|BIEGER, KATRIN - Aarhus University|
|ALLEN, PETER - Baylor University|
|GAO, JUNGANG - Texas Agrilife Research|
|CERKASOVA, NATALJA - Texas A&M University|
|PARK, SEONGGYU - Texas A&M University|
|Bosch, David - Dave|
|YEN, HAW - Texas A&M Agrilife|
|OSORIO, JAVIER - Texas Agrilife Research|
Submitted to: Journal of the American Water Resources Association
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 8/4/2022
Publication Date: 8/23/2022
Citation: White, M.J., Arnold, J.G., Bieger, K., Allen, P.M., Gao, J., Cerkasova, N., Gambone, M.A., Park, S., Bosch, D.D., Yen, H., Osorio, J.M. 2022. Development of a field scale SWAT+ modeling framework for the contiguous U.S. Journal of the American Water Resources Association. 1-16. https://doi.org/10.1111/1752-1688.13056.
Interpretive Summary: The Soil and Water Assessment Tool (SWAT) model is commonly used to predict how agriculture impacts water. In this research, we develop the National Agroecosystems Model (NAM), a very detailed SWAT+ model that covers the lower 48 states. NAM contains 7 million computational land surface units and 3 million individually identifiable stream segments and more than 5,000 reservoirs. This framework is tested in a case study of the Little River Watershed, Tifton, GA. NAM is developed using only public data and can be shared with other government agencies, universities and other research groups.
Technical Abstract: The Soil and Water Assessment Tool (SWAT) model is commonly used to predict the impacts of agricultural practices on water quality and quantity. Although widely applied, the data framework used to drive SWAT in the U.S. is fragmented and inconsistent, varying by user and model interface. This research describes the development of the National Agroecosystems Model (NAM), which provides a unified field to national scale modeling computational framework for research and decision support by using the latest SWAT platform, dubbed SWAT+. NAM has sufficient detail to capture field-level processes and management actions, and spans the full extent of the contiguous U.S. NAM contains 7 million computational units, 4 million of which represent specific cultivated fields. It contains 3 million individually identifiable stream segments and more than 5,000 reservoirs. This work describes the individual data sources, assumptions, and the processing steps for their inclusion. NAM is constructed with 2,121 individual SWAT+ models which can be executed in a parallel hierarchical structure to dramatically improve runtime. This framework is tested in a case study of the Little River Watershed, Tifton, GA. NAM is developed using only publicly available data sources such that subsets of it can be shared to support research with other government agencies, universities, and others in the public domain.