UC San Diego

Fresh water is a critical building block to natural ecosystems and human societies. With increased anthropogenic consumption and shifting hydroclimate patterns, there is a need to understand the spatio-temporal distribution of water resources. However, quantifying terrestrial water storage – the sum of surface water, groundwater, soil moisture, snow, ice, and canopy water – can be challenging due to the lack of direct field measurements, as well as the high complexity required to simulate the interaction between different hydrologic components through numerical modelling. Recent advancements in geodesy provide an innovative method in directly estimating terrestrial water storage. At weekly to multi-year time-scales, the redistribution of water mass along Earth’s surface induces transient elastic surface deformation that is observableusing precise satellite geodetic instruments. In particular, Global Positioning System (GPS) measurements have the potential to supplement conventional hydrologic methods in constraining water storage dynamics with its high temporal resolution and dense spatial coverage.

This dissertation surrounds to use of satellite observations in quantifying transient deformation driven by hydrologic processes at different spatial scales. In chapter 1, we systematically review the current state of vertical crustal motion in the contiguous United States, dissecting how different geophysical processes contribute to the observed deformation patterns, includingthe signature of surface mass loading. In chapter 2, noise characteristics in transient GPS measurements are examined, and we find considerably amount of long-wavelength noise associated with reference frame motion that may affect the accuracy of water storage estimates. A simple framework is proposed to correct for such effects. Chapter 3 demonstrates the ability of using GPS in estimating water storage in California’s Sierra Nevada, and how it can be used to close the mountain-scale water balance in conjunction with hydrometeorological data. Lastly, chapter 4 explores the use of synthetic aperture radar and multi-spectral imaging in mapping surface property changes due to vegetation growth, and how the two dataset can be incorporated together to build a robust classification model for large-scale crop type mapping.