Remote Sensing Big Data and Physical Models for Resilience to Geohazards in the Water-Energy Nexus

Abstract

The interaction of natural and anthropogenic factors rapidly transforms Earth's geophysical systems, posing significant risks to critical infrastructure, ecosystems, and communities. My dissertation tackles two central challenges in geosciences: the climate-driven alterations in geosystems within the water-energy nexus and the dynamics of human-induced seismicity. By integrating satellite-based remote sensing, and physics-based models, my work offers new insights into these interconnected phenomena. It introduces innovative forecasting methodologies to mitigate their impacts. Thus, this dissertation investigates the coupled impacts of climate variability and human activity on groundwater systems and seismic hazards using satellite geodesy and geomechanical modeling. By integrating observations from Interferometric Synthetic Aperture Radar (InSAR), the Gravity Recovery and Climate Experiment (GRACE and GRACE-FO), and supporting datasets, I examine how water mass redistribution and anthropogenic stress perturbations drive surface deformation and influence seismic processes across three geophysically and socioeconomically significant regions. In Mexico City, I quantify extensive groundwater loss from 2014 to 2021 by combining GRACE/GRACE-FO gravity observations with high-resolution Sentinel-1 InSAR data. I detect up to ~35 cm/yr of land subsidence within the city and up to 2 cm/yr of uplift in surrounding regions. Elastic and poroelastic models, constrained by local geological conditions, estimate groundwater volume loss ranging from 0.86 to 8.06 km3/yr, revealing the compound effects of long-term aquifer depletion on urban infrastructure and regional water security. At Lake Mead, I assess drought-induced hydrologic deformation between 2020 and 2022 using InSAR and elastic loading models. A maximum uplift of 6 mm/yr is observed near the lake's center, consistent with crustal rebound due to significant surface water loss. Inverse modeling reveals total water storage loss of up to 3.03±0.25 km3/yr, with groundwater contributing approximately 0.94±0.32 km3/yr to this deficit. Time-lag analysis and hydraulic diffusion modeling indicate a dynamic hydraulic connection between the lake and adjacent aquifers, with diffusivities ranging from 3.2 to 86 m2/s, underscoring the spatial heterogeneity of subsurface hydrologic properties. Finally, I revisit the Parkfield Paradox on the San Andreas Fault, where the anticipated M6.0 earthquake occurred 16 years later than predicted. By integrating seismicity records, creep measurements, and long-term fluid injection data from the nearby San Ardo oil field, I show that anthropogenic stress perturbations likely delayed the rupture by approximately seven years. Probabilistic modeling now suggests the next Parkfield earthquake may occur between 2029 (57% probability) and 2050 (97%). Collectively, these case studies demonstrate the power of satellite geodesy to uncover hidden subsurface processes and emphasize the need to account for both hydrologic and anthropogenic influences in managing water resources and forecasting seismic hazards. The findings contribute to a deeper scientific understanding of coupled Earth systems and offer actionable insights for enhancing infrastructure resilience and environmental sustainability in a changing climate.