AWRA SYMPOSIUM ON GIS AND WATER RESOURCES
Sept 22-26, 1996
Ft. Lauderdale, FL
HYDROLOGIC MODELING TO AID IN LOCATING MONITORING SITES
W. D. Rosenthal (1) and D. W. Hoffman (2)
ABSTRACT
With the increasing importance and awareness of non-poin t source pollution, critical siting of water quality monitoring stations becomes imperative. In this study, the hydrologic model SWAT (Soil and Water Assessment Tool) was used to simulate flows, sediment and nutrient loadings into streams. The model is a continuous, daily time step model that predicts surface runoff, percolation, lateral subsurface flow, groundwater flow, transmission losses and flood routing. The Geographic Resources Analysis Support System (GRASS) and available NRCS (National Resource Conservation Service) databases and the GRASS (Geographic Resources Analysis Support System) provided input into SWAT. Subwatersheds demonstrating areas of significant loading were identified from model output. These areas were correlated to current land use and management practices. Modeled output of streamflow, nitrogen, phosphorus, and sediment loadings were analyzed. Average annual contribution from the entire basin was 3.9 kg ha-1 NO3 and 0.03 kg ha-1 soluble P. The correlation coefficient between observed and simulated streamflow was 0.83. KEY TERMS: Streamflow, loading, nonpoint source, model, TexasINTRODUCTION
Non-point pollution of streams has become a critical concern in many areas. This concern has prompted the U.S. Environmental Protection Agency to commit support to help demonstrate improved stream water quality through implementation of best management practices (BMPs). The current effort towards BMP implementation has largely focused on confined animal feeding operations, such as dairies. To help quantify the success of these practices, water quality samplers have been installed. Locating monitoring sites can be relatively simple in small watersheds. However, the task of locating monitoring sites in much larger watersheds, can be difficult. Consequently, alternative techniques need to beemployed in large watersheds in locating appropriate monitoring sites.Over the past few years, a variety of hydrologic models have been developed, some of which can estimate flow and nutrient concentrations throughout a large watershed. One such model is the Soil and Water Assessment Tool (SWAT) (Arnold et al., 1993). It is a distributed parameter, continuous time model that has been used in many applications (Arnold et al., 1987). The purpose of this study was to demonstrate that the SWAT model, along with available geographical information system (GIS) databases, can aid in locating appropriate monitoring sites in a large watershed.
WATERSHED DESCRIPTION
The Leon River watershed in central Texas covers 9000 km2 and covers five counties--Bell, Coryell, Hamilton, Comanche, and Eastland. A significant number of large dairies have started recently in Hamilton and Comanche counties. Cropland and pasture are the other significant land uses in the watershed. By using the 1:250,000 digital elevation model (available from U.S. Geological Survey)in conjunction with GRASS, 62 subwatersheds were identified (Figure 1).
Figure 1. Subbasin map of the Leon River watershed.
Subwatershed sizes range from 6 to 477 km2. The soils range from loamy in the upper reaches of the watershed to clayey in the central and lower parts.
SWAT DESCRIPTION
SWAT uses a command structure similar to the structure of the Hydrologic Model (HYMO) (Williams and Haan, 1973) for routing runoff and chemicals through a watershed. Commands are included for routing flows through streams and reservoirs, adding flows, and inputting measured data or point pollution sources. The routing command language allows the model to simulate a basin subdivided into grid cells or subwatersheds. Also, output data from other simulation models can be input to SWAT. Using the transfer command, water can be transferred from any reach or reservoir to any other reach or reservoir within the basin. The user can specify the fraction of flow to divert, the minimum flow remaining in the channel or reservoir, or a daily amount to divert. The user can also apply water directly to a subwatershed as irrigation. For the Leon River watershed, manure from the dairies was applied as fertilizer on improved pastureland.SWAT estimates surface runoff volume from a modification of the SCS curve number method (USDA--Soil Conservation Service, 1972). The curve number varies nonlinearly from less than 50 for dry conditions to near 100 for saturated conditions. Percolation occurs when the field capacity of a soil layer is exceeded provided the soil layer below is not saturated. The percolation rate is governed by the saturated conductivity of the soil layer. Percolation from the bottom of the root zone is recharge to the shallow aquifer. A recession constant, as described by Nathan and McMahon (1990) is used to lag flow from the aquifer to the stream. The weather variables necessary for driving SWAT are daily precipitation, air temperature, and solar radiation. If daily precipitation data are available, they can be input directly into SWAT. If not, the weather generator can simulate daily rainfall and temperature. Solar radiation is always simulated. One set of weather variables may be simulated for the entire basin, or different weather may be simulated for each subwatershed within the basin.
RESULTS
Streamflow and sediment and nutrient loadings were estimated using actual weather data from 1965-1994. Figure 2 shows simulated and USGS measured average monthly streamflow at Gatesville, a site approximately 45 km northwest of Lake Belton, the outlet of the watershed. Results indicate that the model does a reasonably good job of estimating flow, as indicated by the simulated vs. USGS measured average daily streamflow at Gatesville, the outlet of the watershed. (Figure 2). Average annual loadings were 3.9 kg ha-1 NO3 and 0.03 kg ha-1 soluble P. The model output indicate that areas in Hamilton and Comanche counties can contribute a significant fraction of the total. From this information, water quality samplers have been installed in the higher-yielding subwatersheds in these counties. Samplers have also been installed in the Fort Hood military reservation subwatershed, where the model indicated significant contributions of sediment.
Figure 2. Average daily streamflow at Gatesville for each month during 1972-1974 (r=0.83).
ACKNOWLEDGMENTS
We wish to acknowledge the U.S. Environmental Protection Agency and the Texas State Soil and Water Conservation Board for funding this project.REFERENCES
Arnold, J., B. A. Engel, and R. Srinivasan. 1993. A continuous time, grid cell watershed model. IN Application of Advanced Information Technologies for Management of Natural Resources. Spokane, WA, 17-19 June 1993. American Society of Agricultural Engineers.Arnold, J. G., M. D. Birchet, J. R. Williams, W. F. Smith, and H. N. McGill. 1987. Modeling the effects of urbanization on basin water yield and reservoir sedimentation. Water Res. Bull. 23(6):1101-1107.
Nathon, R. J. and T. A. McMahon. 1990. Evaluation of automated techniques for base flow and recession analyses. Water Res. Research 26(7):1465-1473.
USDA--Soil Conservation Service. 1972. National Engineering Handbook, Hydrology section. Chapters 4-10. Washington, D.C.
Williams, J. R. and R. W. Haan. 1973. HYMO: Problem oriented computer language for hydrologic modeling. USDA ARS-S-9. Washington, D. C. 1. D. Rosenthal, Assistant Professor, Texas Agric. Expt. Stn., Blackland Research Center, 808 E. Blackland Rd., Temple, TX 76502 2 D. W. Hoffman, Research Scientist, Texas Agric. Extension Service, Blackland Research Center, 808 E. Blackland Rd., Temple, TX 76502
home pageMaintainer: AWRA Webserver Team
Last modified: 24Nov99 gaw
Copyright(c) 1996, American Water Resources Association