ABSTRACT Understanding and quantifying hydrologic alteration is essential for effective watershed management and ecological conservation. This study developed a modeling framework to predict reference hydrology and assess and model hydrologic alteration across Texas using publicly available geospatial datasets. A two‐stage random forest regression approach was used to model reference flow conditions at gaged locations based on natural predictors. Alteration was calculated as the deviation of observed values from these modeled reference conditions. A second set of models predicted hydrologic alteration using anthropogenic predictors. Reference hydrology models performed well (mean R 2 > 0.9), with stream classification identified as a key driver. Hydrologic alteration models exhibited more variable performance ( R 2 range: 0.45–0.89), with land use variables consistently ranking highest in predictive importance. Predicted alteration patterns exhibited both spatial and temporal variability, with a general trend toward increasing surplus flow conditions over time, especially in eastern Texas. This modeling framework enables large‐scale prediction of hydrologic alteration and provides insight into its primary drivers, offering a valuable tool for anticipating the impacts of future development on flow regimes.
Lueders et al. (Thu,) studied this question.