Hydrological Modelling of Flow and Nitrate Load Using SWAT+

Abstract

Climate change is increasingly threatening watersheds worldwide (Harris et al. 2024; Mankar et al. 2020; Tran and Lakshmi 2024a; Tran et al. 2024; Yin et al. 2024), leading to prolonged droughts (Sharma et al. 2022; Tapas, Kumar, et al. 2022; Do et al. 2024) and more frequent floods (Prabha and Tapas 2020) that endanger ecosystem health (Mishra et al. 2023). Hydrological modelling has become an essential tool for assessing the severity of these impacts (Tapas et al. 2024c; Murumkar et al. 2025; Marshall et al. 2025a, 2025b; Tran and Lakshmi 2024b). Over the past decade, technological advancements have significantly enhanced hydrological modelling capabilities (Brookfield et al. 2023; Jehanzaib et al. 2022). The Soil and Water Assessment Tool (SWAT) has evolved into SWAT+, offering a more user-friendly environment and improved process representations (SWAT+ 2020; Tran et al. 2023). While recent studies have utilised SWAT+, there remains considerable potential to further explore its effectiveness in nutrient modelling.

In this study, we employed the SWAT+ model for the Tar-Pamlico River Basin, developed by Tapas (2024a, 2024b), to provide a detailed depiction of flow and nitrate distributions across the basin (Tapas et al. 2024c). The model was constructed using an array of datasets, including the Digital Elevation Model (DEM), observed flow data from the United States Geological Survey (USGS), land cover data from the National Land Cover Database (NLCD), soil data from the Soil Survey Geographic Database (SSURGO), weather data from the Global Precipitation Measurement Integrated Multi-satellite Retrievals for GPM (GPM IMERG), and water quality data from the North Carolina Department of Environmental Quality (NCDEQ) (Tapas, Etheridge, et al. 2022; Tapas et al. 2023, Tapas 2024b, Tapas, Etheridge, et al. 2025, Tapas, Howard, et al. 2025).

Tapas et al. (2024) optimised the SWAT+ model for monthly flow and monthly nitrate load at Washington, North Carolina, using a two-year warm-up period (January 2001 to December 2003), a calibration period from January 2003 to December 2011, and a validation period from January 2012 to December 2019. The model was also soft-calibrated for annual average hydrological response unit (HRU) scale flow, nitrate loss, yield, and denitrification. Lastly, the model was cross-validated for monthly flow at two additional upstream locations in the watershed (Greenville and Tarboro, NC) (Video 1).

To effectively communicate the model results, dynamic animations illustrate the spatiotemporal dynamics of flow and nitrate load throughout the Tar-Pamlico River Basin. These visualisations capture the variability of streamflow and nitrate transport over time, highlighting seasonal and interannual trends. Additionally, a comparative analysis of observed and simulated nitrate loads is presented using graphical representations and statistical indices (NSE and PBIAS) to assess model performance. The integration of remote sensing datasets and SWAT+ simulations provides a comprehensive depiction of hydrological and water quality patterns in the basin.