Hakan Basagaoglu, Chetan Sharma, Debaditya Chakraborty, Icen yoosedoost, F. Paul Bertetti

Study region
The Edwards Aquifer Region is located in south-central Texas, United States.

Study focus
The paper focuses on the development and implementation of a data-inspired heuristic drought identification scheme to (i) quantify the intensity, duration, and frequency of precipitation deficit- and high temperature-driven meteorological droughts (PMet- and TMet-droughts), and (ii) link their propagation to groundwater droughts (GW-droughts) using baseline hydroclimatic measures and prevailing drought conditions derived from historical climate data and regional mitigation strategies.

New hydrological insights for the region
Based on the intensity, duration, and timing of PMet- and TMet-droughts in the semi-arid karstic region, we identified three distinct GW-droughts, including persistence-driven, preconditions-driven, and intensity-driven droughts. The analysis revealed that successive heavy precipitation events are needed to end GW-droughts in the region. The scheme also identified TMet-droughts with the longest dry spells, TMet- and PMet-droughts with the highest intensity, and GW-drought with the second-highest intensity on record all occurred over the past 15 years. These findings provide evidence for a warming climate, intensified meteorological droughts, and increasing stress on the aquifer. Among the artificial intelligence models used, Extremely Randomized Trees (ERT) predicted time series of intensity & duration of GWdroughts from hydroclimatic features with high accuracy. Moreover, the ERT classifier revealed that the duration of PMet droughts and the intensity of TMet droughts are the topmost decisive features in predicting GW-drought intensity in the region.