Groundwater sustainability in Central and Southern Arizona is threatened by prolonged droughts, rising temperatures, reduced surface water supplies, and groundwater overdraft. Recent studies indicate accelerating declines in regional groundwater storage. While the anthropogenic drivers of these declines are relatively well understood, the role of natural hydroclimatic variability in groundwater storage changes has received less attention. In this study, we utilize NASA’s Gravity Recovery and Climate Experiment (GRACE) and its successor, GRACE Follow-On (GRACE-FO), together with NASA’s Western Land Data Assimilation System (WLDAS) to evaluate the hydroclimatic controls on Groundwater Storage Anomalies (GWSA) variability in Central and Southern Arizona. We quantify long-term standardized trends, subbasin-scale correlations, and dominant modes of natural variability using Principal Component Analysis (PCA) from 2004 to 2021. Principal components are then used to group groundwater subbasins using k-means clustering, an unsupervised machine learning algorithm, and results are compared with in situ measurements and local management strategies. Our results show substantial spatial heterogeneity across the region. This heterogeneity is characterized by recharge-responsive northern and central subbasins, which are influenced primarily by precipitation and subsurface runoff, and loss-dominated southern subbasins, associated with weaker natural recharge and stronger atmospheric demand. PCA shows that natural hydroclimatic variability is statistically aligned with approximately 16% of the inter-subbasin spatial variance in GRACE/FO groundwater storage trends. Within this natural component, total evapotranspiration (~ 29%), precipitation (~ 23%), and subsurface runoff (~ 20%) represent the largest contributors to the explained variance. The remaining spatial variance may reflect anthropogenic influences, geologic heterogeneity, and residual observational or modeling uncertainties. Our diagnostic framework identifies groundwater subbasin clusters driven by shared hydroclimatic modes. It has the potential to serve as a transferable tool for recharge feasibility analysis, groundwater sustainability assessments, and future local groundwater planning in the Lower Colorado River Basin.
Mohajer et al. (Tue,) studied this question.