Using hydro-chemical watershed modeling to quantify the impact of water management practices and future climate on selenium fate and transport in a semi-arid watershed

Abstract

Selenium is a naturally occurring element in soils and is often referred to as a "double-edged sword element" due to the narrow margin between its essential nutritional levels and toxic concentrations, for humans and animals. This characteristic becomes particularly problematic in semi-arid irrigated watersheds underlain by seleniferous soils and bedrock, in which intensive irrigation can accelerate the mobilization of soluble selenium species, namely selenate (SeO4) and selenite (SeO3), leading to water quality degradation and posing risks to aquatic organisms and wildlife sensitive to elevated selenium levels. Effectively managing selenium within the context of water management is a complex task, further compounded under future climate change scenarios, which introduce uncertainties in water availability, land use patterns, irrigation demands, and associated selenium transport processes and flux pathways. Gaining a clearer understanding of selenium sources and its mobilization pathways, particularly under changing climatic conditions, can support policymakers in addressing current selenium concerns and planning for future risks. Recognizing the importance of informed selenium management, the overarching objective of this dissertation is to provide actionable insights for policymakers, achieved through three specific aims: i) Developing a modified SWAT Selenium hydro-chemical watershed model to assess the environmental factors that control selenium fate and transport within a watershed, to identify key system parameters influencing selenium contamination in irrigated stream - aquifer systems; ii) Quantifying hydrological fluxes in a large, semi-arid, irrigated region in which selenium contamination occurs, and how water management affects these fluxes; and iii) Evaluating selenium storage and transport in soils, aquifers, and streams under the influence of future climate scenarios and water management strategies in an intensively irrigated, selenium-contaminated watershed. To gain a deeper understanding of selenium fate and transport in semi-arid irrigated areas, it is essential to develop a robust watershed model that can simulate selenium behavior and identify the key environmental and hydrologic drivers influencing its fate and transport. To address this need, a selenium module was developed and integrated into the SWAT hydrologic model to simulate reactive selenium transport through major hydrologic pathways (surface runoff, infiltration, soil lateral flow, tile drainage outflow, recharge, groundwater discharge to streams, streamflow) within a watershed. The model was applied to a selenium contaminated 1,118 km2 irrigated stream aquifer system in the Lower Arkansas River Valley in southeastern Colorado. A two-stage global sensitivity analysis (GSA) was conducted: the first stage assessed the influence of selenium concentration and reaction parameters alone, while the second incorporated selected sensitive SWAT hydrologic parameters. Findings revealed that while selenium specific factors such as the sulfure:selenium (S:Se) ratio in Cretaceous shale , oxidation of shale by dissolved oxygen (O2), and first-order chemical reduction of SeO4 are influential, selenium concentrations in stream, soil and groundwater are predominantly governed by hydrologic conditions, underscoring the importance of water management strategies in mitigating selenium mobility in such systems. To assess selenium conditions at a larger scale and incorporate the influence of a reservoir and tributaries that drain desert landscapes, a SWAT model covering 15,900 km2 was developed for a section of the Arkansas River Watershed between Pueblo Reservoir and Catlin Dam for the period 1990-2014. In recognition that selenium fate and transport are likely controlled by hydrological processes, the first step is to perform model calibration and testing for hydrological conditions. The model evaluated key components of the water balance, including surface runoff, evapotranspiration, soil moisture, lateral flow, and groundwater discharge to streams. Using the model, the impact of various management scenarios on hydrologic fluxes was analyzed: transitioning from flood to sprinkler irrigation; implementing canal lining at different extents. Additionally, wet and dry years were identified and hydrologic flux patterns within these two year types were examined under these scenarios. Transitioning from flood to sprinkler irrigation resulted in a dramatic reduction in surface runoff - over 90% - dropping from 72 mm/yr under the baseline scenario to just 1 mm/yr, without compromising crop water availability in the irrigated areas. Canal sealing also significantly reduced groundwater return flow, with a 15% decrease observed under 20% sealing (from 108 mm/yr to 93 mm/yr) and a 57% reduction under 80% sealing (down to 46 mm/yr). The results indicated that management interventions can significantly influence both surface and subsurface water dynamics, highlighting their potential for improving water resource management in semi-arid, irrigation-intensive regions. In the second step, the model was amended with the selenium module and run to quantify the impact of future climate and water management strategies on selenium mass fluxes and in-stream selenium concentrations. Morris sensitivity analysis suggested that Oxygen oxidation rates and shale to Selenium to Sulphur ratio particularly in the Pierre and Smoky hill shale types have dominant roles in controlling selenium transport in the study area. Based on the sensitivity analysis results, the model for the historical period (1990-2014) is calibrated and tested against in-stream concentrations in Fountain Creek and along the Arkansas River. Mass balance analysis of the Arkansas River indicated substantial inputs at Pueblo (6.9 kg/day) and from Fountain Creek (11 kg/day), while other tributaries contributed minimally due to low flows and limited presence of selenium-rich Pierre Shale. Selenium mass in groundwater return flows were highest in intensively irrigated subbasins, notably Subbasin 20 (6 kg/day) and Subbasin 26 (5.1 kg/day). Canal diversions removed substantial selenium loads, including 6.4 kg/day from Reach 20 and 3.2 kg/day from Catlin Canal. Results highlight irrigation practices and return flows as primary drivers of selenium transport, underscoring the need for targeted management in high-loading areas. To further explore selenium fluxes, additional modifications were made to the SWAT-Se module and it was applied to the calibrated large scale SWAT model under two climate change scenarios (RCP 4.5 and RCP 8.5). Simulations were conducted for the baseline condition as well as for management scenarios involving the conversion of flood irrigation to sprinkler systems and varying degrees of canal sealing under both climate projections. Selenium fluxes were then quantified for each scenario to assess the impact of climate and management interventions. Management interventions yielded mixed outcomes. Canal sealing consistently reduced selenium concentrations across sites, by up to 60 -75%, bringing Catlin Dam down to 14.3 µg/L and Avondale and Nepesta to 6.8 µg/L and 6.3 µg/L, respectively. Sprinkler irrigation, however, reduced selenium at Catlin Dam under RCP 8.5 (to 21.2 µg/L) but increased concentrations in some locations, including Nepesta (16.9 µg/L under RCP 4.5). These findings suggest that while canal sealing is a robust mitigation strategy, sprinkler irrigation may enhance subsurface Se leaching depending on local hydrogeology. The study underscores the dominant role of groundwater in Se transport and the critical need for adaptive irrigation strategies under changing climate conditions. The SWAT-Se model offers a robust framework for evaluating future selenium risks and guiding management decisions in selenium-impacted watersheds.