1. General Model Information

Name: Soil and Water Integrated Model

Acronym: SWIM

Main medium: terrestrial
Main subject: hydrology, biogeochemistry
Organization level: Ecosystem
Type of model: not specified
Main application: research
Keywords: watershed, basin scale, spatially distributed, runoff,groundwater, water quality, crop growth, nutrient cycling, nutrienttransport, erosion, sediment transport, climate change, land usechange, continuous-time, multiple subbasins, multiple hydrotops, threelevel spatial disaggregation, Priestley-Taylor evapotranspiration, modified Curve-Number-runoff,GIS, GRASS interface, soil database, SWAT


Valentina Krysanova
Potsdam Institute for Climate Impact Research
P.O.Box 601203, Telegrafenberg
14412 Potsdam, Germany

Phone: +49-(0)331-288-2515
Fax: +49-(0)331-288-2600
email: valen@pik-potsdam.de


SWIM:V. Krysanova, D.-I. Müller-Wohlfeil, A.Becker
Potsdam Institute for Climate Impact Research
SWAT-Modules:J.G. Arnold, P.M. Allen, G.T. Bernhardt, R. Srinivasan,R.S. Muttiah, C. Walker, P.T. Dyke, 1993, USDA & Texas A&MUniversity
MATSALU-Modules: V. Krysanova, A. Meiner, J. Roosaare, A. Vasilyev, 1989, Estonian Ac. Sci.


Model Objectives: to simulate hydrological cycle, erosion, vegetation growth and nutrient transport in mesoscale watersheds (from 100 km2 to 20,000 km2); to analyse climate change and land use change impacts on hydrology and water quality at the regional scale.

Approach: A three-level scheme of spatial disaggregation "basin - subbasins - hydrotops" is implemented. The soil profile is represented by maximum ten soil layers. SWIM/GRASS interface is used to initialize the model by extracting distributed parameters of elevation, land use, soil, climate, and to create hydrotop structure and routing structure files. The weather parameters necessary to drive the model are daily precipitation, air temperature (average, minimum and maximum), and solar radiation. Weather data can be taken from meteorological stations or produced by a weather generator based on the monthly statistical data. The model operates with a daily time step. SWIM can be used either for hydrological modelling only, or for integrated hydrological/crop, hydrological/erosion, hydrological/water quality modelling.

Background: SWIM is based on two previously developed models - SWAT (Arnold et al, 1993) and MATSALU (Krysanova et al., 1989). Direct application of the both models for German watersheds was not possible due to several reasons. The main reason is connection of SWAT to specific American data sets (especially for soil, weather, and crop rotation parameters), and not sufficient transferability of MATSALU (which is a system of four coupled models developed for the Matsalu Bay watershed in Estonia). SWIM includes modules from both predecessors, trying to combine their advantages (hydrological submodel and GRASS interface from SWAT; spatial disaggregation scheme and nutrient modules from MATSALU), and to avoid overparametrization.

Processes: The following hydrological processes are included: precipitation, snow melt, evapotranspiration, surface runoff, lateral subsurface flow (interflow), percolation to ground water, ground water contribution to streamflow, streamflow routing. The following geo- and hydrochemical processes are included: input of fertilizers, mineralization, denitrification and nitrification, sorption/desorption (for phosphorus), crop uptake of nutrients, leaching to ground water, transport with surface flow, transport with subsurface flow.

The snow melt is calculated using a simple degree-day equation with the snow melt rate equal to 4.57 mm/degree/day. Melted snow is then treated the same as rainfall for further estimation of runoff and percolation.

Potential evapotranspiration is estimated using Priestley-Taylor method (1972), which requires solar radiation and air temperature as input. The actual evapotranspiration is estimated following the Ritchie (1972) concept separately for soil and plants. Plant transpiration is simulated as a linear function of potential transpiration and leaf area index, taking into account soil water availability.

Surface runoff is estimated using a modification of the SCS Curve Number method (USDA-SCS, 1972, Arnold et al, 1990). Surface runoff is predicted as a nonlinear function of precipitation and retention coefficient. The retention coefficient depends on land use, soil type, soil water content and slope. Despite of its empirical nature, after this modification the method is quite reliable, and was successfully tested in US watersheds (SWAT) and in the Elbe subbasins (SWIM). It is possible to eliminate the CN method in SWIM by excluding the dependence of the retention coefficient on land use and soil type.

Lateral subsurface flow and percolation to ground water are calculated using the same storage routing technique as in the SWRRB model (Arnold et al., 1990), which is different from the current version included in SWAT model. Downward flow occurs when field capacity of the soil layer is exceeded, and the layer below is not saturated. The flow rate is governed by the saturated hydraulic conductivity of the soil layer. Lateral flow occurs when the storage in any layer exceeds field capacity after percolation. A nonlinear function of lateral flow travel time is estimated from saturated conductivity and vertical equivalent hydraulic conductivity.

Groundwater component in SWAT and SWIM is based on the equation for return flow from Smedema and Rycroft (1983). It assumes that the variation in return flow is linearly related to the rate of change of the water table height. In a finite difference form, the return flow is a nonlinear function of ground water recharge and the reaction factor, which is a direct index of the intensity with which the groundwater outflow responds to changes in recharge. The reaction factor can be estimated using the base flow recession curve.

Streamflow routing is based on the ROTO model (Arnold, 1990). Flow rate and average velocity are calculated using Manning's equation for the full channel depth, for a depth of 1.5 times the full depth, and for a depth of 0.1 the full depth. Travel time is then related to flow in these two intervals (0.1, 1) and (1, 1.5) using the nonlinear relationship TTIME = a* Q**b. Then the storage coefficient is estimated as a nonlinear function of travel time (Williams and Hann, 1972). The method was slightly modified for SWIM in order to get better fit to the measured discharge with daily time step in our case studies.

Sediment yield is calculated for each subbasin with the Modified Universal Soil Loss Equation (MUSLE) (Williams and Berndt, 1977). The equation includes the runoff factor, the soil erodibility factor, the crop management factor, the erosion control practice factor, and the slope length and steepness factor. To estimate the daily rainfall energy in the absence of time-distributed rainfall, the assumption about exponential distribution of the rainfall rate was made. This allows a simple substitution of rainfall rates into the equation. Soil erodibility factor can be estimated from the texture of upper soil layer. The slope length and steepness factor can be estimated from the DEM.

Crop growth. A simplified EPIC approach (Williams et al., 1984) is used for simulating all the crops and natural vegetation, using parameter values for each plant type from the database. Interception of solar radiation is estimated with Beer's law equation (Monsi and Saeki, 1953) as a function of solar radiation and leaf area index. The potential increase in biomass is estimated as the product of intercepted energy and a specific crop parameter for converting energy into biomass. The potential biomass is adjusted daily if one of the stress factors (water, temperature, N, P) is less than 1.0.

Mineralization. An approach used in CREAMS and MATSALU models is accepted. First, the nitrogen associated with the soil humus is divided into active and stable pools. Mineralization occurs only in the active pool and is simulated as a function of organic N, soil temperature and water content. Besides, nitrogen from the stable pool is allowed to flow slowly into the active pool.

Denitrification. An empirical formula the same as in CREAMS and MATSALU models is used to define the denitrification rate as a function of soil temperature, organic matter, soil wetness, and mineral nitrogen content.

Nitrification is simulated as in MATSALU model, using an empirical formula of Bieleck (Balance..., 1986). The process is limited by temperature and a soil wetness index (soil water content / field capacity).

Sorption / desorption of P. The approach of Jones et al. (1984) is used to describe the distribution of mineral P among three pools: labile, active mineral and stable mineral. Flow between the labile and active mineral pools is governed by temperature, soil moisture, and a sorption coefficient. Flow between the active and stable mineral P pools depends on the concentrations in each pool and the sorption coefficient.

Crop uptake of nutrients. A supply and demand approach is used. The daily crop N and P demand is estimated as the product of biomass growth and optimal nutrient concentration in plant. The latter depends on the stage of plant development. Soil supply of N and P (labile) is defined by current content in the root zone. Actual nutrient content is the minimum of supply and demand.

Sediment transport of P is estimated based on the concentration of sediment-bound P in the top soil layer, the sediment yield, and the enrichment ratio.

Transport with surface, subsurface flow and leaching. The amount of NO3-N in runoff is estimated by considering the top soil layer only. Amounts of NO3-N in direct runoff, lateral subsurface flow and percolation are estimated as the products of the volume of water and the average concentration.

Spatial disaggregation and integration with GIS: A three-level disaggregation scheme implies basin, subbasins, and hydrotops. As a first step, the whole basin area has to be subdivided into subbasins (or raster cells) of a reasonable average area. The river network connects the subbasins and is used for routing water, sediments, and nutrients. An average subbasin area, where the effect of the river network may be neglected, should be in a range of 10 to 100 km2. The subbasin boundaries can be taken from existing maps or derived in GIS (for example, using the r.watershed program in GRASS).Then hydrotops are delineated within every subbasin, based on land use and soil types. The hydrotop is a set of (disconnected) units in a subbasin, which have a unique land use and soil type. The SWAT/GRASS interface (Srinivasan, Arnold, 1993) is adopted and modified (Steps 3 & 4) for SWIM to extract spatially distributed parameters of elevation, land use, soil types, groundwater table and to create hydrotop structure and routing structure files. The interface creates a number of input files for the basin and subbasins.

1. Subbasin attributes. Using a given subbasin map, the program calculates area, resolution, coordinate boundaries for the basin and each subbasin, and the fraction of each subbasin area to the basin area.

2. Topographic attributes. The program estimates the stream length, stream slope and geometrical dimensions, accumulation area, and aspect. The weighted average method is used to estimate the overland slope and slope length. Finally, the channel factors K and C of the Universal Soil Loss Equation (USLE) are estimated using a standard table.

3. Hydrotop structure. The program defines the basin structure by overlaying the subbasin map with land use and soil layers.

4. Weather attributes. The program selects the closest weather/precipitation station to the subbasin. Then either actual weather information, or weather generator can be used.

5. Ground water attributes. The ground water parameters are estimated for each subbasin using the alpha layer, which defines the time lag needed to the groundwater flow as it leaves the shallow aquifer to return to the stream.

6. Routing structure. The routing structure is created for subbasins, using the elevation map. Also, it defines the channel width and depth using a neural network that is embedded in the interface, based on the drainage area and average elevation of a subbasin. There is a possibility to edit the routing structure in case of any problems caused by insufficient accuracy of the DEM.

Spatial and relational data: Spatial data necessary to run the model, are:

The recommended DEM resolution can be estimated using the "thousand-million" rule (Maidment, 1996). A subbasin map must be created in advance. In addition, the following maps can be used. The river network map is useful for checking the routing structure. The map of ground water tables is needed if ground water height has to be included in the model outputs. Maps of climate/precipitation stations and river gages are more important for larger basins. A map of point sources of pollution is necessary in case they contribute a significant part in the river load and must be taken into account. All the maps should be provided in ARC/INFO or GRASS format.

The full list of necessary relational data includes

climate data:

hydrological data:

soil data base, including the following parameters for every of maximum ten soil layers:

crop management parameters:

water quality data at the basin outlet.

In case if weather generator has to be used instead of actual weather data, monthly statistical data for climate stations must be provided. In case if saturated conductivity is not available, it can be estimated in the model as dependent on soil texture, bulk density and organic carbon content.

Modelling procedure: After the input parameters are read from the input files, the three-step modelling procedure is applied:

  1. water and nutrient balance are calculated for every hydrotop
  2. the outputs from hydrotops are averaged (weighted average) to estimate the subbasin output, not accounting for lag time in the case of surface runoff, and taking the average for subbasin lag time for subsurface flow,
  3. routing procedure is applied to the subbasin outputs, taking into account transmission losses.

II. Technical Information

II.1 Executables:

Operating System(s): UNIX uses the Geo Information System GRASS.

II.2 Source-code:

Programming Language(s): programming languages:

II.3 Manuals:

SWAT: Arnold, J.G., J.R. Williams, R. Srinivasan, K.W. King, andR.H. Griggs, 1995: SWAT - Soil and Water Assessment Tool: DraftUsers Manual, USDA-ARS, Temple, TX. SWIM: not yet available SWATGRASS MANUAL ON-LINE

II.4 Data:

see the description in section I

III. Mathematical Information

III.1 Mathematics

III.2 Quantities

III.2.1 Input

III.2.2 Output

IV. References

Krysanova, V., D.I. Müller-Wohlfeil & A. Becker 1998: Development and test of a spatially distributed hydrological / water quality model for mesoscale watersheds. Ecological Modelling 106 (1-2), 261-289.

Krysanova, V., Becker & Klöcking, B. 1998: The linkage between hydrological processes and sediment transport at the river basin scale. In W.Summer E.Klaghover, W.Zhang (eds.) Modelling Soil Erosion, Sediment Transport and Closely Related Hydrological Processes. IAHS Publications no. 249, p. 13-20.

Krysanova, V. & A. Becker 1999: Integrated Modelling of Hydrological Processes and Nutrient Dynamics at the River Basin Scale, Hydrobiologia 410, 131-138.

Krysanova,V., Wechsung, F., Becker, A., Poschenrieder W. & Gräfe J. 1999: Mesoscale ecohydrological modelling to analyse regional effects of climate change, Environmental Modelling and Assessment 4, 259-271.

Krysanova, V., Gerten, D., Klöcking, B., & Becker, A. 1999: Factors affecting nitrogen export from diffuse sources: a modelling study in the Elbe basin. In: L. Heathwaite (ed.) Impact of Land-Use Change on Nutrient Loads from Diffuse Sources, IAHS Publications no. 257, p. 201-212.

Krysanova, V., F. Wechsung, A. Meiner & A. Vasilyev, 1999: Land use change in Europe and implications for agriculture and water resources. In: Ü. Ennuste, L. Wilder (eds.) Harmonisation with the western economics: Estonian developments and related conceptual and methodological frameworks, Estonian Institute of Economics at Tallinn Technical University, 361-384.

Wechsung, F., Krysanova, V., Flechsig M. & Shaphoff, S. 1999: May land use change reduce the water deficiency problem caused by reduced brown coal mining in the state of Brandenburg? Landscape and Urban Ecology (in print)

Krysanova, V., Williams, J., Bürger, G. & Österle, H. 1999: Linkage between hydrological processes and sediment transport at the river basin scale - a modelling study. UNESCO Technical Report on Hydrological Processes and Soil Erosion (accepted).

Arnold, J.G., P.M. Allen, and G.T. Bernhardt, 1993: A ComprehensiveSurface-Groundwater Flow Model: Journal of Hydrology, v. 142,p. 47-69.

Arnold, J.G., B.A. Engel, and R. Srinivasan, 1993: A ContinuousTime, Grid Cell Watershed Model: in Proceedings of Applicationof Advanced Technology for the Management of Natural Resources,Sponsored by American Society of Agricultural Engineers, June17-19, 1993, Spokane, WA.

Arnold, J.G., J.R. Williams, R. Srinivasan, K.W. King, and R.H.Griggs, 1995: SWAT - Soil and Water Assessment Tool: Draft UsersManual, USDA-ARS, Temple, TX.Krysanova, V. & Luik, H. (eds.) 1989a: Simulation modellingof a system watershed - river - sea bay. Tallinn, Valgus, 428pp, (in Russian).

Krysanova, V., Meiner, A., Roosaare, J., & Vasilyev, A.1989b: Simulation modelling of the coastal waters pollutionfrom agricultural watershed. Ecological Modelling, 49, 7-29.

Krysanova V., Müller-Wohlfeil D.I., Becker A., 1996.Mesoscale Integrated Modelling of Hydrology and Water Qualitywith GIS Interface. In: Proceedings of the Third InternationalConference and Workshop on Integrating Geographical InformationSystems and Environmental Modeling, CD-ROM. Publisher: Santa Barbara,CA: National Center for Geographical Information and Analysis.

Krysanova, V., D.-I. Müller-Wohlfeil, A. Becker, 1996.Integrated Modelling of Hydrology and Water Quality in MesoscaleWatersheds. PIK Report No.18, July 1996, 32p.

Krysanova, V., D.-I. Müller-Wohlfeil, A. Becker, 1996.Entwicklung eines integrierten Modells zur Hydrologie undGewässerg¸tein mesoskaligen Einzuggebieten und dessen Anwendung in Teileinzugsgebietender Elbe, Tagungsband zum Vorseminar "Gewässer -Informationssysteme"- Datenmanagement und Modellierung von Stromlandschaften, Budweis,Tschechische Republik, 82-84.

Srinivasan, R., and Arnold, J.G., 1993. Basin scale water qualitymodelling using GIS. Proceedings, Applications of AdvancedInform. Technologies for Manag. of Nat. Res. June 17-19, Spokane,WA, USA.

Srinivasan, R., and Arnold, J.G., 1993. Basin scale water qualitymodelling using GIS. Proceedings, Applications of AdvancedInform. Technologies for Manag. of Nat. Res. June 17-19, Spokane,WA, USA.

V. Further information in the World-Wide-Web

selected other examples for GIS-based watershed modeling are

VI. Additional remarks

Information and remarks about the application of this model

Last review of this document by: 17.2.1997
Status of the document: Updated and revised by Dr. Valentina Krysanova
last modified by Tobias Gabele Wed Aug 21 21:44:51 CEST 2002

Go back to Register of Ecological Models (R E M)