Development of a coupled water quality model

Authors: WANG, LILI - Purdue University; Flanagan, Dennis; CHERKAUER, KEITH - Purdue University

Submitted to: Transactions of the ASABE
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 5/8/2017
Publication Date: 8/31/2017
Citation: Wang, L., Flanagan, D.C., Cherkauer, K.A. 2017. Development of a coupled water quality model. Transactions of the ASABE. 60(4):1153-1170. doi: https://doi.org/10.13031/trans.12002.
DOI: https://doi.org/10.13031/trans.12002

Interpretive Summary: High production agriculture systems rely upon substantial amounts of agricultural chemicals, most notably fertilizers and pesticides. When rainfall, runoff, and soil erosion events occur, almost always some of these applied chemicals are transported off of the fields in the water or attached to the sediments. With growing concerns related to chemical losses to streams and lakes, particularly losses of phosphorus (P) and nitrogen (N), there is a need to be able to estimate how much of these chemicals are lost, and if there are ways that lands can be managed to minimize or eliminate these losses. Monitoring of fields is one possible way, but is difficult, time-consuming, expensive, and infeasible for large numbers of fields. Alternately, computer simulation models can be used to estimate these chemical losses and the effects of land management practices on reducing them. In this study, a new coupled water quality model was developed, using the process-based WEPP model, and water quality components from another model, the Soil and Water Assessment Tool (SWAT). WEPP allows for very detailed descriptions of the land management practices, predictions of runoff, and predictions of soil loss and sediment yields. The combination of WEPP and the SWAT routines resulted in a coupled WEPP-WQ model that was able to very accurately predict losses of N and P from experimental field studies. This research impacts scientists, students, conservation agency personnel, extension agents, farmers, and others involved in environmental modeling and assessment studies. Application of the WEPP-WQ model may allow for more accurate determination of chemical losses, as well as better management practices to reduce or eliminate the pollutants.

Technical Abstract: Non-point source (NPS) pollutants, especially from agriculture, continue to be a primary source of water quality degradation problems. Effective land management decisions at the field scale must be made to minimize nutrient losses that could pollute streams. Existing NPS models often cannot directly estimate the impacts of different land management practices or determine the effectiveness of combined Best Management Practices (BMPs) in a distributed way at the farm scale. In many cases they rely on application of the Universal Soil Loss Equation (USLE) or its improved versions, which represent fields in a lumped fashion and using empirical rather than process based modeling methodologies. Here a coupled Water Erosion Prediction Project and Water Quality (WEPP-WQ) model is completed, updated, improved and evaluated for simulations of hydrology, soil erosion and water quality. The WEPP model is a well-established process-based model that simulates runoff and erosion processes from a hill slope. The water quality components are based on those of the Soil and Water Assessment Tool(SWAT) model. A single Overland Flow Element (OFE) on a hill slope is used to represent a single land use management on a hillslope. The WEPP-WQ model is tested by comparing simulated values from the coupled model with observed nutrient and sediment concentrations in surface runoff following storm events at an experimental site near Waterloo in northeastern Indiana and at the Throckmorton Purdue Agricultural Center in west central Indiana. Time series evaluation of the WEPP-WQ model was done with observed nutrient and sediment losses from an experimental plot near Tifton, GA. The model performed quite well in simulating nutrient losses for single storm events with the R2 higher than 0.8, Nash-Sutcliffe efficiency (NSE) higher than 0.65, and percent bias (PBIAS) less than 31% for runoff, sediments, nitrate nitrogen, total nitrogen, soluble phosphorus, and total phosphorus losses. In predicting time series nutrient loss, the WEPP-WQ model simulated nitrate nitrogen losses adequately with the ratio of the root mean square error to the standard deviation of measured data (RSR) greater than 0.65, NSE higher than 0.51, and PBIAS greater than -35%. Comparisons between simulated soluble phosphorus, total phosphorus and literature reviewed results were performed due to the absence of an available observational dataset.