UC San Diego

Timeseries of ground surface displacement derived from Interferometric Synthetic Aperture Radar (InSAR) provide critical information about aquifer characteristics and groundwater sustainability. We observe subsidence and uplift in Lehi, Utah corresponding closely to industrial groundwater extraction at a semiconductor manufacturing plant (Chapter 1). We compare the surface deformation with public timeseries of extraction volumes to provide a sustainability analysis demonstrating that the aquifer has been pumped at a generally sustainable level but also that some permanent compaction has occurred. To investigate the effect of noises on InSAR timeseries, we compare InSAR with Global Navigation Satellite Systems (GNSS) timeseries in the Mojave desert, using a formal approach which allowing us to place statistical error bounds on the InSAR-based temporal resolution of the post-seismic motion of the 2019 Ridgecrest earthquake (Chapter 2). We find spatial referencing schemes introducing non-uniform spatial patterns of uncertainty and seasonality in the bias and amplitude of turbulent and stratified tropospheric noise. Finally, we introduce a theoretical framework of uncertainty quantification for InSAR timeseries based on the Kronecker product of spatial and temporal covariance matrices (Chapter 3). We translate common InSAR processing techniques, such as the SBAS inversion, timeseries smoothing, common-scene stacking and spatial referencing into this framework. This uncertainty quantification is both powerful and practical for common use for InSAR timeseries processing.