Monitored Groundwater Storage Change: an Overlooked, Essential Driver of Groundwater Management

Groundwater overdraft is a global problem demonstrating a general inability of modern civilization to manage groundwater demand. Climate change only makes this problem more acute and challenging. A key reason for chronic overdraft is the relative invisibility of groundwater and pumpers’ effects on it. Here I assert (1) the need for real-time monitoring of an overlooked metric – change in groundwater storage (Δs), (2) its transformative potential for groundwater management and (3) a simple yet new approach to monitor Δs in real time. The hydrology community has failed to learn a key lesson from the impressive successes of measuring sub-continental scale Δs with GRACE satellite technology. The lesson: monitoring Δs (rather than simply changes in groundwater levels, Δh) increases dramatically peoples’ awareness of groundwater overdraft and their motivation to better manage groundwater. Unfortunately, the sub-continental resolution (~400 km) of GRACE makes it ineffective as a tool for monitoring and managing groundwater at a functional basin scale, which is typically on the order of 10’s of km and rarely as large as ~400 km. Although monitoring of Δs in shallow, unconfined aquifers is relatively simple, it is much more challenging in most sedimentary basins that contain most of the world’s major aquifer systems, and where myriad, interbedded aquifer and aquitard layers create depth-variable degrees of semi-confinement that complicate the relationships between Δs and Δh. Here I demonstrate a simple approach in which a calibrated groundwater flow model is used to translate data on Δh into real-time estimates of Δs, despite massive (104) spatial variations in the effective storage coefficients (Δs/Δh). This meta-modeling approach means that it is feasible today to monitor Δs with conventional hydrogeologic data and tools, highlighting missed opportunities for more effective, science-based groundwater management. I will also articulate the need for much greater research emphasis on new methods for monitoring Δs, including combined use of machine learning methods leveraging diverse datasets along with conventional data and models.