Investigating best management practices to reduce selenium and nitrate contamination in a regional scale irrigated agricultural groundwater system: Lower Arkansas River Valley, southeastern Colorado

Abstract

The Lower Arkansas River Valley (LARV) is well known for its rich agricultural production, with 109,000 ha of irrigated area. Due to agricultural production extending for more than 100 years, the LARV now faces challenges of soil salinity, water logging from shallow groundwater tables, and a high concentration of selenium (Se, both within the alluvial aquifer system and within the Arkansas River and its tributaries). Se originates primarily from bedrock and outcropped marine shale, released due to chemical oxidation in the presence of dissolved oxygen and nitrate. Se is a dynamic element that is biologically essential for plants, animals and humans. However, it is known that Se can be harmful at elevated concentrations. Therefore, elevated concentration levels in the surface water and groundwater in the LARV are considered problematic, and methods must be found to decrease groundwater concentrations and Se loadings from the aquifer to the Arkansas River. This thesis assesses plausible methods that will decrease Selenium (Se) contamination in groundwater and surface water in the LARV. Best management practices (BMPs) to reduce selenium and nitrate mass loadings to the River Arkansas in a 55,200 ha area downstream of John Martin reservoir in the LARV were explored and analyzed using 18 scenarios. The UZF-MODFLOW and UZF-RT3D numerical models, calibrated against extensive sets of field data in the region, were used to simulate groundwater flow and the physical and chemical processes governing the fate and transport of Se and N species. Specific BMPs include reduction in the seasonal application of N fertilizer; decrease in concentration of selenate (CSeO4) and nitrate (CNO3) in canal water, representing treatment of water before application as irrigation water; reduction in irrigation application volumes; and combinations of these practices, along with fallowing of irrigated land. These practices are applied for a long term period (40 years) to observe the effects of each BMP on groundwater CSeO4 and CNO3 and on mass loadings from the aquifer to the Arkansas River. The BMPs are applied at varying levels: less aggressive (20%) to very aggressive (40%) of each practice. Results indicate that the highest aggressive combined scenario 40% reduction in N fertilizer reduction, 40% reduction in canal concentration, and 35% reduction in irrigation volume, with 25% irrigated land fallowing result in the highest decrease of mass loadings of SeO4 into the Arkansas River with 22.7%, followed by the less aggressive and highest aggressive combined scenarios of N fertilizer reduction, canal concentration reduction, and irrigation volume reduction with land fallowing showing decrease of mass loadings from 15% to 21%. For individual scenarios: the irrigation volume reduction scenario (13.1% to 13.4%) is followed by the canal concentration reduction scenario (3% to 6%); whereas the N fertilizer reduction scenario shows a minimum percent reduction (1.5% to 2.7%) as compared to the Baseline (“do-nothing” scenario). Similarly for NO3, results show that the highest aggressive combined scenario 40% reduction in N fertilizer reduction, 40% reduction in canal concentration, and 35% reduction in irrigation volume, with 25% irrigated land fallowing result in the highest decrease of mass loadings of NO3 to the Arkansas River with 34.7% followed by the less aggressive and very aggressive combined scenarios of N fertilizer reduction, canal concentration reduction, and irrigation volume reduction with land fallowing showing reduction of mass loadings from 15.5% to 30%. The results of individual BMPs is as follows: 35% irrigation volume reduction scenario (14.9%) is followed by 40% N fertilizer reduction scenario (14.5%); 20% irrigation volume reduction scenario (12%); 20% N fertilizer reduction scenario (8.3%); whereas 20% and 40% canal concentration reduction scenarios show minimum percent reduction (0.6% to 1.1%). The results are compared with results from a similar study recently performed in the Upstream Study Region of the LARV to observe the differences in BMP practices and their reduction of Se contamination in the study areas.