components in a small
drainage basin
Steffen Möller, Wolfhard Symader
15:00-15:20
The role of infiltration conditions for storm runoff generations at the hillslope and small
catchment scale
Axel Bronstert, Daniel Katzenmaier
15:20-15:25
Episodic rising in nitrate concentration of streamwater by partial dieback of pine forest
in Japan: runoff generation process as a controlling factor of the seasonality
Nobuhito Ohte, Naoko Tokuchi, Masanori Katsuyama, Satoru Hobara, Yuko Asano,
Keisuke Koba
15:25-15:30
Runoff generation process and chemical weathering in small headwater catchments,
Tanakami, Japan
Yuko Asano, Taro Uchida, Nobuhito Ohte
15:30-15:35
Temporal and spatial variation of runoff generation and its influences in the Chilliwack
river basin, British Columbia, Canada
Henry M. Sichingabula
15:35-16:00
Coffee Break
16:00-16:20
The use of cumulative data for characterising hydrological systems
Chris Barnes
16:20-16:40
Integrated approach of investigating and modelling runoff formation considering
information from environmental and artificial tracer experiments
Andreas Herrmann, Matthias Schöniger, Sybille Schumann
16:40-17:00
Runoff generation and scale in a hyperarid region
A.P. Schick, J. Lekach, N. Greenbaum, T. Grodek
17:00-17:05
Percolation Response to Rainfall computed from separated baseflow
Hartmut Wittenberg
17:05-17:10
Soil water variability and the conservation of wet, species-rich grasslands in Devon,
United Kindom
Francien van Soest, Andrew Williams, Robert Parkinson
17:10-17:15
Identification
of
hydrotopes
(Dürreychtal, Black Forest, Germany)
Guido Waldenmeyer, Markus Casper
in
a
small
forested
catchment
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
17:15-17:20
8
Soil moisture variability and landuse in a seasonally arid environment
Williams, A.G., Ternan, J.L., Fitzjohn, C., de Alba, S
Perez-Gonzalez, A.
17:20-18:00
Finale discussion of the session 1
At the beginning, the chairmen are asked to summarize the essentials
18:00-late
Welcome Reception
Snacks and beverages form the region are served at the Institute of Hydrology
st
(building Fahnenbergplatz, 1 floor)
DAY 2 (Tuesday, October 10, 2000)
Session 2:
Runoff Generation
Headwater Scale
Processes
at
the
Hillslope
Session Chair:
Günter Blöschl
9:00-9:20
Conceptual modelling of runoff, groundwater and oxygen-18 in a till hillslope
and
Jan Seibert, Allan Rodhe, Kevin Bishop
9:20-9:40
Towards a better process representation of catchment hydrology in conceptual runoff
modeling
Jan Seibert, Jeff McDonnell
9:40-10:00
Analysis of flowpath dynamics at a forested headwater catchment, Central Japan
Taro Uchida, Yuko Asano, Nobuhito Ohte, Takahisa Mizuyama
10:00-10:20
Experimental analysis of rainfall-runoff events in two small adjacent mountainous
catchments
Christoph Sambale, Gerd Peschke
10:20-10:25
Runoff generation on artificial slopes in a Mediterranean-Continental environment:
Teruel Coalfield, Spain
José-Manuel Nicolau
10:25-10:30
Simulation of water flow in heterogeneous layered hillslope
Mikko Jauhiainen, Tuomo Karvonen
10:30-10:35
Runoff generation processes on Reclaimed Opencast Coal-lands, Wales
M. J. Haigh, M. P. Kilmartin
10:35-10:40
A study on ponding time and infiltration capacity during steady rainfall using a rainfall
simulator
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
9
Kazumasa Fujimura, Yosihisa Ando, Fumio Yamada, Akihiro Nagaiwa
10:40-11:00
Finale discussion of Session 2.
At the beginning, the chairmen are asked to summarize the essentials
11:00-11:30
Coffee Break
11:30-13:00
Poster Session
All poster authors are ask to be available near their poster.
13:00-14:00
Lunch
(Places are reserved in the Stadthotel Kolping, Karlsstr. 7, (5 minutes walk); choice
between one vegetarian and one normal meal; costs are 15,- DM and 18,- DM,
respectively)
Session 3:
Process oriented Modeling at different Scales
Session Chair:
Stefan Uhlenbrook
14:00-14:20
Internal validation of the HBV model
Sten Bergström, Göran Lindström
14:20-14:40
A
knowledge-based
system
to
improve
the
preprocessing
of
distributed
precipitation-runoff models
S. Zimmermann, J. Töpfer, G. Peschke
14:40-15:00
Modeling runoff generation in a forested catchment in Southern Finland
Harri Koivusalo, Teemu Kokkonen, Tuomo Karvonen
15:00-15:20
Integrating tracer studies with semi-distributed models: recent insights from a
sub-arctic catchment in the Cairngorm Mountains, Scotland
Chris Soulsby, Sarah M. Dunn, Rob Malcolm
15:20-15:25
A distributed model of runoff generation in the permafrost regions
L. S. Kuchment, A. N. Gelfan, V. N. Demidov
15:25-15:30
Non-calibrated models for the dry hydrologic regime - from the arid to the
semiarid/mediterranean zone
Jens Lange, Christian Leibundgut, Samar Husary, Marwan A. Hassan, Asher P. Schick
15:30-16:00
Coffee Break
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
16:00-17:30
10
Small Group Discussion
4small groups will discuss at different locations
Group
Chairman
Reporter
BLUE
Jeff McDonnell
Stefan
Uhlenbrook
RED
Günter Blöschl
Jens Lange
GREEN
Alfred Becker
Andreas
Schumann
YELLOW
Felix Naef
Christoph
Sambale
Questions for the Discussion:
17:30-18:00
Chairs and reporters are asked to summarize to the main group. Discussion and
comparison of the outcomes.
(Lecture hall Fahnenbergplatz)
DAY 3 (Wednesday, October 11, 2000)
Session 3:
Process oriented Modeling at different Scales (continued)
Session Chair:
Norman E. Peters
9:00-9:20
Modelling and validation of runoff and its components in Swiss Pre-Alpine and Alpine
catchments
Joachim Gurtz, , Karsten Jasper, Herbert Lang, Massimiliano Zappa, Alexandre
Badoux, Mark Verbunt, Tomas Vitvar
9:20-9:40
Improving the representation of water balance and runoff processes in mesoscale
hydrologic models for water management modelling
Joachim Geyer, David C. Garen, Andreas H. Schumann
9:40-10:00
Towards reduced conceptual rainfall-runoff modelling uncertainty
Thorsten Wagener, Matthew J. Lees, Howard S. Wheater
10:00-10:05
Optimizing land use management - a new approach for hydrologic modelling
Manfred Fink, Clemens Beckstein
10:05-10:10
Mathematical model of runoff formation for small river
Yu. I. Denisov, A. I. Sergeev
10:10-10:15
Simulation of runoff and streamflow on local and regional scales
Heinz-Theo Mengelkamp, Kirsten Warrach, Gerard Kiely
10:15-10:20
Integrated water quality modeling by coupling client-server based DBMS, GIS and the
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
11
solute model WASMOD
Ulrike Bende-Michl
10:20-11:00
Coffee Break
Session Chair:
Gerd Peschke
11:00-11:20
Physically-based rainfall-runoff models; setbacks and solutions
T.H.M. Rientjes, P. Reggiani
11:20-11:40
Extrapolating parameters of a conceptual catchment scale model to different
catchments without calibration
Sarah M. Dunn, Carmen Bernzen, Chris Soulsby, Rob Malcolm, Allan Lilly
11:40-12:00
How far from the modelling optimum? Analysing a process-based catchment model by
Artficial Neural Networks
Gunnar Lischeid, Stefan Uhlenbrook
12:00-12:20
New studies for stream-subsurface hydrologic exchange (bank storage)
Tobias Conradt, Matthias Schöniger
12:20-12:25
Retrieval of Hydrological Parameters from Remote Sensing Data
Volker Hochschild
12:25-12:30
A SVAT-scheme combined with a simple regionalized baseflow separation for large
scale groundwater recharge modelling
Volker Armbruster, Markus Haas, Christian Leibundgut , Siegfried Demuth
12:30-12:35
Operational application of the water balance model LARSIM in the Neckar Basin
M. Bremicker, K. Gerlinger
12:35-12:40
Modelling runoff generation processes for the regions with seasonal snow cover at plot
and catchment scales
Yeugeniyi M. Gusev, Olga N. Nasonova
12:40-13:00
Finale discussion of Session 3.
At the beginning, the chairmen are asked to summarize the essentials
13:00-14:00
Lunch
(Places are reserved in the Stadthotel Kolping, Karlsstr. 7, (5 minutes walk); choice
between one vegetarian and one normal meal; costs are 15,- DM and 18,- DM,
respectively)
14:00-18:00
Excursion to the Dreisam Research Basin,
guided by Stefan Uhlenbrook
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
12
The field trip leads to the nearby Dreisam basin. Several experimental test sites will be
visited. Afterwards a short trip through the countrysite (Black Forest and foothills) will
bring us to the location where the Conference Dinner will be served.
19:00-late
Conference Dinner
The dinner and the wine tasting will take place in a classical wine cellar in a small
village close to Freiburg.
DAY 4 (Thursday, October 12, 2000)
Session 4:
Tracer and Modeling Studies for Investigating Runoff
Generation Processes at Different Scales
Session Chair:
Felix Naef
9:00-9:20
Modeling hydrologic responses in a small forested watershed by a new dynamic
TOPMODEL (Panola Mountain, Georgia, USA)
Norman E. Peters, James Freer, Keith Beven
9:20-9:40
Effects of the variations of groundwater levels on the hydrochemical processes in a
headwater catchment, Japan
Masanori Katsuyama, Nobuhito Ohte
9:40-10:00
Mean
groundwater
residence
times,
subsurface
contact
times
and
runoff
recession times: methodology & applications in the Neversink watershed, Catskill Mts,
New York
T. Vitvar, D. A. Burns, J. McDonnell, G. B. Lawrence, D. M. Wolock
10:00-10:20
Uncertainties of two-, three- and five-component hydrograph separations in a
mountainous catchment
Simon Hoeg, Stefan Uhlenbrook
10:20-10:40
Verification of flow processes in soils with combined sprinkling and dye tracer
experiments
Markus Weiler, Felix Naef
10:40-11:10
Coffee Break
11:10-12:30
Small Group Discussion
4 small groups will discuss at different locations:
Group
BLUE
RED
GREEN
YELLOW
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
Chairman
Reporter
Norman
E. Günter Blöschl
Peters
Markus Weiler
Allan Rodhe
Jeff
McDonnell
Volker
Armbruster
13
Stefan
Uhlenbrook
Bruno Merz
Questions for the Discussion:
12:30-13:00
Chairs and reporters are asked to summarize to the main group. Discussion and
comparison of the outcomes.
(Lecture hall Fahnenbergplatz)
13:00-14:00
Lunch
(Places are reserved in the Stadthotel Kolping, Karlsstr. 7, (5 minutes walk); choice
between one vegetarian and one normal meal; costs are 15,- DM and 18,- DM,
respectively)
Session Chair:
Jeff McDonnell
14:00-14:20
The separation of flow pathways in a sandstone catchment of the Northern Black
Forest using a nested approach
Markus Casper; Erich J. Plate, Jens Mehlhorn, Holger Volkmann, Guido Waldenmeyer
14:20-14:40
Humic substances and agrochemicals in the discharge components of Scheyern
K.-P. Seiler, C. Hellmaier
14:40-14:45
18
Hydrograph separation of selected runoff events using DOC and O as natural tracers
Holger Volkmann, Markus Casper,
14:45-15:00
Finale discussion of Session 4.
At the beginning, the chairmen are asked to summarize the essentials
15:00-15:30
Finale Discussion and Evaluation of the Outcomes
Summary by the chairmen regarding the publication of proceedings in the Freiburger
Schriften zur Hydrologie and the special issue of Hydrological Processes
(final reviews, deadline for the manuscripts, anticipated publication date, etc.)
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
14
5 Abstracts
The abstracts were reviewed thematically by the scientific advisory committee. However the
authors are responsible for the contents.
DAY 1 (Monday, October 9, 2000)
Session 1:
Physical Processes and Mechanisms of Stream Flow
Generation at the Catchment Scale
A decision scheme to identify dominant flow processes at the plot-scale for the
evaluation of contributing areas at the catchment-scale
Simon Scherrer
Horat & Scherrer AG, Hydrology and Flood Defence Strategies, 4103, Bottmingen, Switzerland
Felix Naef
Institute of Hydromechanics and Water Resources Management, Swiss Federal, Institute of Technology, 8093 Zurich,
Switzerland
Artificial rainfall of high intensity was applied to hill-slope plots of 60 m2 to study the mechanisms of runoff
formation during intense rainfall. Different flow processes e.g. infiltration excess or saturation overland flow
and subsurface flow etc. could be observed. Based on experiments and simulations with a two-dimensional
model QSOIL, the conditions and criteria for the occurrence of the different flow processes could be
identified. These criteria and knowledge gained by other field studies were used to develop a decision
scheme to evaluate the dominating flow process at the plot-scale as well as to estimate the hydrological
behaviour of sub-catchments and catchments by delineating contributing areas. The scheme might be used
for practical investigations or for instructional purposes.
Spatial variability and scale dependent runoff generation
Bernd Diekkrüger, Helge Bormann
Geographical Institute, University of Bonn, Meckenheimer Allee 166, 53115 Bonn, Germany
Christian Renschler
USDA – ARS, National Soil Erosion Research Laboratory, Purdue University, West Lafayette, IN 47907-1196, Indiana,
USA
The spatial and temporal variability of input data has a major impact on the calculation of runoff generation. The
information needed to apply sophisticated simulation models is provided by measurements (local scale) or by
standard data sets (regional scale). At the regional scale knowledge about variability is limited by the availability
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
15
of measurements. Therefore in this study the concept of effective parameter and methods for data
disaggregation as well as the uncertainties implied by the different approaches are investigated.
The application of the effective parameter concept seems to depend on the process. While the regional
calculation of evapotranspiration and soil water flux leads to reasonable results, determination of surface
runoff generation is strongly affected by the accuracy and resolution of the boundary conditions. If water flux
simulations are based on a soil map, the parameter uncertainty from applying soil texture classes is
considerably high. Disaggregating a texture class using different strategies may lead to a change in the
predicted dominant runoff generation process. An analysis of the requirements and implications of methods
for temporal rainfall disaggregation shows that the mean calculated long term runoff using disaggregated
rainfall series is similar while single events cannot be predicted as well. The coefficients of correlation
between different runoff simulations for long term simulations are high but the results show a constant bias.
Thus a method which combines the calculation of aggregated soil parameters with an estimation of the
internal variability of the most important parameters for surface runoff generation. Spatial variability and
scale dependent runoff generation needs to be applied.
Soil moisture patterns and runoff generation in a small Dartmoor catchment,
Southwest England
Erik Meyles, Andrew Williams, Les Ternan
Department of Geographical Sciences, University of Plymouth, Drake Circus, Plymouth PL4 8AA, UK
John Dowd
Geology Department, The University of Georgia, Athens, GA 30602-2501, Georgia, US.
The relationship between topsoil moisture patterns and storm runoff was investigated. Volumetric soil
moisture was measured at 151 points on a ten by ten metre grid on 19 occasions. Geostatistical analysis of
the soil moisture grid data showed that in dry conditions the pattern is very heterogeneous, with patches of
dry soil adjacent to wet areas. In wet conditions, the pattern is more uniform as the wet areas become
interconnected. A soil moisture content of about 0.55 to 0.60 divided the heterogeneous and more uniform
conditions. In dry antecedent conditions small rainfall amounts yield little runoff. In wet conditions, rainfall
greater than 20 mm is associated with large increases in runoff. The relationship between volumetric soil
moisture and runoff is exponential. These findings suggest that as landmanagers can affect the soil and
vegetation in an area, grazing management in upland headwaters may considerably influence the
hydrological behaviour.
Space-time patterns of runoff generation in the Löhnersbach catchment
Robert Kirnbauer, Günter Blöschl, Peter Haas, Gabriele Müller
Institute of Hydraulics, Hydrology and Water Resources Management, Vienna University of Technology, Karlsplatz 13/223,
A-1040 Vienna, Austria
Bruno Merz
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
16
GeoForschungsZentrum Potsdam, Telegrafenberg, D-14473 Potsdam, Germany
2
An analysis of runoff observations in the highly instrumented 16 km alpine Löhnersbach catchment, Austria,
indicates that a number of runoff event types exist. These types include a bimodal event hydrograph
consisting of a flashy peak due to saturated overland flow and a delayed peak due to subsurface stormflow
occuring in response to long duration - low intensity storms, and a unimodal flashy event hydrograph
occuring in response to short duration - high intensity storms. One or two particular mechanisms dominate
each of these types while the contribution from other mechanisms is not significant. Runoff generation is
found to be very different between the northwestern and southeastern parts of the Löhnersbach catchment.
During low flows, the former contributes below average to runoff and the latter contributes above average,
however this trend is reversed for high flows and large events. These differences can be interpreted in terms
of the dominating runoff generation mechanisms and are fully consistent with the spatial patterns of dynamic
indices such as stream network density and the presence of saturated zones. However, standard GIS
catchment descriptors such as topographic attributes, landuse and geology do not suffice to explain these
spatial patterns of runoff generation.
Neither macropores nor interflow: generation of spiky discharge peaks in small
forested catchments and its implications on stream water chemistry
Gunnar Lischeid
BITÖK, Department of Hydrogeology, University of Bayreuth, D-95440 Bayreuth, Germany
Small catchment runoff often exhibits a clear response of discharge and solute concentration within a few
minutes after the onset of rainstorms. These patterns are investigated in two forested catchments. In spite of
very similar hydrographs the dynamics of nitrate, sulphate and silica concentration differ substantially between
the two sites. A detailed study confirmed the variable saturated area concept. Preferential flow does not even
play an important role at the site with most pronounced dynamics,. The discrepancy between similar
hydrometric and dissimilar hydrochemical hydrographs is ascribed to the discrepancy between pressure
propagation and mass transport. During discharge peaks the shallow groundwater level increases and
pre-event water is mobilized by piston-flow displacement. As surface runoff is reduced in the occasionally
saturated area, more stream water is generated by discharging deeper groundwater, even during storm flow,
and the correlation between discharge and solute concentration becomes more less pronounced.
Spatial distribution and temporal behaviour of dry weather flow components in a
small drainage basin
Steffen Möller, Wolfhard Symader
Hydrological Department, University of Trier, D-54286 Trier, Germany
Dry weather flow can only be understood, if the whole drainage basin is considered. It is useful to divide the
basin into subunits corresponding to the longitudinal profile of electric conductivity and major ions. By
applying this method the spatial distribution of important flow components can be found out. Considering of
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
17
temporal behaviour of these sources starts in the next step. Temporally invariant endmembers could be used
for a mixing approach.
In a small drainage basin in the southern Eifel mountain, Germany, nine subunits were found. Six of them
have a major influence on runoff generation under dry weather conditions, but only two show the properties
of characteristic endmembers.
The role of infiltration conditions for storm runoff generation at the hillslope and
small catchment scale
Axel Bronstert
Institute of Geo-Ecology, University of Potsdam, P.O. Box 60 15 53, 14415 Potsdam, Germany
Daniel Katzenmaier
Potsdam Institute for Climate Impact Research, P.O. Box 60 12 03, 14412 Potsdam, Germany
The influence of infiltration conditions on storm runoff generation is investigated by means of a modelling
study focusing on three important factors: (1) macropore flow, (2) soil siltation and (3) small-scale variability
of soil moisture as well as temporal variability of precipitation intensity during periods of heavy rainfall.
Physically based hydrological modelling is carried out at the hillslope and catchment scale respectively. The
case studies presented here investigate the implications of these three factors for runoff generation on
hillslopes and in catchments covered by loess. The first study illustrates the influence of macropores on
storm runoff generation and water retention at the catchment scale. The second study investigates the
effects of soil siltation on catchment runoff. The third study deals with the uncertainty introduced by temporal
variability of rainfall intensity. The results show that soil surface conditions influence the generation of
infiltration-excess overland flow significantly. In particular, a macroporous soil usually reduces direct runoff
and increases water retention in the catchment area. But this effect is only important for rainfall events with
high precipitation intensities. Siltation is modelled by using a simple adoption of empirical findings on siltation
obtained at the plot scale. Simulation results indicate that such a procedure lead to a drastic overestimation
of quick hydrological response at the catchment scale when re-infiltration is not considered. Temporal
variations of rainfall intensity can be very significant for storm runoff generation at the hillslope and
micro-catchment scale, resulting in an increase of runoff production compared to averaged values.
Episodic rising in nitrate concentration of streamwater by partial dieback of pine
forest in Japan: Runoff generation processes as a controlling factor of the
seasonality
Nobuhito Ohte, Naoko Tokuchi, Masanori Katsuyama, Satoru Hobara and Yuko Asano
School of Agriculture, Kyoto University, Kyoto, 606-8502, Japan
Keisuke Koba
School of Informatics, Kyoto University, Kyoto, 606-8501, Japan
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
18
-
Nitrogen dynamics, especially the temporal and spatial changes in NO3 concentration of soil and ground
waters, was investigated in a forest catchment to describe the influences of partial dieback of dominant tree
stands on nutrient cycling and streamwater chemistry. Decreases in N uptake by pine roots and N supply
from litter fall increased the nitrate concentration of streamwater and subsurface groundwater threefold. It
-
was found that seasonal peaks in stream NO3 concentration during the rainy season (July to August)
occurred during 1992-1996. This seasonal variation corresponded directly to groundwater NO3 levels at the
riparian zone near the catchment outlet. This suggested that seasonality in groundwater level is the
-
dominant factor controlling temporal changes in streamwater NO3 concentration.
Runoff generation process and chemical weathering in small headwater
catchments, Tanakami, Japan
Yuko Asano, Taro Uchida, Nobuhito Ohte
Laboratory of Forest Hydrology, Graduate school of Agriculture, Kyoto University
Kitashirakawa-Oiwakecho Kyoto 6068502, Japan
Temporal δD variations in rainfall, throughfall, soil water, groundwater, streamwater and bedrock
groundwater were analyzed to evaluate the residence times of waters and their spatial patterns in Rachidani
(non-forested) and Fudoji (forested) catchments. Using the nearly sinusoidal variations of input and output
δD, the exponential flow model was applied. The followings are the results. (1) Spatial patterns in the age of
waters were more sensitive to the sampling depth compared to the upstream length from the ridge in most
part of two catchments. (2) In the small area near to the channel head in Fudoji, the residence time increased
remarkably reflecting the oldest bedrock groundwater contribution. (3) The ages of waters were younger in
forested catchment than in non-forested catchment when compared at the same sampling depth. It was
suggested that the plant uptake of water during the dry periods increase the ratio of new water content in the
soil pore water during the rainfall events.
TEMPORAL AND SPATIAL VARIATION OF RUNOFF GENERATION AND ITS
INFLUENCES IN THE CHILLIWACK RIVER BASIN, BRITISH COLUMBIA, CANADA
Henry M. Sichingabula
Department of Geography, School of Natural Sciences,University of Zambia, P. O. Box 32379, Lusaka, Zambia
This study investigated the influence and effects of factors controlling temporal and spatial variation of runoff
generation and related processes in the Chilliwack River basin southwestern British Columbia, Canada,
between 1965-1975. Chilliwack River basin, 1230 km2 in size, has a good coverage of weather, hydrometric
and sediment stations. It was previously glaciated and experiences snow and rainfall types of precipitation
throughout the year. The objectives were to determine: (i) the rainfall-runoff ratios in the basin; (ii) major
factors contributing to runoff generation; (iii) temporal and spatial variation of runoff generated by individual
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
19
rainfall and snowfall events; and (iv) to assess the effects of the variations in runoff generation on sediment
transport and occurrence of extreme events.
Analysis revealed that seasonal rainfall-runoff ratios in the Chilliwack basin ranged from moderate to high
owing to the large number of high intensity rain storms experienced in summer and fall months. But this was
low in the winter months due to low and sub-zero temperatures which inhibit runoff generation. The effect of
this was high water storage capacity of the basin and its subsequent rapid release in spring or during periods
of warm air masses intrusions. The major factors contributing to runoff in Chilliwack River included high
precipitation (snow and rainfall) by virtue of its location, and high relief which promotes rapid movement of
runoff on valley slopes. Temporal and spatial variations in the distribution runoff over the catchment were
attributed to various controlling factors. Temporal variation in runoff distribution was understood to be the
consequence of delayed generation of runoff caused by different types of vegetation cover. The conifers
delay runoff generation due to snow and ice storage in canopies for varied periods of time in winter.
Cottonwood, spruce and other broad-leafed trees delay runoff due to rain water retention and its partial loss
to evaporation. Ground ice and ice lenses inhibit infiltration to underground pathways but promoted quickflow
during snowmelt, extremes of temperature were responsible for locking water away into ice during winter
and liberating it as runoff in spring.
These variations in runoff generation influence rates of discharge and sediment movement in rivers and on
slopes due to ice formation in winter; increased frequency of flooding in spring and during warm air intrusions
in winter, especially when combined with occurrence of intense rain storms and blizzards. Prediction of
intense storms and flood events requires better understanding of the influence of these interrelated factors in
order to reduce risks of loss of life and property.
The importance of individual meteorological and
hydrological events on runoff generation and their effects on sediment transport and flooding will be
illustrated by analysis of the interactions using specific examples. Modelling of factors controlling runoff
generation processes, sediment transport and occurrence of extreme events at catchment scale would
provide the best solution to enhanced understanding of this multifaceted hydrological problem.
It is
concluded that physically based models are required for better understanding of basin scale factors
controlling variations in runoff generation and their influences.
The use of cumulative data for characterising hydrological systems.
Chris Barnes
Dept. Engineering, Australian National University
Rainfall runoff relationships for small to medium sized catchments (<1000 km2) are typically determined
from data collected at a single gauging point for both rain and discharge. Variability in the spatial uniformity
of rainfall is probably the greatest source of uncertainty in deriving accurate hydrograph models, with the
consequence that hydrographic response is often poorly characterised (R2<80%), and both model structure
and parameters are only weakly determined. Also, lack of effective techniques for computing error bounds
on parameter estimates significantly compromises any efforts to detect or quantify management effects such
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
20
as land use change on hydrograph properties.
The use of cumulative data is generally disparaged, because although the resulting model fits give
exceedingly high values for R2 (typically >0.99), this is rightly attributed to the smoothing effect of the
inherent integration over time. However, further consideration of this technique of data analysis allows a
more realistic goodness-of-fit comparison, and indicates that minimisation of the cumulative Residual Sum of
Squares (cum RSS) results in somewhat different estimates of model parameters than the usual RSS
minimisation estimates. Moreover, utilising an approximate standard error technique suitable for non-linear
hydrographic models, it is possible to demonstrate that certain key parameters are much more precisely
determined from the cumulative data than from the hydrograph itself.
The technique is particularly suitable for poor quality or sparse data, and permits the concomitant use of
water quality or tracer data in defining catchment behaviour characteristics. The combination of these two
techniques in certain instances allows a quantitatively more precise estimate of management effect, or
alternatively, a significantly earlier detection of a change in system behaviour. These ideas are illustrated
using rainfall and runoff data from a number of Australian catchments, including additional hydrometric and
tracer data where available.
Integrated approach of investigating and modelling runoff formation considering
information from environmental and artificial tracer experiments
Andreas Herrmann, Matthias Schöniger & Sybille Schumann
Institute of Geography and Geoecology, Technical University of Braunschweig, Langer Kamp 19c, D-38106 Braunschweig,
Germany
Profit drawn from an integrated hydrological catchment system approach is demonstrated on the example of
the Lange Bramke catchment in the Harz Mountains, Germany. After introduction on experimental tools,
methods and monitoring systems applied (catchment water balance, environmental isotope and artificial
tracer techniques, hydrologeological survey, geophysical prospecting) respective findings will be
summarised and evaluated. The second main issue concerns GIS and mathematical modelling techniques
used, to make experimental findings suitable for process identification and parameterisation. Finally, it is
shown, how tracer and related spatial data may be used to calibrate hydrological catchment models.
RUNOFF GENERATION AND SCALE IN A HYPERARID REGION
AP Schick, J Lekach, N Greenbaum, T Grodek
Dept of Geography, The Hebrew University of Jerusalem
Detailed, long-term data on rainfall, infiltration and runoff from areas spanning six orders of magnitude (from
0.25 to 500,000 sq.m.) demonstrate a very large variability in time and space and thus highlight the great
difficulties in any future attempts to model the underlying processes in a deterministic way. The data, which
will be presented in detail in the paper, are mainly based on the Nahal Yael Research Project, in continuous
existence since 1965, but use also data taken from other sources in the hyperarid Southern Negev Desert,
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
21
Israel. They include close to 100 simulated rainfall infiltration tests on various terrain surfaces, 5-12 (varying
over periods) high resolution rainfall recorders, 5 high resolution streamflow gauging stations, and an
experimental, limnigraph monitored check dam designed for the verification of the in-catchment records.
In general, the results over the years substantiate the concept of a very large variability both in time and in
space of the main components which regulate runoff: extreme variability of rainfall, the great, lithologically
controlled changes in infiltration capacity over very short distances, and the only partially understood but
probably variable effect of channel alluvium floodflow absorption.
While simple averaging over space and time may be the only way open for portraying the inputs and outputs
in a practically meaningful way, models based exclusively on too coarse areal or temporal unit elements (e.g.
radarscan cells which represent units of one sq.km. and over) are liable to a wide departure from the
underlying physical processes and therefore should be regarded with caution. This word of warning may not
be limited only to arid areas, in which simplicity of processes (e.g. lack of interflow, almost no soil or
vegetation) is ‘compensated’ by compound logistics and paucity of information), but, under various
constraints, to moister areas as well.
Percolation Response to Rainfall computed from separated baseflow
HARTMUT WITTENBERG
Fachhochschule Nordostniedersachsen - University of Applied Sciences, Dept. of Civil Eng., 29599 Suderburg, Germany
Abstract Due to a relative fast infiltration through macro-pores the level of the upper unconfined aquifer of
many river basins rises within days after heavy rainfalls. Groundwater outflow into the rivers reacts
accordingly. After analysis of observed flow recession curves a nonlinear reservoir algorithm is applied for a
separation of baseflow originating from the groundwater reservoir. The use of an inverse reservoir routing
procedure allows to compute time series of daily values of storage and inflow to the aquifer, i.e. percolation
and groundwater recharge. Considering the seasonal variation of flow recession it is also possible to
determine fluxes from the groundwater by evapotranspiration losses. Thus the main components of the
groundwater balance of a basin, namely, discharge, evapotranspiration loss, storage and recharge can be
quantified. Comparison of the computed recharge hydrographs with measured daily rainfalls revealed typical
travel time distributions of water percolating through the vadose zone. Unit responses of groundwater
recharge were found for these data by application of the linear unit hydro-graph theory. The shapes of the
estimated unit response functions showed no significant seasonal variation. Though there is a strong
seasonal variation in the recharge-rainfall ratio of the events, the shape of the determined functions
throughout the year is practically time invariant. These results are similar to those obtained by other
researchers with quite different approaches as lysimetric analyses and piezometric observations. Methods
and results are demonstrated for observed data of different catchments.
Soil water variability and the conservation of wet, species-rich grasslands in
Devon, United Kingdom
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
22
Francien van Soest, Andrew Williams
Department of Geography, University of Plymouth, Drake Circus, Plymouth, PL4 6AS, UK.
Robert Parkinson
Department of Agriculture and Food Studies, Seale Hayne Faculty, University of Plymouth, Newton Abbot, Devon, TQ12
6NQ, UK.
This paper reports an investigation of the soil hydrology of areas dominated by Rhôs pasture, a species-rich
2
wetland community, formerly occupying much of North Devon and Cornwall, UK. At a small scale (3500 km )
the major landscape features that influence the position of Rhôs pasture are identified. At the medium-sized
2
catchment scale (< 60 km ) rainfall-runoff modelling is used to quantify soil saturation periods and relate this
to Rhôs pasture. The majority of the wet grassland sites were found on gentle, concave slopes situated at
sites with medium to high upslope contributing area. The catchment study indicated a preference of medium
to high topographic and soil saturation index values. The study concluded that a strong relationship exists
between Rhôs pasture location and topographic features.
Identification of hydrotopes in a small forested catchment (Dürreych, Black
Forest, Germany)
Guido Waldenmeyer
Institute of Geography & Geoecology (IfGG I), University of Karlsruhe, Kaiserstr. 12, D-76131 Karlsruhe, Germany
Markus Casper
Institute of Water Resources Planning, Hydraulics & Rural Engineering (IWK), University of Karlsruhe, Kaiserstr. 12,
D-76131 Karlsruhe, Germany
This paper discusses the spatial distribution of areas with similar runoff generation (hydrotopes) in the small
forested Dürreych catchment area in the northern Black Forest region of Germany. The hydrograph is
determined by rapid runoff from wet areas of the flat headwater and interflow from steep slopes. Derivations
from the digital elevation model and the forestry site map are combined to identify the hydrotopes. The forestry
site map is more suitable for identifying wet areas than the common topographical index. A satisfactory
correspondence between the proportions of wet areas and coefficients of discharge underscores the
applicability of this approach. The forestry site map, however, does not adequate indicate interflow. Therefore
an interflow index combining slope and soil stratification based on the digital elevation model and hydraulic
conductivities is developed. The first efforts are based on a simplified classification and do not as yet
correspond sufficiently with interflow observations made in the field.
Soil moisture variability and landuse in a seasonally arid environment
Williams, A.G1., Ternan, J.L.1, Fitzjohn, C. 1, de Alba, S2 Perez-Gonzalez, A. 2
1
Department of Geographical Sciences, University of Plymouth, PL4 8AA, UK.
2
Departamento de Geodinamica, Universidad Complutense de Madrid, Espagne.
Spatial patterns in surface soil water content were monitored at six locations in central Spain during
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
23
1998-1999. Volumetric water content, determined using time domain reflectometry, was measured at sites
chosen to reflect both the traditional and recent developments in landuse across a suite of topographic units.
A grid sampling strategy was used to measure soil moisture content and the grids varied in size from 143 to
767 measurement points with node intervals ranging from 2.5 to 5 metres. The spatial pattern of soil moisture
was recorded on two occasions, once during a dry period and once during a wet period. A soil sampling
strategy based on the spatial pattern of soil moisture was developed and samples were characterized in
terms of texture, organic carbon, hydraulic conductivity and soil moisture characteristic curve results. A
digital elevation model was available for each grid and watershed at a 5m contour interval. Multivariate
analysis of the data sets was used to identify the relative importance of land cover, soil property and
topography in controlling soil water content variability.
Ranges of spatial correlation derived from
geostatistical analysis of the soil moisture data sets were determined and used as a guide to indicate the size
of area of different moisture contents and thus the spatial frequency for each land use. The spatial patterns
for selected landuses for the two dates were presented as kriged interpolated contour plots. The spatial
variability in soil moisture shows a clear dependency upon land use type as well as a temporal dependency
related to antecedent conditions. The relationship between the spatial patterns in soil moisture and runoff
generating areas has been assessed using both rainfall simulation and in-field runoff detectors. The
upscaling methodology to predict soil moisture variability for an entire watershed or administrative area
based on slope scale variability is still under discussion. However, the method will take the primary controls
on soil moisture variability into account such as landuse, soils and topography. Strengths and weaknesses
of the proposed approach will be elaborated.
DAY 2 (Tuesday, October 10, 2000)
Session 2:
Runoff Generation
Headwater Scale
Processes
at
the
Hillslope
and
Conceptual modelling of runoff, groundwater and oxygen-18 in a till hillslope
Jan Seibert1, Allan Rodhe2 and Kevin Bishop3
1
Oregon State University, Department of Forest Engineering, Corvallis, OR 97331, U.S.A.
2
Uppsala University, Institute of Earth Sciences, Hydrology, S-752 36 Uppsala, Sweden
3
Swedish University of Agricultural Sciences, Department of Environmental Assessment, S-750 07 Uppsala, Sweden
A geostatistical analysis of groundwater level data from a Swedish till hillslope demonstrated that
groundwater levels in an area close to the stream followed the dynamics of the runoff, while this correlation
dropped off markedly at distances more than ca 60 meter from the stream. A model concept was proposed in
which the hillslope was subdivided into a riparian and a more distant part. Due to very shallow groundwater
in the hillslope, special emphasis had to be put on the interaction between saturated and unsaturated
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
24
storage. This model concept improved the simulations of both runoff and groundwater dynamics. To test the
model structure and investigate whether calibration against a conservative tracer could reduce parameter
uncertainty,
18
18
O-concentrations in the stream were simulated and compared to measured values.
O-concentrations in the runoff were reproduced acceptably, but the simulation of
18
O-flux required too
many new parameters to reduce parameter uncertainty.
Towards a better process representation of catchment hydrology in conceptual
runoff modelling
Jan Seibert, Jeff McDonnell
Oregon State University, Department of Forest Engineering, Corvallis OR
97331, U.S.A.
The use of field data for model calibration is often limited beyond the use of streamflow information despite
the general acceptance that internal state informal are necessary for ensuring internal model consistency.
Hydrologists often have a highly detailed yet highly qualitative understanding of dominant runoff
processes-thus we usually know much more about a catchment than we use for calibration of a model. We
present a new method where weak information (i.e., qualitative knowledge from the experimentalist, which
cannot be used as exact numbers) is made useful through fuzzy measures of model-simulation and
parameter-value acceptability. A three-box model was developed for the Maimai catchment in New Zealand,
a particularly well-studied process-hydrological research catchment. The boxes represent riparian, hollow
and hillslope zones. These zones differ significantly in their groundwater dynamics, physical soil
characteristics, stable isotope composition and end member chemistry. The model was calibrated against
hard data (runoff and groundwater-levels) as well as a number of criteria derived from the weak information.
Parameter sets were evaluated via: (1) the traditional comparison of simulations with observations for hard
data such as time series of runoff and groundwater levels, (2) the acceptability of the parameter values with
respect to field knowledge (e.g., size of the riparian zone or ratio of the hydraulic conductivity in the riparian
zone compared to that in the hollow zone) and (3) the acceptability of the simulations with regard to field
observations (e.g., size of events with groundwater occurrence in hollow zone or peak groundwater levels in
the different zones). Using a comparatively large number of criteria for model evaluation helped to reduce
parameter uncertainty and to ensure a better process representation. The proposed method of using weak
information for model calibration and validation is also a way to encourage dialogue between the modeler
and the experimentalist. It is also useful for comparing the value of different field measurements in support of
modeling.
Analysis of flowpath dynamics at a forested headwater catchment, central Japan
Taro Uchida, Yuko Asano, Nobuhito Ohte and Takahisa Mizuyama
Graduate school of Agriculture, Kyoto University, Kyoto, 606-8502, Japan
To evaluate spatial and temporal aspects of flowpath in a steep headwater catchment, the simultaneous
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
25
measurements of runoff, soil pore water pressures, soil temperatures and water chemistry were conducted.
During the baseflow period, two flowpaths (infiltration in soil layer and water emerging from bedrock) meet in
a small area near to channel head and form a perennial saturated zone. This small saturated area is main
source area of the spring water during the baseflow period. When a rainstorm occurred, the rainwater and
the shallow soil water had large impacts on the formation of transient saturated groundwater at the upper
hillslope. This transient groundwater at the upslope was delivered to the spring via lateral preferential
flowpaths. Consequently, the contribution of bedrock groundwater to the spring water decreased with the
increase of rainfall magnitude.
Experimental analysis of rainfall-runoff events in two small adjacent mountainous
catchments
Christoph Sambale, Gerd Peschke
International Graduate School Zittau, Markt 23, 02763 Zittau, Germany
Process-related knowledge seems to be necessary to get more comprehensive models of runoff formation.
For this reason we observed precipitation-runoff events in two small adjacent basins. More than twenty
events differing in precipitation intensity, duration and amount and antecedent soil moisture have been
analysed. Three general types of runoff reaction can be distinguished: (1) During extreme rainfall the basin
responses are dominated by the rainfall, the catchment features can be neglected and we observe the
maximal expansion of saturated areas. (2) Short but intense rainfall events on dry basin soils cause only a
short and very steep runoff hydrograph. (3) A second runoff wave delayed in relation to the first one is mainly
caused by an interflow component. This runoff type is also influenced by the actual soil moisture content and
by transpiration. We observe these three types in both basins, but there are differences due to the specific
basin features and event characteristics. Such experimental results provide a valuable base for a detailed
modelling of saturation overland flow and interflow.
Runoff
Generation
on
Artificial
Slopes
in
a
Mediterranean-Continental
Environment: Teruel Coalfield, Spain
José-Manuel Nicolau
Departamento Interuniversitario de Ecología. Universidad de Alcalá. 28871 Alcalá de Henares, Madrid. Spain.
Runoff generation mechanisms and their controlling factors on the hillslope scale were investigated on
artificial slopes derived from surface coal mining reclamation in a Mediterranean-Continental area. Rainfall
and runoff on both the inter-rill and rill network (micro-catchment) scale were recorded over a one-year period.
Runoff coefficient and runoff continuity from inter-rill areas to micro-catchment outlet were higher in the
overburden substratum than in topsoil and higher in the more developed rill network than in the less developed
one. Total rainfall is the major rainfall parameter explaining runoff response in overburden. Rainfall volume, I30
and I60 explain runoff on topsoil-grass-covered slopes on the inter-rill scale, but no correlation was found at
micro-catchment level. Hydrological response is controlled by soil surface state, rill network density and soil
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
26
moisture content in overburden, showing great seasonal variability according to soil surface changes during the
year. Plant cover and soil moisture content control a more homogeneous response on topsoiled slopes.
Simulation of water flow in heterogeneous layered hillslope
Mikko Jauhiainen, Tuomo Karvonen
Laboratory of Water Resourcers Helsinki University of Technology P.O.Box 5200 FIN-02015 HUT Finland
Water balance of a forested hillslope was studied. The aim of the work was to determine the path by which
hillslope runoff will reach a stream channel. Humus mat of the site is a water repellent media which has to be
taken into account when calculating water balance of the site. As getting wet the humus layer of the podzolic
soil act as a large water storage that can not be neglected in calculations. When the humus mat is dry it acts
like a sieve with many parallel macropores. At that stage the Richards’ equation can not be used in
calculations. Knowledge of the topography of the bedrock under the soil is also of crucial importance when
calculating runoff of the hillslope. Often it is not known much of the topography of the bedrock. Using the
mathematical models in calculations we can get more understanding of the runoff processes or the pathways
water uses before it gets to a stream channel. In this study two additional features were added to standard
solution of the Richards’ equation. Firstly swelling and shrinking of the humus mat was included. Secondly
certain fraction of the runoff can be converted to direct bypass flow to the groundwater.
Runoff Generation Processes on Reclaimed Opencast Coal-lands, Wales.
Martin J. Haigh[1] and Marianne P. Kilmartin [2],
1. Oxford Brookes University, Oxford, England
2. Open University in the South, Oxford, England
Traditionally, the clayey, compacted, pasture-lands created by the reclamation of former surface-coal mines
in upland Wales have been conceived as simple, surface-runoff-dominated, hydrological systems. Runoff
generation has been controlled by overflow from the thin active layer provided by the relatively
loose-textured artificial topsoil covers, which are created on the sites as part of the land reclamation process.
This artificial topsoil overlies a dense impermeable clay-silt cap, which is created from the weathered shales
and mudstones that dominate local mine-spoils and which covers the loose, open-textured, bouldery
mine-fill beneath. Where these lands replace open moorland, covered with degraded peaty- and
peaty-gleyed podsolic soils, the result is a marked change in the hydrologic regime of affected streams,
especially in dry weather. This is characterised by increased surface and near-surface runoff, higher and
sharper flood peaks, reduced base flow and shorter flow duration. However, there are different patterns in
the behaviour of infiltrometers established on such sites. Six years of measurement at the catchment scale
has shown that runoff generation is controlled by three discrete hydrological process suites, which are
controlled by soil piping and surface ponding. This work also introduces the possibility of hydrological
connectivity, through fissures, between the surface hydrological system and waters in the buried minespoils.
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
27
A study on ponding time and infiltration capacity during steady rainfall using a
rainfall simulator
Kazumasa Fujimura1, Yosihisa Ando2, Fumio Yamada3, Akihiro Nagaiwa3
1)
Department of Civil Engineering, Meisei University, 2-1-1 Hodokubo, Hino City, Tokyo 191-8506, Japan
2)
Department of Civil Engineering, Graduate School of Engineering, Tokyo Metropolitan University, 1-1 Minami-ohsawa,
Hachioji City, Tokyo 192-0397, Japan
3)
Industrial System Control Technology Group, Power and Industrial Systems Research and Development Center,
Toshiba Co., 1 Toshiba-cho, Fuchu City, Tokyo 183-8511, Japan
The infiltration model proposed recently by DISKIN & NAZIMOV (1995) is a fairly simple conceptual model,
and it takes into consideration the rainfall intensity. In this presentation, the validity of this infiltration model is
investigated for infiltration data measured under steady rainfall generated using a rainfall simulator. The
relationship between ponding time and rainfall intensity is studied. Moreover, the simulated infiltration curves
are compared with the observed infiltration curves measured using a rainfall simulator for 40 plots over the
past 10 years. The results support the applicability of this model under field conditions.
Session 3:
Process oriented Modeling at different Scales
Internal validation of the HBV model
Sten Bergström and Göran Lindström
Swedish Meteorological and Hydrological Institute, SE-601 76 Norrköping, Sweden
The expanding use of hydrological models outside the scope for which they were once developed has
stressed the need for evaluation of their physical relevance through internal validation of their basic
subroutines. This process is described from experience by numerous applications of the HBV model since
the early 1970s and onwards. Examples of internal validations by observations of snow, soil moisture,
groundwater, hydrochemistry and stable isotopes are given and their value as grounds for validation is
discussed. Finally, the use of integrated internal model calibration, in contrast to sequential calibration, is
proposed and exemplified.
A knowledge-based system to improve the preprocessing of distributed
precipitation-runoff models
S. Zimmermann, J. Töpfer & G. Peschke
International Graduate School Zittau, Markt 23, D-02763 Zittau, Germany
Basins are subdivided formally into raster cells or into hydrological response units. In the process of
upscaling it is necessary to aggregate these elementary areas to reduce their number as well as the number
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
28
of parameters. The required aggregating approach should be accomplished by methods of Artificial
Intelligence because our knowledge of runoff generation is gained to a large extent through empirical
knowledge which contains significant uncertainties.
We have developed an XPS (expert system) as an instrument of regionalization in order to identify
process-related areas with a dominant runoff generation process. The GIS ARC/INFO solves the problem of
data acquisition and management and generates the smallest areal units from the overlay of different
geo-information. The XPS represents process understanding and is able to determine how a precipitation
runoff model should be parameterized to simulate basin runoff. Such a discretization procedure offers an
improvement on the preprocessing of distributed precipitation runoff models.
Modeling Runoff Generation in A Forested Catchment in Southern Finland
Harri Koivusalo, Teemu Kokkonen, Tuomo Karvonen
Laboratory of Water Resources, Helsinki University of Technology, P.O.Box 5200, FIN-02015 HUT, Finland
2
The modelled generation of runoff in a forested catchment (0.18 km ) is assessed against measurements of
groundwater level and streamflow, and against results of isotope tracer studies. The rainfall-runoff processes
are modeled using a vertical two-dimensional hillslope model. The pressure head distribution in the
unsaturated zone is estimated using a DRAINMOD-type approximation (SKAGGS, 1980). According to the
model runoff generation is dominated by saturation-excess overland flow which occurs when soil water
exfiltrates into the surface or when rainfall/snowmelt cannot infiltrate into the saturated soil. The model is
validated against measurements of groundwater table elevation and streamflow. The soil water procedure is
used to compute runoff during another time period for which estimates of event and pre-event water fractions
in the streamflow exist. The model produced satisfactory results in terms of groundwater elevation and
streamflow reproduction, but the contributions of pre-event water were underestimated.
Integrating tracer studies with semi-distributed models: recent insights from a
sub-arctic catchment in the Cairngorm Mountians, Scotland
C. Soulsby, S. Dunn1 and R. Malcolm
Department of Geography, University of Aberdeen, Aberdeen, Scotland, AB24 3UF.
1
Macaulay Land Use Research Institute, Craigiebuckler, Aberdeen, Scotland, AB15 8QH
Hydrochemical tracers (alkalinity and silica) are used in an end member mixing analysis (EMMA) of runoff
2
sources in the 10 km Allt a’ Mharcaidh catchment. A 3-component mixing model was used to separate the
hydrograph and estimate, to a first approximation, the annual contribution of surface runoff, subsurface
stormflow and groundwater. The model output is compared with that of the semi-distributed catchment scale
model, DIY, which produced a similar hydrograph separation. The integration of the two approaches as a
2
means of reducing structural and parameter uncertainty in DIY and aiding upscaling to larger (>200km )
catchments is discussed.
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
29
A DISTRIBUTED MODEL OF RUNOFF GENERATION IN THE PERMAFROST
REGIONS
KUCHMENT L. S., GELFAN A. N. , and DEMIDOV V. N.
Water Problems Institute of Russian Academy of Sciences, 117735, 3 Gubkin Str., Moscow, Russia
A physically based distributed model of snowmelt and rainfall runoff generation in the permafrost regions has
been developed. Main difference of this model from models of runoff generation for regions with moderate
climate is a small role of infiltration of water into soil. The choice of the structure of the model is based on the
analysis of the long-term observations on the runoff generation at the Kolyma water balance station and is
oriented to the available standard hydrometeorological information in the cold regions. The following scheme
of runoff generation is accepted: the snowmelt water, first of all, fills up the free storage capacity in
topographic depressions, the peat mats and the ground where this water freeze; excess of snowmelt and
rainfall water over the free storage capacity forms overland flow; ice melting and evaporation of soil moisture
begin at the snow free areas of the river basin; the melt of ice in the ground and in the depressions produces
the subsurface flow and increases the basin storage capacity; the subsurface flow occurs on the frozen
layer of the ground. The model is based on the finite-element schematization of the catchment area and
describes the processes of
snow cover formation, snow melt, freezing and thawing of the ground,
evaporation, basin water storage dynamics, overland, subsurface and channel flow. An important feature of
the model is taking into account the effects of the depth of thawed ground on basin water storage and the
water input to the surface and subsurface basin capacities. The case study of the proposed model has been
2
performed for the Upper Kolyma River basin ( the catchment area is 99,400 km ). Most of parameters and
dependencies which have been used for this basin were received on the basis of hydrometeorological
observations on small watersheds. It has been shown that for the region under consideration the hypothesis
of self-similarity of spatial distribution of snow cover for scales from hundreds meters up tens kilometers can
be applied. The sensitivity of the developed runoff generation model to the description of the subgrid
variations of snow cover and the basin storage has been analyzed. Some of the parameters have been
calibrated. Validation of the proposed model for the Upper Kolyma River basin has shown that the model
simulates satisfactory main peculiarities of runoff generation in cold regions.
Paper from Jens Lange et al.
DAY 3 (Wednesday, October 11, 2000)
Session 3:
Process oriented Modeling at different Scales (continued)
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
30
Modeling and validation of runoff and its components in Swiss Pre-Alpine and
Alpine catchments
Joachim Gurtz, Karsten Jasper, Herbert Lang, Massimiliano Zappa
Institute for Climate Research, Swiss Federal Institute of Technology (ETH) Zurich, Winterthurerstr. 190, CH-8057 Zürich,
Switzerland
Alexandre Badoux
Swiss Federal Institute for Forest, Snow and Landscape Research (WSL) Birmensdorf (CH)
Mark Verbunt
Department of Env. Sciences, Sub-department Water Resources, Wageningen University (NL)
Tomas Vitvar
State University of New York, College of Env. Science and Forestry, Syracuse (USA)
In pre-alpine and alpine catchments the quality of runoff modelling strongly depends on the assessment of
the spatial differences in the generation of various runoff components and flow paths as coupled with the
amount and intensity of precipitation and/or the snow melting. The different structured catchment models
2
PREVAH and WaSiM-ETH were applied on the pre-alpine Rietholzbach research catchment (3.2 km ) and
2
the high-alpine Dischmabach catchment (43 km ) for simulation of hydrological processes and runoff
hydrographs. The simulation results are discussed in comparison with the observed catchment discharges,
with measurements of evapotranspiration, soil moisture and outflow of a lysimeter and with groundwater
levels in 3 boreholes. The results point out the interflow as the main runoff component and that the quality of
the snow melt model is essential in these mountainous catchments for hydrological simulations runoff
generation.
Improving the representation of water balance and runoff processes in mesoscale
hydrologic models for water management modelling
Joachim Geyer, Andreas H. Schumann
Institute for Hydrology, Water Management and Environmental Engineering, Ruhr-Universität Bochum, Universitätsstraße
150, D-44780 Bochum, Germany
David C. Garen
U. S. Dept. of Agriculture, Natural Resources Conservation Service, National Water and Climate Center, 101 SW Main St.,
Suite 1600, Portland, Oregon 97204-3224, USA
A spatially distributed daily water balance model for mesoscale catchments in midelevation to alpine
mountainous regions is described. The development of the model structure and parameterization is based
on the integration of spatially distributed data layers of relevant catchment characteristics by employing
geographic information systems. The spatially distributed simulation results offer many opportunities for the
verification of the model structure and its parameterization by field measurements and tracer experiments.
With regard to this, the conception and intention of a new research project are outlined.
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
31
Towards reduced conceptual rainfall-runoff modelling uncertainty
Thorsten Wagener, Matthew J. Lees, and Howard S. Wheater
Department of Civil and Environmental Engineering, Imperial College of Science, Technology and Medicine, London, SW7
2BU, United Kingdom
Conceptual rainfall-runoff model structures suffer from a lack of parameter identifiability. The result is
ambiguity in estimated parameter values, which in turn is propagated as uncertainties into model predictions.
It has been suggested that the improved use of information in available time-series could lead to improved
parameter identifiability. It is possible to derive a dynamic identifiability measure using an extension of
Regional Sensitivity Analysis. The paper describes this approach and shows how the identifiability of
different model parameters varies with the response mode of the system, and hence how parts of the
observed response can be associated with dominant components of the model structure. This information
forms the basis for a multiple-objective calibration, resulting in reduced parameter and prediction uncertainty.
Optimizing land use management – a new approach for hydrologic modelling
Manfred Fink, Sven Kralisch, Wolfgang-Albert Flügel
Institute of Geography, Friedrich-Schiller-University Jena
Clemens Beckstein
Institute of Computer Science, Friedrich-Schiller-University Jena
Managing a catchment for drinking water supply with a high aerial amount of agricultural land use is a difficult
task if one has to maintain a good balance between water quality and restrictions for the farming. In this
paper we present a neural net based method for finding good approximations to solutions of this problem.
The method allows not only to simulate land use scenarios like hydrologic models do. Moreover it can search
systematically for land use scenarios that fulfill certain desired properties without worrying about complexity
of combinational optimization.
MATHEMATICAL MODEL OF RUNOFF FORMATION FOR SMALL RIVER
YU. Ì. DENISOV, À. I. SERGEEV
Central Asian Research Hydrometeorological Institute (SANIGMI), 72 K. Makhsumov Str., 700052 Tashkent, Uzbekistan
It has been studied the formation of small river runoff using mathematic modeling. As a basis of the creation
of the model of river formation has been taken the follow hypothesis. Small river runoff consists of the two
volumes. The first volume is allocated on the surface of catchment area and plays the main role in surface
runoff formation, the second one is in the soil layer under the first volume and determines water infiltration
from upper volume. If intensity of water input per catchment area unit is more than possible value of
infiltration surface runoff is formed , which is collected by river system of the basin. Infiltrated water
accumulates underground one, which is afterward drained by the system of basin streams. Numerical
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
32
solution of obtained system of equations and correllation of computation results with natural measurements
data on individual basins have shown aligibility of accepted hypothesic for quantative assesment of small
mountain rivers runoff formation.
Simulation of runoff and streamflow on local and regional scales
Heinz-Theo Mengelkamp, Kirsten Warrach, Cord Ruhe
Institute of Atmospheric Physics, GKSS Research Center, D-21502 Geesthacht, Germany
Gerard Kiely
University College Cork, Cork, Ireland
Individual hydrological processes are included in the 1-dimensional land-surface scheme SEWAB (Surface
Energy and Water Balance) to simulate runoff generation on a range of spatial and temporal scales. For
macroscale hydrological studies SEWAB is linked to a horizontal routing scheme which describes the transport
of locally generated runoff into river systems and along the river channel network.. The implementation of the
macroscale hydrological model SEROS (SEWAB linked to a Routing Scheme) to the Odra drainage basin is
described. The model is calibrated for a two years period. Simulated and observed streamflow at various
gauging stations show a good agreement. The topographic homogeneity and smallness of a catchment in
Ireland allow a direct comparison of the simulated runoff with streamflow data. The hydrograph is
characterized by an immediate response to individual rainfall events. This behaviour is simulated by
explicitely including the ponding and infiltration process for surface runoff generation.
Integrated water quality modeling by coupling client-server based DBMS, GIS and
the solute model WASMOD
Ulrike Bende-Michl,
Friedrich-Schiller University, Institute of Geography, Dept. of Geoinformatics & Modelling, Loebergraben 32, D-07743
Jena, Germany
For water quality management on a catchment scale there is a great need to build up innovative computing
techniques such as client-server toolsets for better understanding, predicting and presenting complex
hydrologic and solute flow systems.
In this study the client-server based Date Base Management System (DBMS) Oracle8, the GIS facilities of
Arc/Info and the solute model WASMOD were coupled to explain and predict the hydrochemical runoff
generation in the mesoscale Broel river. This approach is assumed to be userfriendly for storing, accessing
and analysing of relational structered, heterogenous data types (point & spatial data) in a multi-tier
environment and especially proved for producing methodic synergism.
As modeling tool it is achieved for predictive simulation of water and nutrient catchment response on the
base of spatially distributed chemical response units (CHRUs). Hence the integrative approach can be
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
33
integrated into a decision support systems interlinking simulations for best management practices (e.g.
what-if scenarios by changing landuse pattern) which is dedicated for future research.
Physically-Based Rainfall-Runoff Models; Setbacks and Solutions
T.H.M. Rientjes and P. Reggiani
Hydrology section, Delft University of Technology, Delft, The Netherlands
The application of distributed physically-based rainfall runoff models to catchment systems is questionable
since it is known that the prediction reliability for these models is low. In order to improve the model
performance, parameter estimation methodologies have to be applied that give information on the structure
of the parameter set and on how well parameters can be identified. Calibration methodologies also must
enable the modeller to define 'effective' parameter values at the scale of a grid element. The theoretical
framework to achieve these goals is presented in this paper. For this study, a quasi three-dimensional
physically-based runoff model is developed while automated parameter estimation is achieved by the
'Gauss-Marquart-Levenburg’ algorithm. In this algorithm, multiple parameters can be estimated by use of
multiple state variables.
Extrapolating parameters of a conceptual catchment scale model to different
catchments without calibration
Sarah M. Dunn
Macaulay Land Use Research Institute, Craigiebuckler, Aberdeen, AB15 8QH, Scotland
Carmen Bernzen
Ruhr University Bochum, Bochum, Germany
Chris Soulsby, Rob Malcolm
Department of Geography, University of Aberdeen, Aberdeen, AB24 2UF, Scotland
Allan Lilly
Macaulay Land Use Research Institute, Craigiebuckler, Aberdeen, AB15 8QH, Scotland
A technique has been developed to calculate conceptual parameter sets of a semi-distributed hydrological
model using a soil hydrological classification in addition to topographic data. The method is tested in this
paper by attempting to model a new catchment without calibration of the parameters. The simulations are
successful in representing most features of the hydrograph, but under-predict the groundwater response for
the catchment. Calibration of the groundwater recharge fraction leads to acceptable results, that are
comparable with the results of a hydrochemical mixing model in terms of flow separation.
How far from the modelling optimum? Analysing a process-based catchment
model by artificial neural networks
Gunnar Lischeid
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
34
BITÖK, University of Bayreuth, Dr.-Hans-Frisch-Str. 1-3, D-95440 Bayreuth, Germany
Stefan Uhlenbrook
Institute of Hydrology, University of Freiburg, Fahnenbergplatz, D-79098 Freiburg, Germany
2
A process-based, conceptual runoff model is developed and calibrated for the 40 km Brugga catchment
(Black Forest Mountains, southwest Germany) that explicitly takes into account different runoff generation
processes and components. In addition, artificial neural network (ANN) models are applied using the same
data set. Both approaches yield comparably good results, similar to three other process-based models. This
provides evidence that the models were close to the optimum achievable with the given data set. Silica
concentration in the catchment runoff is used to check the runoff components mixing approach of the
conceptual model. Using modelled discharge values instead of measured data markedly deteriorates the
silica predictions. On the other hand, the ANN model substantially improves with input of hourly instead of
daily values, thus allowing mapping the observed discharge-silica concentration hysteresis. Antecedent
moisture conditions proved to be crucial, since in the dryer summer season both the dependence of silica
concentration on discharge and the hysteresis of this relationship is much more pronounced than in the
wetter winter and spring seasons.
New studies for stream-subsurface hydrologic exchange (bank storage)
Tobias Conradt, Matthias Schöniger
Department of Physical Geography and Hydrology, Institute of Geography and Geo-ecology, Technical University
Braunschweig, Langer Kamp 19c, D-38106 Braunschweig, Germany
Hydrological models of river systems should consider river-groundwater correspondence. For this purpose it
is necessary to know about the spatial distribution of the water exchange due to water level variations. A
method for estimating its intensity has been developed. The main idea is to take statistical correlations
between water table measurements in observation wells and the neighbouring part of the river and to strip
these data from the dominating effect of distances between river and wells. Additionally a modelling example
is presented: a flood routing model based on Saint-Venant equations coupled with a bank storage module.
Water levels, exchange fluxes and their influences on runoff can be computed dynamically.
Retrieval of Hydrological Parameters from Remote Sensing Data
Volker Hochschild
Institut für Geographie, Friedrich-Schiller-Universität Jena, Germany
Distributed hydrological models are a key issue in water resources management. They serve as tools for
applied hydrological decision making. Since they are based to a large amount on areal data, remote sensing
data is a powerful technique to supply models with actual areal input. The number of hydrologically important
parameters derived from remote sensing data is manifold and ranges from precipitation estimates,
evapotranspiration over land use and Digital Terrain Models (DTMs) to soil moisture and vegetation
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
35
parameters. Especially in remote areas like most parts of rural Africa with lack of field measurements of
components of the hydrological cycle, remote sensing data is the only way to provide areal input. Major
drawback of all these high resolution satellite systems in the hydrological sense is the limited time resolution.
The requests of hydrological systems analyses to data derived from remote sensing and the possibilities of
remote sensing for the parameterization of hydrological models (MMS/PRMS, ACRU) are the major
objectives. It is realized within the methodological chain of optical, microwave backscatter, microwave phase
(interferometry) and multifrequent and multipolarimetric SAR data evaluation. These methods are applied to
examples from different climatic zones and different relief energy: Thuringia (Germany), Sardinia (Italy) and
Southern Africa (South Africa, Swaziland, Zimbabwe).
The remote sensing evaluations serve as input data into the hydrological models and concentrate on water
balance and water quality. It include the following topics:
-
detailed and in terms of hydrology enhanced mesoscale land cover classifications of catchments up to
15 000 km² with optical (Landsat TM, SPOT, IRS-1C) and microwave (ERS-1, ERS-2) datasets,
-
multisensoral analysis of flood and low flow river channels,
-
derivation of high resolution Digital Terrain Models from stereo SPOT imagery and aerial photography to
derive the extension of floodplains,
-
detection of rural settlements and gully erosion to assess their contribution to water quality,
-
derivation of LAI and biomass as input for the modelling of interception and evapotranspiration,
-
detection of short-time weather induced alterations by interpretation of coherence images (ERS Tandem
Mission),
-
multitemporal, multifrequent and multipolarized air- and spaceborne SAR data evaluation to derive soil
moisture distributions and weather independent agricultural land cover classifications.
Integration of the data into a GIS and connection with a DBMS inside a decision support system called
IWRMS (Integrated Water Resources Management System) in Southern Africa are shown, as well as the
role of remote sensing data for the validation of simulation results. The use of remote sensing data though
enables the extrapolation and regionalisation of hydrologic models.
A SVAT-scheme combined with a simple regionalized baseflow separation for
large scale groundwater recharge modeling
Armbruster, Volker; Haas, Markus; Leibundgut, Christian, Demuth Siegfried
Institute of Hydrology, University of Freiburg, Fahnenbergplatz, 79098 Freiburg
A complex SVAT-scheme is applied at 20 different lysimeter stations. The results show, that the model is
capable to simulate groundwater recharge on a monthly basis under different conditions for flat alluvial aquifers.
Furthermore, a method is described to account for additional losses by fast runoff components in more complex
terrain (mountainous regions). This method uses a baseflow separation in a dynamic sense and relates the
baseflow to the total discharge. To obtain a statistical model which estimates this ratio (baseflow/total
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
36
discharge) for ungauged catchments a multiple linear regression analysis is performed. The resulting model
estimates the ratio in dependence on physiographic catchment characteristics, which are known to influence
runoff generation, with an acceptable accuracy.
Operational application of the water balance model LARSIM in the Neckar basin
Manfred Bremicker
State Institute for Environmental Protection of Baden-Wuerttemberg, PO Box 21 07 52, D-76157 Karlsruhe, Germany
Kai Gerlinger
Ludwig Consultant Engineers, Herrenstr. 14, D-76133 Karlsruhe, Germany
The water balance model LARSIM was applied to the Neckar catchment in the southwest of Germany (area
approx. 14000 km²). The model was set up with a raster-based discretisation of 1 km². It was calibrated and
verified using a ten-year series of hydrometerological input data and measured discharges on daily time
steps. Currently it is used in a test phase as a real-time forecast model for low and mean flow with hourly time
steps at the State Institute for Environmental Protection of Baden-Wuerttemberg. Input data is the forecasted
hydrometeorological data by the German Weather Service over a time period of 48 h and the currently
measured data.
Modelling runoff generation processes for the regions with seasonal snow cover
at plot and catchment scales
Yeugeniyi, M. Gusev & Olga N. Nasonova
Institute of Water Problems, Gubkina St.3, Moscow 117971, Russia.
The aim of the paper is to investigate the ability of a physically based land surface model SWAP to simulate
runoff generation processes on a plot and catchment scales for the regions with seasonal snow cover and
seasonally frozen soil. The model SWAP which operates on a plot scale was coupled with a two-dimensional
kinematic wave equation to simulate catchment hydrological response. Validation of the coupled model was
performed using a set of hydrometeorological data measured during 18 years at the Usadievskiy catchment
(grassland) situated in the central part of Valdai Hills (Russia). The results obtained have shown that SWAP
parameterizes runoff generation processes occurring in mid-latitude grassland quite reasonable.
DAY 4 (Thursday, October 12, 2000)
Session 4:
Tracer Studies for Investigating
Processes at different Scales
Runoff
Generation
Modeling hydrologic responses in a small forested watershed by a new dynamic
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
37
TOPMODEL (Panola Mountain, Georgia, USA)
NORMAN E. PETERS
U.S. Geological Survey, 3039 Amwiler Rd., Suite 130, Atlanta, GA, 30360
JAMES FREER & KEITH BEVEN
Institute of Environmental and Natural Sciences, Lancaster University, Lancaster LA1 4YQ, UK
Preliminary modeling results for a new version of rainfall-runoff model TOPMODEL, Dynamic TOPMODEL
are compared to those of the Original TOPMODEL formulation for predicting streamflow at the Panola
Mountain Research Watershed, Georgia. Dynamic TOPMODEL uses a kinematic wave routing of
subsurface flow, which allows for dynamically variable upslope contributing areas. Each models
performance was assessed using a 30-minute time step for each of three water years (October through
September: WY93, WY94, and WY98) and results were compiled for wetting up, wet, drying out and dry
periods within each year. A likelihood measure of predicted streamflow was similar for Original TOPMODEL
2
and Dynamic TOPMODEL and depending on the year and the sub period, the r of the best models ranged
from 0.5 to 0.7.
Each model underpredicted the peak streamflow, and during recession generally
overpredicted streamflow in wet periods and underpredicted in dry periods. However, during recession, the
difference between the observed and Dynamic TOPMODEL predicted streamflow was generally less than
that for the Original TOPMODEL. The distribution of transmissivity for the Dynamic TOPMODEL simulations
generally is more stable for the different wetness periods, except during dry periods, than for the Original
TOPMODEL simulations, i.e., a clearly defined relation between SAE and T0 having a narrow band of
transmissivities for minimum SAE simulations. The Dynamic TOPMODEL prediction bounds were broader,
and the lower bound more closely follows recession streamflow than that of the Original TOPMODEL.
Effects of the variations of groundwater levels on the hydrochemical processes in
a headwater catchment, Japan
Masanori Katsuyama, Nobuhito Ohte
Division of Environmental Science and Technology, Graduate School of Agriculture, Kyoto University, Kitashirakawa,
Oiwake-cho, Sakyo-ku, Kyoto 606-8502, Japan
In order to understand the effects of the variations of groundwater levels on the temporal variation of
streamwater chemistry, intensive water sampling with hydrological observations was conducted during
rainstorms at a small headwater catchment, Japan. The three end-members contributed to the stormflow
were: (a) the groundwater in the saturated zone that prescribes the chemistry of the baseflow (b) the
throughfall that dilutes the streamwater and (c) the groundwater in the transient saturated zone prescribed
which was dependent on the groundwater level. When the groundwater level was lower, only the two
components, groundwater in the saturated zone and throughfall, affected the streamwater chemistry. When
the groundwater level rose and saturated zone spread, the groundwater in the transient saturated zone
became the third component. When the third component contributed to the discharge, this component
became the dominant source.
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
38
Paper from Tomas Vitvar et al.
Uncertainties of two-, three- and five-component hydrograph separations in a mountainous
catchment
Simon Hoeg
Interdisciplinary Centre for Scientific Computing, Reactive Flows Group, University of Heidelberg, INF 368, D-69120
Heidelberg, Germany
Stefan Uhlenbrook
Institute of Hydrology, University of Freiburg, Fahnenbergplatz, D-79098 Freiburg, Germany
The hydrograph separation technique using natural tracers, (where different runoff components are quantified in
dependence of their chemical or isotopic signature) is a suitable method for investigating runoff generation
processes. The first objective of this study is to demonstrate equations for separating three and five runoff
components using
18
O and dissolved silica as tracers. The second is to describe the uncertainty of the
hydrograph separations with Gaussian error estimators.
The study took place in the mountainous Zastler catchment (18.4 km², southern Black Forest, Germany). During
several storm events an interaction of three main runoff components having distinct silica concentrations was
evident: (1) the low silica component from riparian zones and impermeable areas, (2) the medium silica
component from periglacial and glacial debris and drift cover and (3) the high silica component from crystalline
detritus and crystalline hard rock. The first two components could be separated into event and pre-event water
fractions, so that five components are considered in all. It is demonstrated that large relative uncertainties must
be considered for the quantification of runoff components. These are caused primarily by analytical errors and
the spatial variability of end member concentrations. Nevertheless, reasonable runoff generation behaviour of
the catchment could be shown.
Verification of flow processes in soils with combined sprinkling and dye tracer
experiments
Markus Weiler & Felix Naef
Institute of Hydromechanics and Water Resources Management, Swiss Federal Institute of Technology Zürich, CH-8093
Zürich, Switzerland
Water flow in soils is strongly influenced by macropores, the heterogeneity of the soil matrix and the
exchange of water between the macropores and the soil matrix (interaction). The degree of interaction
strongly influences the runoff generation process in soils with macropores during extreme rainfall events. A
low interaction in combination with low permeability in the subsoil produces rapid overland flow. Low
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
39
interaction in combination with permeable lateral flow paths in the subsoil results in bypassing of the topsoil
and the generation of subsurface flow or the opening of the storage capacity of lower soil layers. The water
movement during infiltration and the resulting flow paths were studied with sprinkling experiments and dye
tracing under different rainfall intensities and soil moisture conditions. The dye tracer was continuously
2
applied with the rainfall on 1 m plots. After the sprinkling, cross-sections of the soil were prepared for
surveying horizontal and vertical dye patterns, thus displaying the cumulated flow path in the soils. These
experiments were carried out on sites with different infiltration behaviour. The results are used to illustrate
and study the influence of the water exchange between macropores and the soil matrix on runoff generation.
These demonstrate the dependence of flow processes on properties like water repellence near the soil
surface and soil structure. Thus, the verification of flow processes in the field can be improved with combined
sprinkling and dye tracer experiments.
The separation of flow pathways in a sandstone catchment of the Northern Black
Forest using a nested approach
Markus Casper; Erich J. Plate; Jens Mehlhorn
Institut für Wasserwirtschaft und Kulturtechnik (IWK)
Holger Volkmann
Institut für Ingenieurbiologie und Biotechnologie des Abwassers (IBA)
Guido Waldenmeyer
Institut für Geographie und Geoökologie (IfGG), University of Karlsruhe (TH), Kaiserstr. 12, D-76128 Karlsruhe, Germany
Knowledge of runoff generating processes is fundamental for testing and improving of predictive rainfall
runoff models. In this study, the different types of runoff from a mountainous catchment in the northern Black
Forest were explored. The observed runoff was caused by rainfall events with different volumes and
intensities occuring with different pre-event catchment conditions. The hydrographs were separated by
methods adjusted to the expected time scale of the underlying processes: (a) catchment scale: Isotope
(18-O) and DOC based separation, (b) baseflow from springs using hydrometric and isotopic data (3-H,
18-O), (c) separation of saturation excess flow and hillslope runoff by direct measurements from 4 subcatchments, (d) description of flow pathways for saturation areas through detailed isotope and DOC
measurements and (e) additional hillslope investigations on a plot scale to distinguish between rapidly
responding pipeflow in the upper soil layer and a slower subsurface flow in the lower soil horizons of the
podzols. By combining all results we were able to roughly separate the hydrograph at the catchment outlet
into 4 components: base flow from deep spring; saturation excess flow from the flat saturation areas of the
tops, runoff from the hillslopes and runoff from the alluvial plain.
Humic substances and agrochemicals in the discharge components of Scheyern
K.-P. Seiler, C.Hellmaier
GSF-Institute of Hydrology, D-85758 Neuherberg, Germany
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
40
In the unsaturated zone, bypass-flow and its feedback with matrix and interflow significantly govern the
subsurface export of solute and particle matter. At the FAM Research Station, it was observed that the
concentration of agrochemicals in the interflow discharge component increased proportional by the amount
-
of precipitation wherever rain events exceeded 2 to 4 mm/day. The distribution of DOC, nutrients (e.g. NO3 ),
-
pesticides and additives of fertilizers (e.g. Cl ) within the effective root zone (0 to 100 cm below surface)
reflected the current application, consumption and leaching of agrochemicals. Due to slow seepage
velocities in the unsaturated zone below, only past applications were observed. Since a lesser amount of
agrochemicals are presently applied, their concentrations are lower in the effective root zone than in the
unsaturated zone. - In the study area DOC export from forested areas is higher as compared to agricultural
areas; obviously DOC leaching is either more pronounced in agro-lands or DOC production is higher in
forest soils. - Since interflow produced by storm events export insoluble agrochemicals such as phosphates
and pesticides, and since DOC increases during storm events, a significant particle-favoured export of
insoluble agrochemicals must occur. The export of agrochemicals via storm events provides a significant
contribution to groundwater protection, favours pulse impacts to surface waters and causes income losses
to farmers.
Hydrograph separation of selected runoff events using DOC and
18
O as natural
tracers
Holger Volkmann
Institut für Ingenieurbiologie und Biotechnologie des Abwassers, Universität Karlsruhe, Am Fasanengarten, Postfach
6980, 76128 Karlsruhe, Germany.
Markus Casper
Institut für Wasserwirtschaft und Kulturtechnik, Universität Karlsruhe, Postfach 6980, 76128 Karlsruhe, Germany.
Investigations of the spatial and temporal dynamics of dissolved organic carbon (DOC) export and
18
O
(SMOW) in precipitation, soil water and runoff were carried out within the Duerreych catchment, in order to
distinguish runoff generation processes. Hydrograph separations of selected precipitation events were
constructed for a small sub-catchment. The DOC in soil water showed a clear decrease in concentration with
increasing soil depths. The DOC-rich water originates from the upper soil horizons and acts as a rapid runoff
component. Using the temporal variability of DOC concentrations in runoff water, the dominating runoff
processes during precipitation events were ordered chronologically. The study catchment was found to be
dominated by fast runoff components with water percolating through the upper soil horizons with a major
portion of runoff water being (old) pre-event water. Saturation excess flow plays a minor role except for
strong storm events.
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
41
6 Posters
The following posters are presented during the workshop. The poster session is
scheduled Thuesday 11:30-13:00. All poster authors are asked to be available during
this session in front of there posters.
USE OF SO4 AS A NATURAL TRACER FOR THE HYDROGRAPH SEPARATION OF
A KARSTIC SPRING
Iñaki Antigüedad and Inma Mugerza.
University of the Basque Country.
In the Basque Country a lot of rivers cross karstic regions in a way that surface waters and groundwaters are
strongly interrelated. In fact, some rivers, mainly in its headbasin (head waters catchment) areas, are karstic
groundwater dominated rivers.
2
In the Albiztur catchment (27 km ), in Gipuzkoa province, we are carrying an investigation on the mutual
influence between surface waters and groundwaters in order to study runoff generation processes (correct
separation of the effective precipitation into different runoff components) in areas where karstic waters
influence is very great.
This study is based on a combination of hydrometric measurements (river, sinking creeks, main springs,
boreholes) and natural and artificial tracer methods. In fact, we use the differences in natural tracer
concentrations of the two end members (inflow river water - groundwater in the outflow river; quick flow delayed flow in the spring) as an effective way to achieve a reliable hydrograph separation.
Concretely SO4 is the most important natural tracer in this catchment, taking into account the different
composition for this element in waters from cretacic limestones (less sulphate) and from jurasic
limestones-dolomites (more sulphate). SO4 contents in inflow river waters varies from 130 mg/l in low waters
to 20 mg/l in high waters and in outflow river waters from 60 mg/l in low waters to 15 mg/l in high waters.
According to the preliminary results we conclude that in this catchment about 85 percent of the total outflow
river discharge is caming from the aquifer (Salubita big spring), varying from 100% in very low waters to 20%
in near peak moments of very high waters. Contribution of the quick flow to the spring discharge is very
changeable according to the peak importance. On the other hand, comparison between inflow into swallow
holes and outflow from spring, together with tracing test, allows us to evidence the influence of epikarst in
separation of runoff components. We are now quantifying this influence.
PECULARITIES OF RUNOFF GENERATION IN DESERTS (AS AN EXAMPLE OF
KIZILKUM)
S.A. Azimbaev & B.S.Nurtayev
Institute of Water Problems,Uzbek Academy of Sciences 25a, F.Hodjaeva Street,Tashkent,Uzbekistan,700000
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
42
Water scarcity is one of the major cause that constrains of agricultural development of deserts.Since ancient
times population of such areas, attempted to accumulate and secure rain water runoff for their needs.
Nowadays improvement methods on accumulation of runoff, study Á rules of runoff generation can
significantly improve present situation. Precipitation is the main water source at such area.Average annual
rainfall in Kizilkum desert varies from 100 up to 150 mm or about 10-15 cu.km,it is the major cause of
formation a temporary indigenous water. Based on a method of relief plastic (scale 1:300000) at the Central
part of Kizilkum by autors were fulfiled regionalization of districts in accordance with dynamic of runoff
generation and accumulation.As a result were subdivided three regions of runoff generation :i) concentrated
runoff ;ii)diffused runoff; and, iii) without runoff.
Delineation of
areas with homogeneous system response (HRUs) for
parameterizing hydrologic relevant processes - a case study from Thuringia Klaus Bongartz
Friedrich Schiller Universität Jena, Institut für Geographie, Lehrstuhl Geoinformatik, Geohydrologie und Modellierung,
Löbdergraben 32, 07743 Jena
Analysing and modeling hydrological catchment dynamics implies dealing with the problem of scaling up
point measurements to a mesoscale spatial distribution. This challange can be met by applying the concept
of hydrological response units (HRUs). Different HRU-Concepts are used to delineate areas which have
similar hydrological properties within the subsystems of atmosphere, biosphere, pedosphere, and
lithosphere. These areas also can be used as model entities for parameterization. The topographic
stream-segment-based HRU delineation after LEAVESLEY (1983) is compared to the more physiographic
based HRU delineation after FLÜGEL (1995). Thereby the spatial distribution of those entities is deliniated
2
using GIS analysis. The concepts are applied to the mesoscale catchment of the river Ilm (A = 895 km ) in
Thuringia. The catchment is divided into three nested subcatchments: (i) to investigate the advantages and
disadvantages of both HRU-concepts on different catchment sizes (ii) to test the possibility of parameter
transfer from one catchment to another and (iii) to analyse the influence of variable precipitation distribution
on catchment response.
The climate is determined by the transition zone between wet oceanic and dry continental climate where the
3
mean annual rainfall ranges between 500 and 1200 mm and the mean annual runoff is 63 m /s. The spatial
database includes land use pattern classified by IRS-1C-Data into the landuse classes agriculture (42%),
coniferous and deciduous forests, (31%), rangeland (12%) and impervious areas (12%).
Furthermore a GIS database containing other physiographic factors like soils, geology and topography was
build up. Additionally on the point scale there is precipitation data from 16 raingauges in the catchment.
Temperature, snow, and radiation data was measured from 2 Stations and runoff data from 3 gages.
The model PRMS/MMS in combination with the GIS WEASEL and ARC/INFO was used for modeling
hydrological catchment dynamics. PRMS is a deterministic, distributed and continuous water balance model
running successfully within the modeling framework of MMS. First simulations show that the physiographic
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
43
based concept was found to be a reliable method for modeling basin dynamics in catchments up to 200 km
2
whereas the topographic based concept has advantages in modeling bigger sized catchments. Future
research should tent to combine both concepts to represent the spatial heterogeniety of the catchment and
connect areas of different hydrological response to the drainage system.
FORMATION OF A MINOR RIVER RUN-OFF UNDER THE CONDITIONS OF
ANTHROPOGENIC IMPACT ON THE CATCHMENT
Yelena V. Fyodorova
Russian Research Institute for Integrated Water Management and Protection
ul. Mira, 23 Yekaterinburg, Russia
Influence of the catchment structure and climatic conditions on the formation of quantitative and qualitative
characteristics of a minor river run-off has been considered. It has been revealed that the flood run-off layer
and rainfall high water values are formed depending on combination of run-off forming complexes at the
catchment that are characterized by a particular accumulating capacity/ Accumulating capacity of every
run-off forming complexes depends both on natural factors (climate and landscape) and on their economic
use. The following natural factors of the run-off formation have been considered: water storage in snow,
freezing depth and pre-winter soil moistening. Such anthropogenic factors as drainage of peat soils and
plough land have been considered. It has been revealed that there is a close linkage of the flood run-off layer
with the sum of positive temperature values during snow melting period and pre-winter moistening of the
catchment. It has been found that drained peat soils are characterized by the most accumulating capacity,
while run-off forming complexes represented by ploughed lands are characterized by the least accumulating
capacity.
Distributed Hydrological Model Parameterization of Large Scale Afforested Areas
in the North East Cape Province, South Africa
Jörg Helmschrot
Dept. of Geoinformatics, Institute of Geography, FSU Jena, Löbdergraben 32, D-07743 Jena, Germany
Water is one of the most important natural resources in South Africa. Because of its variable spatial and
temporal distribution, the continuous assessment and monitoring of hydrological system components is
necessary. Detailed informations on geology, topography, land use, etc. are needed for hydrological
catchment modelling. Particularly the impact of land use changes on the basinwide runoff is of interest to
water resource planners and managers. Remote sensing data provide actual information in large scale
areas which are difficult to monitor using conventional field techniques. The integration of derivates from
remotely sensed data in Geographical Information Systems (GIS) and their analysis can be seen as the
primary tool for operational, efficient acquisition of input parameters for physically-based distributed
hydrological models.
The major objective of this research study is to investigate the impact of large scale afforestation on
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
44
hydrological dynamics of semi-arid upland catchments in the North East Cape Province as well as their
analysis and prognostic modelling. Because of its physio-geographic factors (similar geological and
geomporphological conditions, semi-arid environment, land use changes,the forestry development with its
impact of runoff reduction) as well as socio-economic problems (population growth, deficits in domestic
water supply) the Umzimvubu catchment (19 845 km²) shows a typical hydrological system behaviour and
can be seen as a representative area for larger parts of the sub-tropical South Africa.
The paper focuses mainly on the integration of field studies, remote sensing data and GIS analysis to the
parameterization for hydrological modelling in the Umzimvubu basin. Hence, field based techniques were
developed for the determination of plant-physiological parameters such as LAI and cover density for the
validation of the land use information and the parameterization of the hydrological model. Furthermore
preprocessed stereoscopic SPOT PAN images were appraised to generate a high-resolution Digital
Elevation Model (DEM) of the Mooi river subcatchment by using the photogrammetric ORTHOMAX software.
Mean height accuracies of about 13 to 21 m could be achieved. Furthermore multispectral Landsat TM data
from May 1995 were used for an quantitative determination of the land use patterns in the Umzimvubu
catchment. After the data preprocessing a hybrid approach was applied. First an unsupervised classification
was choosen for estimation of spectral characteristics within the selected testsites. The supervised
classification was utilized by a Maximum-Likelihood classificator and enabled to generate a basinwide land
use map with 11 classes and an enhanced landuse map for the Mooi river subcatchment. Finally the
classification results were improved using several post processing algorithms. In a second step, a TM scene
of April 1999 was used to detect land use changes considering supplement forestry and measured plant
specific data. A precise map of the extent of afforested areas was derived.
Further investigations focus on the integration of the topographic and land use data in the GIS-based
delineation of distributed homogeneous modelling entities and the final modelling system parameterization.
Analysis of high resolution multifrequent, multipolarimetric and interferometric
airborne SAR Data for hydrologic model parameterization
Herold, Martin & Hochschild, Volker
Geoinformtics, Department of Geography, FSU Jena, Loebdergraben 32, 07743 Jena, Germany
Remote Sensing techniques can provide spatial information for investigations on runoff generation and
modelling respectively for parametrization, calibration and validation of hydrological models. Especially
multifrequency, polarimetric and interferometric
microwave remote sensing data are sensitive for
three-dimensional dielectric and structural features of the earth surface allowing the derivation of soil
moisture, vegetation parameters, land use and topographic information.
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
45
In this study airborne SAR data of two flight campaigns using the “Experimental Synthetic Aperture Radar”
(E-SAR) of the German Space Agency (spatial resolution 2 m) have been evaluated in investigations of a
detailed study on hillslope hydrology in the Broel catchment (Rheinisches Schiefergebirge, Germany) and
micro-scale solute transport simulation of
Testsite Kiefer
Testsite
the Thuringian Drinking Water Reservoir
Administration
(TTV)
in
Zeulenroda
(Thueringia, Germany). Good results were
obtained in the derivation of land use
(overall accuracy 88,3 %) and vegetation
a
information (plant water content, biomass,
plant height, correlation with field data:
r=0,95-0,98). The interferometric data
b
were used to calculate a digital surface
model with 5 m spatial and 0,5 m vertical
resolution. For correction of the vegetation
heights the land use information was
c
included in the retrieval of a digital
elevation model. For the derivation of
Figure: Comparison of surface soil moisture distribution as
derived from radar backscatter (a), result of the multiple flow
topographical
index
(b)
and
interpolated
from
TDR
spatial soil moisture distribution the longer
wavelengths
like
L-Band,
partly
penetrating through vegetation into the
soil surface, are most sufficient. The soil
and vegetation induced backscatter parts were separated by the calculation of principal components (PC)
from the L-band polarizations. A regression between the tensiometer field measurements and the values of
the first PC (r=0,79) were applied to estimate a surface soil moisture map. The comparism with other spatial
soil moisture distributions (see figure) derived from the DEM (multiple flow topographical index) and with
geostatistical methods from data of 51 TDR-measurement points (2-3 measurements at each point) shows
the good representation of topography-induced soil moisture variations with wet drainage lines visible in the
microwave data, while the slopes are dryer. At the Kiefer testsite there is a saturated zone in the middle of the
slope resulting from a dip in the underlying geology where the interflow from the upper slope is cumulated
and linearly drained to the valley floor. This area is visible in the radar backscattering (figure a) and the TDR
soil moisture sampling (Fig. 3c), but not in the mf-topographical index (figure b), because it is not
represented in the surface topography.
The effect of interannual snow cover variability on spring flood snow constituent
in the large river basins of Northern Eurasia.
L. Kitaev, T. Wegener, E. Barabanova
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
46
Institute of Geography, Russian Academy of Sciences
The investigations of the role of snow cover in the river run off formation, though having a long history, are
still actual nowadays. The main challenge is to learn the influence of temporal and spatial variability of snow
cover and climatic properties on the character of spring flood in the large river basins of Northern Eurasia.
The data base used contains long-term observational data on snow cover, river run off, air temperature and
precipitation conditions, obtained at hydrological and meteorological stations. For the period under
investigation the snow constituent of the spring flood has been separated out.
Positive (100-180 mm) and negative (less than 70 mm) anomalies of spring flood snow constituent resistant
with space and time are observed. They are determined by the character of snow storage, climate conditions
and orography.
The peculiarities of interannual variability of spring flood snow constituent are revealed. Its greatest
amplitudes (similarly to snow storage) are characteristic for the Yenisey basin (70 mm), the smallest - for the
Amur basin (35 mm). Interannual changes of the snow constituent are opposite in sign to those of snow
storage and climatic parameters. This is determined by the influence of supplementary factors exhibiting
regional features. In particular, in the Siberia the spring flood is essentially influenced by the permafrost, on
the East European Plain it does by seasonal soil frosting and thawing out.
J2000 – A new Modelling System for physically based simulation of hydrological
processes in large catchments
Peter Krause
Research Centre Jülich, Systems Analysis and Technology Evaluation
The process oriented modelling of catchment hydrology in the meso- to macroscale is of increasing
significance due to actual political, economical and ecological questions. Modelling systems which are able
to produce reliable forecasts concerning the influences of global-change scenarios, e.g. landuse or climate
changes on the water quantity and quality in large catchments are needed. For that purpose it is necessary
to model the processes, which are responsible for the runoff generation, spatially differentiated and
physically based. Models for the microscale which are able to simulate the runoff generation with a high
degree of certainty and which can be validated well are already existent. The transferring of this models to
the meso- or macroscale is not possible without further ado, because of the required data on the one hand
and the meaning of the single processes, which can change during the upscaling on the other. Due to this
reason demand for new model systems, which are able to simulate the relevant runoff generation processes
in the macroscale on the basis of the available data resources, with a similar certainty as the microscale
models exists. For this purpose the modelling system J2000 was developed at the Research Centre Jülich.
The J2000 is a modular modelling system which is able to simulate the processes, which are responsible for
the generation and concentration of the runoff in large catchments, spatially differentiated, physically based
in high temporal resolution ( t = 1 day) on basis of hydrological homogenous regions with the goal to
quantify the total runoff and its components (surface runoff, interflow and baseflow) with high accuracy.
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
47
Additionally the model system is capable to regionalize the precipitation and climate data which are gathered
on single points and to calculate further values, e.g. the potential evapotranspiration from the regionalized
datasets. The following processes are simulated by the model and their influence on the waterbalance is
quantified: snow-water-balance, interception, infiltration, soil-water-balance and ground-water-balance. The
runoff concentration and the flood routing are calculated integrative on the basis of subcatchments. The
validation of the modelling quality is done by comparison of the calculated runoff with measured gauge
values. Additional values, e.g. actual evapotranspiration, soilwater-capacity and groundwater-recharge
which are calculated during the model run can be written out spatially differentiated for each time step, which
allows the validation and verification of the model results within the catchment according to measured values.
Throughout the implementation of the model system a mostly open architecture was taken into account, to
make it possible to integrate newer research results into the system to obtain the possibility of further
development.
The model system should be introduced during the presentation and its quality should be demonstrated on
the example of the German part of the catchment of the river Elbe from the Czech border to the river gauge
Barby near Magdeburg (Ac: 40 000 km²).
An analysis of interaction of the climatic factors and annual runoff with empirical
orthogonal functions
Loboda, Nataliya
Department of science and education, Hydrometeorological Institute of Odesa, Lvovskaya 15, Odesa, Ukraine, 65016
The watermanagement transformation of runoff and prospective change of a global climate calls for the
necessity of studying the climatic factors contribution to the runoff formation process. In present paper we
consider and solve the problem of the "climate-runoff" interaction with the use of the empirical orthogonal
functions (EOF) and water-heat balance methods. The approach is illustrated by studying the North-West
Black Sea region , where data about a runoff in a natural condition is not available and, therefore, it is
necessary to develop methods of its account on basis of meteorological information. For modeling of natural
annual runoff series it is possible to use resalts of analysis of annual precipitation and maximally possible
evaporation fields with EOF-method. Spatial distribution of the first four weight expansion coefficients for the
annual flow can be considered as a result of the interaction between the climatic factors in the meso- and
macroscale For river, where there are no measurements of runoff, the coefficients can be calculated as liniar
functions of norms of annual precipitation and maximum possible evaporation. In EOF- model the value of
climatic runoff norm (calculated by the water-heat balance method) is considered as anaverage arithmetical
values of natural runoff .
Matrix and bypass-fluxes in Quaternary and Tertiary sediments of Scheyern
K.-P. Seiler, S. Schneider
GSF-Institute of Hydrology, D-85764 Neuherberg, Germany
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
48
In the unsaturated zone of Tertiary and Quaternary sediments bypass-flow, may become significant as
compared to matrix-flow. Tracer experiments with Deuterium at the FAM Research Station Scheyern show
that bypass-flow exceeds flow velocities of 0.5 m/day, and matrix-flow ranges between 0.7 m/year (Loess),
1.2 m/year (Tertiary gravels and sands) and 2 m/year (Quaternary gravels). In flat areas with a deep
groundwater table in unconsolidated rocks, bypass-flow penetrates to depths lesser than 1m in fine grained
sands and silts and greater than 3m in gravels before it finally incorporates into matrix-flow. In hilly areas with
geologic or manmade permeability interfaces paralleling morphology, bypass-flow generates partly interflow
and partly vertical groundwater recharge; the portion of both has been determined to amount an average of
10: 90% in the research area. Depending on storm intensity and soil water preconditions it may range from 0:
100% to 50: 50 %.
Bypass-flow is favoured by geologic, biologic, pedologic and, in arable lands, also from anthropogenic
factors. It seems to express strongest under very wet and very dry conditions at the soil surface and is less
pronounced in between. It is more dominant in coarse than fine grained-sediments as well as stronger in
fluvial than in lake sediments.
Introducing topological information of Hydrological Response Units for
physically-based distributed catchment modelling
Helmut Staudenrausch
Institut für Geographie, Lehrstuhl Geoinformatik, Friedrich-Schiller-Universität Jena, Löbdergraben 32, 07743 Jena
The basic entity for hydrological analyses is the river basin (catchment, watershed). This unit can be
described as a system with defined input and output. However, its properties controlling the processes
affecting water resources are highly heterogeneous. Therefore, further subdividing this system is a common
method in hydrological applications. A detailed deterministic approach based on Hydrological Response
Units (HRUs, FLÜGEL 1995) is used in this study.
Catchment topology (hydrological connectivity between modelling entities) can be trivial, if simple
distribution methods (e.g. subcatchments) are used. However, such areas are just smaller lumped entities
without unique and homogeneous hydrological properties. Hydrological topology of HRUs, however, can be
complicated, because HRUs are likely to be fragmented throughout the catchment (small number of HRUs,
but large number of polygons); hence the connectivity between HRU polygons and stream segments is a
crucial and complex issue (STAUDENRAUSCH 2000). The aims of this study are (i) to create a methodology to
delineate catchment topology regardless of the distribution method applied, and (ii) testing the topology by
applying it using routing algorithms within a distributed hydrological model.
Creating HRUs is a knowledge-based GIS procedure including selection, reclassification, multilayer overlay,
generalization and assembling input data into an appropriate geodata model, that, however, can be
automated by statistical means. As result one obtains a pattern of distributed, fragmented modelling entities,
all characterized by a unique combination of hydrological properties and parameters. Using this spatial
pattern in conjunction with the stream network, it is possible to derive a tree-shaped network of the
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
49
hydrological topology. This topology creation process is implemented as an automated GIS procedure.
Routing algorithms simulate the downstream attenuation and retention of the hydrological response between
its model entities. In this approach, a complete network setup comprising both the HRUs and the stream
segments, can be considered. Various hydrological (conceptual) routing methods are used to assess the
potential of topological information to improve distributed hydrological modelling on different spatial and
temporal scales. One major advantage of this methodology is the possibility of a 2-level modelling approach:
vertical processes can be simulated on the HRU classes, whereas for lateral flow processes the unique HRU
polygon fragments are considered. This ensures a high degree of detail while keeping the spatial model
setup still relatively simple (STAUDENRAUSCH 2000). The concept is currently tested in two hydrologically
different catchments on varying scales. Initial results are encouraging.
For water resources management applications, in a further step reservoirs, gauging stations, abstraction
points etc. can be incorporated, then considering the HRUs as supply, abstraction points as demand,
gauging stations as validation sites and the network links as transport paths. Based on this network, GIS
network analysis functionality like up-/downstream tracing or location/allocation modelling can be used to
manage water resources (FLÜGEL & STAUDENRAUSCH 1999).
A Comparison of Feedforward Network Topologies for Rainfall-Runoff Modelling
TRUNG-DZUNG NGUYEN
Institute of Irrigation, Drainage and Rural Water Management, University of Rostock, Satower Str. 48, D-18058 Rostock,
Germany.
(From Department of Water Resources Economics, Hanoi Water Resources University (HWRU), 299 Tay-son Street,
Dong-da District, Hanoi, Vietnam)
An important but difficult problem in neural network modelling is the selection of network topology (the
appropriate number of hidden layer and units, the way of connections). In this paper seven feedforward
network topology groups containing 41 typical models were introduced. All topology models are differed in
such aspects: layered/nonlayered, one/two hidden layer, number of hidden units, with/without shortcut
connections, with/without separation of input units in time lag or hydrological stations. All models were
compared on the real data set for rainfall-runoff modelling of Ke-go station (Vietnam) and proved that
standard topology (three-layer feedforward network) is superior to other topologies. The general
(experimental) strategy for selection of appropriate network topology was suggested for further application in
rainfall-runoff modelling.
Modelling water retention at forest sites as a tool for water management
Bott, W., Hofmann, T. & Schenk, D.
Johannes Gutenberg University Mainz, Institute of Geoscience, Division of Applied Geology
The occurrence of floods, with their high economic damages, have become more prominent in the last years.
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
50
Especially the development of measures which decrease the risks of floods is a mean issue of hydrogeologic
research [1].
In this context the significance of forest systems is often figured out concerning their high water retention
potential. In an actual researchproject the approvement of the retention possibilities of two forest sites with
different hydrologic properties under the influence of forestry and hydraulic measures are examined.
Research is focusing on the effect of different wood-types, plant density, road construction, soildegradation,
and drainage network on the runoff. For that purpose the forest system is described by a physically based
hydrological model for continuous simulation of catchment water dynamics (Catflow) [2]. The model
considers parameters which describe the hydraulic behaviour of upper soil layer including the vegetation
canopy and climatic data to calculate runoff, evapotranspiration, interception, infiltration, recharge and
change in soil water content.
The first step was the selection of two forestry catchment test-sites, each representing a typical hydrologic
situation. One area lies within the Soonwald and its build up of Devonian quarzites and clayshistes and their
according weathering products. The second is situated within the Pfälzer Wald build up of Triassic
sandstones. Mean criteria for this selection were the endowment of the testfields (e.g. measurement of soil
water content) and the availablty of landuse informations and longtime series of climatic data. These data are
used as model-input and for the calibration of the output of the model.
The next step was the discretisation of the chosen catchment areas, their drainage geometry and to assign
the climatic and soilspecific input data to the modelsystem. This work was based on a digital height model
and an geographic information system (GIS) including landuse distribution [3].
The first simulations figured out strong relations between changes in landuse patterns respectively of
soilhydraulic parameters and surface runoff. Besides these findings there is a distinct dependence of
waterdynamic in the system and fieldmorphology and climate, especially to precipitation.
Our first results are very encouraging that modelling surface runoff in dependance of forestry measures is
successful. Thus, we shell be able to predict the influence of these measures on stowater-runoff and could
possibly contribute to a sustainable catchment area management.
[1]
[2]
[3]
KOEHLER, G. (Ed.) (1999): Bemessungsabflüsse für kleine Einzugsgebiete.- Bericht 9, Fachgebiet Wasserbau und
Wasserwirtschaft der Universität Kaiserslautern, 395 p.
MAURER, T. (1997): Catflow - A physically based and distributed hydrological model for continuous simulation of
catchment water dynamics, User Guide.- Institute for Hydrology and Water Resources Planning (IHW), University of
Karlsruhe.
GOODCHILD, M. F., STEYAERT, L. T. & BRADLEY, O. P. (Ed.) (1996): GIS and environmental modelling: Progress in
research issues.- Fort Collins, 486 p.
Infiltration study in a soil layer of a dry flood plain using artificial tracers
P. Königer1, M. Karnuth1, L. Jaeger2 & Ch. Leibundgut1
1
University of Freiburg, Institute of Hydrology, Fahnenbergplatz, 79098 Freiburg, Germany
2
Univerity of Freiburg, Meteorological Institute, Werderring 10, 79098 Freiburg, Germany
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
51
Artificial tracing techniques were used to study infiltration in the soil layer of a forested flood plain located in
the upper Rhine valley close to Hartheim, South West Germany (200 m a.s.l). At this site,
micrometeorological observations of evapotranspiration can exceed values calculated by water balance
studies. Due to the flatness of the area there is no surface runoff. Therefore infiltration into deeper zones
seems to be negligible from this point of view. To prove this assumption, investigations were carried out
using oxygen-18 depleted, deuterium enriched and fluorescein traced water, to observe infiltration
processes in the upper soil layer. The investigated area may be classified as almost semiarid, since the
mean annual precipitation and temperature is 667 mm and 9.8°C respectively. River regulation work induced
a lowering of the groundwater table to about 7 m under the surface. These conditions cause water stress for
the pine forest during summer months. The aquifer reaches a thickness of more than 100 m and consists of
quaternary fluvioglacial Rhine gravels. A relatively thin humic silty and sandy soil layer (40 cm on average)
overlays a high permeable unsaturated zone.
Two series of experiments were carried out to follow the movement of infiltrating precipitation in the upper
soil layer under natural and simulated conditions: a deposition of dry fluorescein powder about five cm under
soil surface at four different places and furthermore a sprinkling experiment with deuterium, oxygen-18 and
fluorescein labelled water at two different places. An evaluation of the tracer movement was possible by
excavating the soil profiles to a depth of 1.5 m and an afterwards extraction of soil water in the laboratory.
The soil water was extracted using toluene distillation method for isotope analysis and dimethylformamid for
fluorescein. Additionally, a gravimetric analysis of soil water content in all profiles was done.
The resulting tracer curves in both experiments prove an infiltration of water through the upper soil layer into
the deeper horizon of the unsaturated zone under natural as well as simulated conditions. All tracers used in
the sprinkling experiment could be detected and shows similar behaviours. Due to heavy rainfall in
November, the movement of the tracer under natural conditions was unexpectedly fast. The soil water
profiles give a detailed impression of vertical saturation characteristics.
Investigating real processes like it was done in this study, may contribute to a better understanding of the
complexity of interactions in the unsaturated zone and allow a sound management of the underlying water
resources in many low land areas.
Mean Values of Runoff and Evapotranspiration Calculated for the Hydrological
Atlas of Germany (HAD)
Jankiewicz, P.; Glugla, G.; Rachimow, C.; Richter, K.; Fürtig, G.
Federal Institute of Hydrology, Berlin, Germany
2
Mean values of annual runoff and evapotranspiration for each km in Germany were calculated. Using a
method based on the BAGROV-Relation
æ ETa ö
d ETa
÷÷
= 1 − çç
d Pkorr
è ET max ø
n
,
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
mean actual evapotranspiration
evapotranspiration
ET max
ETa
is determined by corrected precipitation
Pkorr
52
, maximum
depending on land use in case of sufficient water supply and the
BAGROV-parameter n. A great number of measurements with weighable lysimeters all over Germany had
been used to develop the method. Meteorological input data were necessary to determine gras-reference
evapotranspiration, land use parameters were needed to calculate the maximum evapotranspiration. Using
soil parameters (moisture of field capacity and wilting point) the parameter n was set in relationship to
available information being subject in the HAD.
Mean annual runoff finally is the result of the water balance
R = Pkorr − ETa .
These values are evaluated by comparison of measured runoff in defined river basins. Several typical
examples of river basins had been chosen, so that the method is proved to be valid for different cases.
Soil water repellency: a cause of temporal variations in response time to rainfall
events at a range of scales from point to catchment in size?
Doerr, S.H.a, Walsh, R.P.D.a, Shakesby, R.A.a, Ferreira, A.J.D.b, Leighton-Boyce, G.a, and Coelho, C.O.A.b
a
Department of Geography, University of Wales Swansea, Singleton Park, Swansea SA2 8PP, U.K.
b
Departmente de Ambiente e Ordenamento, Universidade de Aveiro, 3810-193 Aveiro, Portugal.
Water repellency is a poorly understood, temporally variable soil property that has been reported from many
regions around the globe. It has been shown to reduce greatly the wettability of soils and has thus been
suggested to cause enhanced overland flow and runoff responses following drier periods when water
repellency is most strongly expressed. This study attempts to demonstrate the impact of water repellency on
the hydrological response of medium-textured forest soils in north-central Portugal using laboratory wetting
and rainfall simulation experiments and runoff response data from long-term runoff plots and small
catchments. Runoff responses were found to be more pronounced for water-repellent soil conditions at all
scales. This effect is particularly marked at a small-plot scale and decreases with increasing size of the area
investigated. This decrease in effect with increased area of investigation is attributed to increased capture of
overland flow by, for example, macropores and areas of wettable soil. It is concluded that soil water
repellency can influence hydrological processes to an extent that they warrant consideration when
investigating and modelling the runoff responses of catchments in regions where soils are prone to water
repellency.
Design floods in small torrential catchments
F. Forster 1), H. Kleindienst 2), St. Vogt 1), A. Badoux 1), Ch. Hegg 1)
1)
Swiss Federal Research Institute WSL, CH-8903 Birmensdorf
2)
GRUNER AG, Ingenieure und Planer, Gellertstrasse 55, CH-4052 Basel
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
53
The estimation of design floods is essential to the planning and design of protection measures in torrents.
The poster shows different possibilities to determine design floods in small catchments. On the one hand the
hydrologic model HQsim is tested and compared with the results of a simple flood estimation procedure and
a statistic extrapolation. The discharge data of four prealpine torrent catchments with areas between 0.75
2
and 3.2 km form the basis of the comparison.
The results show that the extrapolation of measured floods produces rather inaccurate values in small
catchments often characterised by a strong variation of the annual floods and short time series. The simple
estimation procedure seems to be a workable method to quickly get a design flood value. In three out of four
catchments the simple estimation procedure produces the lowest flood values. The application of the
hydrologic HQsim model showed that it cannot be run without calibration, even though the majority of
parameters are physically based. Rather high flood values should be used for a reliable calibration. The time
spent on parameterisation and calibration of the model HQsim is considerable in comparison with the other
two methods. But in this context it is to be mentioned that the model contributes to the better understanding
of hydrological processes in torrent catchments and that it produces besides the peak flow value also a
discharge hydrograph. This hydrograph is essential to answer certain questions dealing with the discharge
volume of a flood.
Out of this situation the demand arose to develop an approach for the definition of design floods combining
the advantages of a simple estimation procedure and those of the model HQsim. The method which is to be
2
applied on small torrent basins (area smaller than 5 km ) should meet the following requirements:
-
different probabilities of occurance can be taken into consideration
-
the time need for its application should not exceed a few days
-
the method enables an objective determination of the design flood, i.e. independent from the user
-
the determination of a discharge hydrograph must be possible
The resulting flood estimation procedure is based on a time-area diagram from the surface which produces
direct runoff and on the subdivision of the whole catchment area into hydrologically homogeneous subareas.
The runoff hydrograph is calculated following the Clark approach. It has be mentioned that the method was
not created to reproduce past flood events but rather to determine design floods and design hydrographs. In
2
general the approach yields satisfying results. In larger catchment areas (> 3 km ) the runoff tends to be
3 -1
-2
overestimated whilst in areas with a high specific runoff (r > 15 m s km ) the discharge is rather
underestimated.
Based on the gained experience a procedure for the determination of design floods in small torrential
catchment areas is being developed.
A comparison of runoff components obtained by measurements of natural tracers
and by digital recursive filtering of hydrographs
K. Eckhardt 1 and S. Uhlenbrook 2
1
Dept. of Agricultural Ecology and Natural Resources Management, Justus-Liebig-University, Giessen, Germany
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
2
54
Institute of Hydrology, University of Freiburg, Fahnenbergplatz, Freiburg, Germany
A recursive digital filter can quickly be applied to a measured hydrograph and give an impression of the mean
groundwater discharge into the stream, but it lacks a true physical basis. By means of the tracer
measurements the filter algorithm developed by Eckhardt (2000) has been tested for a mountain range
2
catchment (Brugga catchment, 40 km , Southern Black Forest Germany) with predominantly shallow soils
above hard rock aquifers. The filter yields a satisfying long-term mean, but is not able to reproduce the
interannual variability of the slow runoff component.
Runoff Generation and Implications for River Basin Modelling, October 9-12, 2000, Freiburg, Germany
55
7 Epilog
The workshop is jointly convened by the International Association of Hydrological
Sciences (IAHS), International Committee on Tracers (ICT), International Committee on
Surface Waters (ICSW) and National Working Group of IHP/OHP-FRIEND/ERB.
The financial support for printing the proceedings from the German Secretariat of
IHP/OHP is gratefully acknowledged.
The chairmen would like to thank the Scientific Advisory Committee for their work.
Beside the chairmen the following persons reviewed abstracts and papers, set up a
common strategy and helped to prepare the sessions and discussion groups:
•
Becker, Alfred, PIK Potsdam, Germany
•
Blöschl, Günter, TU Wien, Austria
•
Lang, Herbert, ETH Zürich, Switzerland
•
Naef, Felix, ETH Zürich, Switzerland
•
Peschke, Gerd, IHI Zittau, Germany
•
Peters, Norman E., USGS, Atlanta, USA
•
Tanaka, Tadashi, University of Tsukuba, Ibaraki, Japan
At last but not least the workshop would not have been possible without the input of
Andrea Ursula Schmitz, Angelika Reichardt, Barbara Heinrich, Franz-Josef Kern,
Ingeborg Vonderstrass, Jens Lange, Jürgen Strub, Paul Königer, Siegfried Demuth,
Volker Abraham, Volker Armbruster, ……..Tilch, Didszun, Tetzlaff, all from the University
of Freiburg, Germany. Thank you!