Evaluating trade-offs among cotton yield, groundwater extraction, and future projections for sustainable water management in the Texas High Plains

The Ogallala Aquifer—the largest freshwater aquifer in North America—has experienced significant depletion over time, primarily due to intensive agricultural irrigation in the Texas High Plains (THP). With minimal natural recharge, the aquifer shows little to no signs of recovery. This study presents a novel data-driven framework that integrates remote sensing and in-situ observations to quantify groundwater extraction (GWE) and evaluate trade-offs with crop water productivity (WPc) in irrigated cotton fields across the THP from 2008 to 2030. To analyze spatial and temporal trade-offs, we developed cotton area and lint yield models using Tasseled Cap Transformation (TCT) indices, Landsat-based RS metrics, and climatic data from 2008 to 2023. The CatBoost model achieved strong performance in cotton extent mapping (overall accuracy = 0.97; F1-score = 0.89) and yielded the best results in lint yield estimation when combining RS and climate data (R² = 0.55; RMSE ≈ 305 kg/ha). WPc was computed as the ratio of lint yield to total water input (precipitation plus groundwater), where groundwater use was estimated as evapotranspiration minus precipitation. Trade-offs were categorized into four classes—overuse, inefficient, efficient, and low input—based on the median values of WPc and GWE. Results from 2008 to 2023 indicate that much of central and northern THP fall under overuse and inefficiency categories, posing a risk to groundwater sustainability. Projections through 2030 estimate an increase in GWE from 2892 to 3439 billion liters. This integrated approach demonstrates strong potential for broader application in other regions experiencing similar groundwater management challenges.