This is not a peer-reviewed paper.

Pp. 123-126 in Research for the 21st Century, Proc. Int. Symp. (3-5 January 2001, Honolulu, HI, USA). Eds. J.C. Ascough II and D.C. Flanagan. St. Joseph, MI: ASAE.701P0007

Watershed Application of WEPP to a Michigan Water Quality Problem

R.C. Vining, D.C. Flanagan, J. Grigar1

Abstract

The USDA Water Project (WEPP) model allows simulation of overland flow and processes at the small watershed scale. Capabilities within WEPP allow evaluation of various parameters influencing water quality. The WEPP model was applied to a small agricultural watershed in southern Michigan to test the feasibility of using its hydrologic and sedimentation outputs to estimate phosphorous loadings at the mouth of the watershed. WEPP provides detailed daily estimates of runoff, soil loss, delivery and sediment particle size distributions. Simple techniques are available to correlate phosphorous loadings to sediment yield and discharge volumes from a watershed and its individual components. Using procedures provided by the Michigan Department of Environmental Quality and outputs from WEPP, we can predict phosphorous loading at the mouth of the watershed, and identify areas within the basin making the highest contributions to the total phosphorous load. With the new user-friendly WEPP Windows interface, users can test the effectiveness of different management systems to reduce runoff and sediment delivery, thereby reducing sediment-bound and soluble phosphorous delivered to the mouth of a watershed. Keywords. Soil erosion, WEPP, Watersheds, Water quality, Phosphorous.

Introduction

The watershed version of the Water Erosion Prediction Project (WEPP) model (Flanagan and Nearing, 1995) provides a tool to allow evaluations of soil erosion and sediment delivery over a combination of hillslopes and concentrated flow channels (Nearing et al., 1989; Ascough, 1997). Loading of sediment-attached contaminants such as phosphorous can be estimated from WEPP watershed outputs. This study demonstrates an application of WEPP to a small agricultural watershed in Michigan, in part using data files and slope measurement techniques from the Revised Universal Soil Loss Equation (RUSLE). WEPP sediment delivery and discharge volume provide the base by which losses of phosphorous from the test watershed can be evaluated.

Model Inputs

Brief Description of Michigan Watershed

The study watershed is located in Kalamazoo County, Michigan, east of the city of Kalamazoo. Size of the watershed is approximately 42 ha. in the watershed are Kalamazoo loam and Oshtemo sandy loam. The climate of the area is classified by Koppen as high latitude, uniform precipitation, warm summer (Dfb) (Griffiths and Driscoll, 1982). Mean annual precipitation is about 84 cm, and mean annual temperature is about 10 °C. Field slopes range from 2-6%, with steeper slopes in woods and pasture from 11 to 18%. Primary agricultural production rotations in the watershed are corn-soybean, corn-wheat-soybean, and corn-hay. Soil phosphorous levels are maintained at 1000 kg/ha.

Application of WEPP to the Watershed

Individual field lengths and slopes were measured using techniques described by Renard et al. 1997. Only critical slope of each hillslope was measured, with only a single slope angle and length being used to represent each hillslope; thus, total watershed area was underestimated with WEPP. Characteristics for channels for concentrated flow routing were measured in the same fashion. Soils data were collected from the National Soil Survey index of soils. Thirty years of climate data were generated using the CLIGEN climate generator (Nicks

1 Roel C. Vining, Hydrologist, USDA-NRCS, National Soil Erosion Research Laboratory, West Lafayette, IN; Dennis C. Flanagan, Agricultural Engineer, USDA-ARS-MWA, National Soil Erosion Research Laboratory, West Lafayette, IN; Jerry Grigar, State Agronomist, USDA-NRCS, Lansing, MI. Corresponding Author: Roel C. Vining, USDA-NRCS, National Soil Erosion Research Laboratory, 1196 SOIL Building, West Lafayette, IN, 47907-1196; tel.: (765) 494-8691; fax: (765) 494-5948; e-mail: .

123 and Gander, 1995) with input data from Ft. Wayne, IN. Management files, including operations, dates, and yield goals, were derived from NRCS Field Office Technical Guide RUSLE management files for Michigan. Figure 1 shows a screen from the WEPP Windows interface, displaying the watershed simulated in this study. The new interface provides an easy way to construct a watershed, by adding and hillslope elements. In an actual model simulation, WEPP “sees” the watershed as a series of rectangles that generate and transport runoff and sediment.

Figure 1. Layout of the Michigan test watershed in the WEPP Windows interface.

Results

The WEPP model predicts runoff, soil detachment and sediment deposition on hillslope profiles, runoff, soil detachment and sediment deposition on channel elements, and runoff and sediment delivery from a watershed outlet. The average annual predicted discharge volume and sediment yield from the mouth of the test watershed are shown in table 1 for both conventional tillage (plow/plant) and conservation tillage (no-till/chisel).

Table 1. WEPP model version 99.5 predicted discharge and sediment yield from the watershed under conventional and conservation tillage Tillage System Discharge Volume (m3/yr) Sediment Yield (tonne/yr) Conventional Tillage 5901 25.9 Conservation Tillage 4178 4.8

Estimated sediment yields for the watershed are rather low, particularly when conservation practices are implemented, reflecting both the low angle of the tilled slopes, and the conservation tillage practices in place in the fields. Also, because RUSLE derived hillslope values represent only the critical eroding portion of the slope, the WEPP measured watershed underestimated the total size of the watershed (19 ha versus 40 ha actual size).

124 Given the sediment yield and the soil phosphorous test level, a gross prediction of attached phosphorous transport can be made using a technique developed by the Michigan Department of Environmental Quality, (1999): Phosphorous loading = tonnes/yr sediment * kg P/kg soil * 1000kg/tonne * 0.85 (1) where 0.85 is a correction factor suggested for loamy soils. Using the predicted sediment yield of 25.9 tonnes/yr predicted from the conventional tillage example, total attached phosphorous load would be 10.1 kg/yr. By applying a conservation tillage system to the tilled fields in the watershed, sediment yield is reduced to 4.8 tonne/yr, reducing the estimated attached phosphorous to 1.9 kg/yr, a reduction in phosphorous load of 8.2 kg/yr, or 81%. Erosion and phosphorous loadings can be estimated for each hillslope in the watershed using the WEPP- watershed output. For brevity, let’s consider only the hillslopes under intensive agricultural rotations (corn- soybean and corn-soybean-wheat) as these are the hillslopes where the majority of erosion is occurring. Model predicted average annual sediment yield from these 10 hillslopes for the two tillage systems is presented in Figure 2.

35000 Conventional Tillage 30000 Conservation Tillage 25000

20000

15000

10000

Sediment Delivery (kg/yr) Delivery Sediment 5000

0 H18 H19 H20 H21 H23 H24 H25 H26 H27 H29

Figure 2. Model predicted average annual sediment yield for tilled hillslopes under conventional and conservation tillage systems.

As would be expected, application of conservation tillage greatly reduced the estimated sediment yield. Greatest reductions were predicted to occur on hillslope 23 (H23), which was the steepest and longest of the hillslopes under a corn-soybean-wheat rotation. The technique applied to the watershed output to estimate phosphorous loading attached to sediment (equation 1) can also be applied to the individual hillslopes. Using H23 as an example, predicted phosphorous loading was reduced from 12 kg/yr under conventional tillage to 0.28 kg/yr with conservation tillage. Phosphorous load reductions would be predicted on all hillslopes noted in Figure 2, although the magnitude of the reductions would be far less than those estimated for H23.

Conclusions

The WEPP model has been shown to provide reasonable estimates of runoff and sediment yield when applied to small agricultural watersheds (Liu et al., 1997; Cochrane and Flanagan, 1999). The WEPP Windows interface provides a relatively simple method of setting up and running the WEPP model for watershed applications, and it is easy to view the outputs, both from the watershed, and from the individual hillslopes and channels. WEPP gave a good analysis of the values needed to make an estimation of phosphorous loading from runoff in the watershed. This study applied phosphorous loadings to the WEPP output after the model had been run; small changes in the WEPP model code could allow for water quality analysis to be conducted during the actual simulation runs.

125 Using RUSLE inputs limited certain aspects of WEPP. Chief among these was applying the method of measuring slope length and angle from RUSLE to WEPP. Whereas when using RUSLE, only the main eroding portion of the slope is measured, WEPP allows for the entire slope, from the watershed boundary to the concentrated flow channel, to be input. Use of the RUSLE inputs resulted in WEPP being run over a 50% smaller area (19.17 ha) than that of the full watershed (about 40 ha). A more representative survey of the watershed hillslopes would provide larger estimates of watershed discharge volume and sediment yield, improve the predicted pnosphorous loads attached to sediment carried from the watershed. Further work in applying WEPP to analysis in this watershed is to conduct a more detailed survey of the hillslopes and channels in the watershed, so that the entire watershed area is represented in WEPP. Additionally, a method for applying WEPP through the automated identification of hillslopes and concentrated flow channels (Cochrane and Flanagan, 1999) would help to reduce the underestimation of the watershed area, while simplifying the problems of identifying individual hillslopes and assigning the appropriate slope and length to each hillslope and channel.

References

Ascough II, J.C., C. Baffaut, M.A. Nearing, B.Y. Liu. 1997. The WEPP watershed model: I. and Erosion Transactions of the ASAE 40(4): 921-933.

Cochrane, T.A., and D. C. Flanagan. 1999. Assessing water erosion in small watersheds using WEPP with GIS and digital elevation models. Journal of Soil and Water Conservation 54(4): 678-685.

Flanagan, D.C. and M.A. Nearing (eds.). 1995. USDA-Water Erosion Prediction Project: Hillslope Profile and Watershed Model Documentation. NSERL Report No. 10. West Lafayette, IN: USDA-ARS National Soil Erosion Research Laboratory.

Griffiths, J.F., and D.M. Driscoll. 1982. Survey of Climatology. Charles E. Merrill Publishing Co. Columbus, OH. 358 pp.

Liu, B.Y., M.A. Nearing, C. Baffaut, J.C. Ascough II. 1997. The WEPP watershed model: III. Comparisons to measured data from small watersheds. Transactions of the ASAE 40(4): 945-952.

Michigan Department of Environmental Quality. 1999. Pollutants Controlled: Calculation and Documentation for Section 319 Watersheds Training Manual. Lansing, MI.

Nearing, M.A., G.F. Foster, L.J. Lane and S.C. Finker. 1989. A process-based soil erosion model for USDA- Water Erosion Prediction Project Technology. Transactions of the ASAE 32(5):1587-1593.

Nicks, A.D. and G.A. Gander. 1995. Chapter 2. Weather generator. In D.C. Flanagan and M.A Nearing (eds.): USDA-Water Erosion Prediction Project: Hillslope Profile and Watershed Model Documentation. NSERL Report No. 10. West Lafayette, IN: USDA-ARS National Soil Erosion Research Laboratory.

Renard, K.G., G.R. Foster, G.A. Weesies, D.K. McCool, and D.C. Yoder, coordinators. 1997. Predicting SoilErosion by Water: A Guide to Conservation Planning with the Revised Universal Soil Loss Equation(RUSLE). U.S. Department of Agriculture, Agriculture Handbook No. 703, 404pp.

126