Abstract
Agricultural water optimization is critical for improving water security for the arid Western USA and other water-scarce regions. Optimization includes assessment of agricultural water diversions within conveyance systems managed by irrigation canal companies (ICC), yet few water balance approaches exist at this level. Satellite data provides insight into historical agricultural water use and the potential for optimization. In this study, historical remote sensing data from OpenET and gridMET were coupled with Geographic Information Systems data (fields/canals distribution and agricultural diversion records) to implement a water balance model for three ICCs in Northern Utah. The monthly water balance model determined the change of water depletion at the root zone (∆Dr) with precipitation, irrigation and capillary rise from groundwater (CR) as inflows; and actual crop evapotranspiration (ET c act ), runoff and deep percolation (DP) as outflows. The water balance residual (WBR) was calculated as the unresolved portion of the water balance and reflect runoff, DP and other unquantified hydrologic exchanges (CR was considered only where supported). The spatial distribution of ∆Dr was captured by a Random Forest model using local observed data (ET c act and soil water content). Across the study area, WBRs varied spatially and temporally (2020–2023) from 0.01 to 2.7 cubic hectometers (12 − 2,210 acre-feet) per month, with the highest values occurring during peak irrigation season (July-August). Seasonally (Apr-Oct), WBR represented 5–29% of the sum of all water balance components across the studied ICCs and should be interpreted as an illustrative screening range. Overall, this study demonstrates the feasibility of approximating an ICC-scale water balance under data-limited conditions, and highlights the need for improved monitoring, data collection, and research.