Towards Management-Scale, Satellite-Based Estimates of Groundwater Depletion

October 23, 2025

Ryan Smith

Hosted by Eric Maloney

Abstract

Groundwater supplies drinking water for billions of people worldwide, and nearly half of all irrigation water. Many aquifers of the world are being depleted, but traditional methods for estimating storage loss can vary by up to an order of magnitude due to uncertainty in aquifer properties and spatio-temporal sparsity of in-situ measurements such as groundwater levels and withdrawals. Satellite observations such as Interferometric Synthetic Aperture Radar (InSAR) and OpenET provide high-resolution estimates of some elements of the water balance that could be leveraged for improved estimates of groundwater storage. However, gaps in the water balance challenge efforts to utilize these datasets for improved groundwater monitoring. In this talk, I will describe two recent projects of my research group that explore new approaches to harness these data. The first makes use of OpenET imagery and physics-informed machine learning to estimate groundwater withdrawals at 2 km resolution over the western US. The second employs a combination of in-situ groundwater measurements and InSAR data to disaggregate groundwater storage change estimates to confined and unconfined aquifers in several depleted aquifers of the western US, including California, Utah and Colorado. These approaches provide a path towards satellite-driven or satellite-supported methods for improved groundwater management.