Prioritization of the agricultural fields in the Little Arkansas river watershed, KS, with a SWAT+ model for off-site BMP program

Abstract

Soil erosion from agricultural land and transport of sediment into streams create serious water-quality problems, especially in watersheds that supply drinking water. The Little Arkansas River watershed (LTAK) in central Kansas supplies water to the City of Wichita and is subject to high sediment loads due to intensive farming, sloping land, and high hydrological connectivity between agricultural fields and stream networks. Addressing soil erosion and the associated transport of sediment into streams requires both clear identification of sediment sources and a reliable framework to prioritize agricultural fields for conservation intervention. The Off-site Best Management Practice (BMP) program, a state-sponsored initiative that incentivizes the implementation of BMPs on agricultural lands, provides the policy and operational context within which targeted conservation intervention can be implemented. The main goal of this study was to develop, calibrate, and apply a SWAT+ watershed model in the LTAK to (1) estimate sediment contributions from agricultural areas, (2) evaluate sediment reduction potential from selected BMPs, and (3) prioritize agricultural fields for BMP implementation to meet sediment reduction goals linked to urban water-quality protection. The SWAT+ model was developed using spatial and environmental datasets, including a 10-m digital elevation model, land-use data from the NASS Cropland Data Layer, KARS sub-CLU agricultural field boundaries, STATSGO2 soils, HUC-12 sub-watershed boundaries, and observed hydro-meteorological data. Model calibration and validation followed a multi-site approach, incorporating observed streamflow from four locations: Alta, Halstead, Emma, and Sedgwick. Model performance was evaluated at both daily and monthly time steps using standard statistical measures, including Nash Sutcliffe Efficiency (NSE), percent bias (PBIAS), Kling Gupta Efficiency (KGE), and root mean square error (RMSE). To evaluate the effectiveness of conservation practices, two management scenarios were developed and analyzed, comparing baseline agricultural conditions under which corn was uniformly applied across all croplands. The alternative conservation scenarios are no-till alone (BMP1) and a combination of no-till and cover crops (BMP2), with sediment reductions estimated at field and HUC-12 sub-watershed scales. Sediment outputs were converted from Hydrologic Response Units (HRUs) to field-based estimates and analyzed across multiple spatial scales to capture the spatial variability of erosion processes across the landscape. Additional analysis was conducted by selecting the 10 highest sediment-producing subbasins and further refining these areas using clustered hotspot identification. The results revealed strong spatial heterogeneity in sediment production, with a relatively small number of areas contributing a large share of total sediment yield within the watershed. Building on these spatial findings, a Monte Carlo-based field selection framework was implemented to evaluate BMP implementation strategies. Field selection was performed across three spatial conditions: (1) all subbasins within the watershed, (2) selected high sediment-producing subbasins, and (3) clustered hotspot subbasins. For each condition, 1000 simulation runs were conducted, randomly selecting fields until a sediment reduction target of 160 tons per year, representing the estimated quantity of offsite sediment reduction credits needed each year for the City of Wichita’s Offsite program, was achieved. The results consistently demonstrated that spatially informed selection strategies, particularly the clustered subbasins, reduced both the number of fields and the total treated area required to meet the target, indicating improved implementation efficiency. This analysis considered exclusively cropland areas, with uniform corn cultivation applied across the entire land use to ensure consistent production conditions. Field prioritization was assessed at watershed, HUC-12 subbasin, and individual field scales, with emphasis on higher sediment-producing areas based on landscape and spatial characteristics. This study offers a practical, model-based approach for identifying priority agricultural fields and guiding effective Off-site BMP implementation to protect water resources in the LTAK.