Targeting salinity solutions: a national geospatial framework with regional salinity characterization for agricultural desalination planning

Abstract

Water scarcity and rising salinity in agricultural areas present a major threat to the long-term sustainability of irrigation and the quality of water in river systems across the United States. This study aimed to identify suitable locations for implementing desalination technologies to enhance the availability of irrigation water and improve water quality in river systems throughout the contiguous United States, while addressing the insufficient characterization of salinity in these systems. GIS tools, machine learning, a river mass balance approach, and a hydrological model were utilized to generate spatial suitability models and subsequently to characterize salinity in a specific region in greater detail, both temporally and spatially. For the national assessment, national datasets on watershed boundaries, surface water salinity, groundwater salinity, irrigated lands, cropland cover, solar photovoltaic potential, and wind capacity factor were utilized to characterize watersheds and evaluate their suitability for implementing agricultural desalination powered by renewable energy sources, based on thresholds established in the literature. Several hotspot regions deemed suitable for agricultural desalination were identified, primarily in the western regions of the United States, including parts of the Rocky Mountain states (Colorado, Idaho, Montana, Utah, and Wyoming), most of Texas and Oklahoma in the Southwest, and throughout the Plains region in Kansas, Nebraska, Minnesota, North Dakota, and South Dakota. Additionally, sections of California and Florida, along with the lower Mississippi states. This emphasizes the extensive issues of freshwater salinization across the country and the potential of agricultural desalination as a source of freshwater. For one of the suitable regions (Colorado's Lower Arkansas River Valley), a detailed salt mass balance was applied along the river between Pueblo Reservoir in Colorado and Coolidge in Kansas (total distance =317.5 km) to quantify daily groundwater salt ion loadings for the 2000-2020 period, and determine the relative influence of upstream flow, canal diversions, tributary inflow, and groundwater discharge on salt mass within the Arkansas River. Random Forest modeling was used to estimate daily salt ion loads at gaging sites. A SWAT-MODFLOW model was then used to determine the influence of on-farm desalination on groundwater salt loadings to the Arkansas River. Two scenarios were selected: the first with no desalination applied, and the second with half desalination, meaning 50% of the salt ions in the irrigation water are removed. Although agricultural desalination showed no significant impact on the salinity of the river system, this research provides a framework for applying agricultural desalination and establishes a decision-support foundation for managing agricultural water salinity challenges.