Assessing controls on selenium fate and transport in watersheds using the SWAT model

Pratikshya Neupane; Ryan T Bailey; Saman Tavakoli-Kivi

doi:10.1016/j.scitotenv.2020.140318

Assessing controls on selenium fate and transport in watersheds using the SWAT model

Sci Total Environ. 2020 Oct 10:738:140318. doi: 10.1016/j.scitotenv.2020.140318. Epub 2020 Jun 18.

Authors

Pratikshya Neupane¹, Ryan T Bailey², Saman Tavakoli-Kivi³

Affiliations

¹ Department of Civil and Environmental Engineering, Colorado State University, 1372 Campus Delivery, Fort Collins, CO 80523-1372, United States. Electronic address: pratikshya.neupane@colostate.edu.
² Department of Civil and Environmental Engineering, Colorado State University, 1372 Campus Delivery, Fort Collins, CO 80523-1372, United States. Electronic address: rtbailey@colostate.edu.
³ Department of Civil and Environmental Engineering, Colorado State University, 1372 Campus Delivery, Fort Collins, CO 80523-1372, United States. Electronic address: saman@colostate.edu.

PMID: 32806359
DOI: 10.1016/j.scitotenv.2020.140318

Abstract

Selenium toxicity in groundwater and surface water in many regions globally has resulted in toxicity in the food chain thus harming terrestrial and aquatic animals. Here we assess the environmental controls on selenium fate and transport in soils, groundwater, and rivers at the watershed scale using the SWAT watershed simulation model. We modify the SWAT modeling code to simulate selenium fate and transport in surface runoff, lateral flow, soil percolation, groundwater flow, streamflow, and irrigation water within a watershed subject to oxidation-reduction reactions. This modified SWAT model is applied to an irrigated watershed with selenium problems in the Arkansas River Valley, Colorado. The model is calibrated and tested using observed groundwater and surface water selenium data in different locations of the study area, whereupon global sensitivity analysis is performed to assess controls on selenium fate and transport in soil water, groundwater and streams. The impacts of reaction rates of selenate, selenite, oxygen and nitrate along with 14 hydrologic parameters are assessed. For the baseline model, selenium loading to streams during growing season is mostly via groundwater discharge (365 kg/ha; 85% of total), followed by surface runoff (14%), and soil lateral flow (1%), which are about 4-fold the loading that occurs during non-growing season. Results from the sensitivity analysis indicate that hydrologic factors (timing of recharge, soil hydraulic conductivity, and groundwater storage thresholds) are the principal controls on selenium content in soils, groundwater, and stream water, followed by the first-order reaction rates of selenate and selenite, amount of selenium-bearing shale in the aquifer, and sulfur to selenium ratio in the shale material. These results suggest that selenium mitigation procedures should focus on water management practices rather than influencing microbial redox reactions. Overall, the SWAT-Se model introduced here can be used to assess Se contamination and mitigation practices in watersheds worldwide.

Keywords: Groundwater; SWAT; Selenium; Sensitivity analysis; Watershed modeling.