Interferometric synthetic aperture radar (InSAR) is a powerful geodetic technique capable of measuring deformation at fine resolution. Radar data's two-dimensional structure along with the pair-wise nature of interferometry allow InSAR to capture both the spatial and temporal extent of deformation. This dissertation focuses on improving spatio-temporal modeling techniques for InSAR data to better describe the observed subsidence at several geothermal fields in the Western U.S. The first chapter focuses on refining the spatial analysis of deformation observed at Brady Hot Springs, Nevada by introducing a parameterization which directly relates displacement at the Earth's surface to subsurface reservoir volume change. Geostatistical inversion in a Bayesian framework identifies thermal contraction of the rock matrix as the dominant driving mechanism of the observed subsidence. The second chapter extends this modeling to multiple interferometric pairs to explore the deformation's temporal nature. Joint time-series analysis of volume change rates estimated from InSAR and Global Positioning System (GPS) data determines the dependence of deformation on well operations. The third chapter measures transient deformation at Coso geothermal field, California using InSAR and GPS data acquired between 2004 and 2016 to quantify relationships between deformation, pumping, and seismicity. Changes in subsidence rate, reservoir contraction, and estimated sink depth after 2010 found from spatial and temporal deformation modeling are attributed to changes in injection protocol corresponding to sustainability efforts implemented in late 2009. The last chapter quantifies the spatio-temporal dependence of the subsiding region at San Emidio geothermal field, Nevada by modeling InSAR data from 1992 to 2010.