The Application of Satellite Radar Interferometry to the Study of Land Subsidence over Developed Aquifer Systems

This dissertation investigates the application of interferometric synthetic aperture radar (InSAR) to the measurement and interpretation of surface displacements over developed aquifer systems. Land subsidence over developed groundwater systems has been observed in a wide variety of hydrogeologic settings worldwide. The phenomenon can be explained with elastic and inelastic deformation of water-bearing material at depth in response to declining pore pressures. The lack of observational data has made it difficult in the past to define the extent of the deforming areas, the magnitude of the surface displacements, and the time-history of the deformation process accurately. Consequently, this has also generally precluded the estimation of aquifer system storage or flow parameters, which relate the surface subsidence to the subsurface pore pressure changes. The development of InSAR techniques using satellite radar data now provides the ability to map surface displacements with centimeter to millimeter precision over extensive areas with great spatial detail (10s of meters). I have used InSAR data to derive detailed maps of the time-varying surface displacement fields over the Las Vegas Valley, Nevada and Antelope Valley, California aquifer systems during several years in the 1990s. The achieved measurement accuracy in the two study areas was typically better than 1cm and was limited primarily by the effect of tropospheric signal delays in the radar images. The availability of satellite acquisitions from closely spaced orbits in the existing data catalog constrains the temporal sampling to 35 days or longer for ERS data. For both aquifer systems studied the InSAR observations enabled a detailed and spatially complete characterization of the highly heterogeneous displacement fields. The structure of the observed subsidence in many cases reflected known or previously v unknown subsurface structure such as faults or changes in sediment thickness, emphasizing the value of these displacement maps in delineating subsurface units. A comparison of surface displacements derived from SAR data realizing different viewing geometries over Antelope Valley indicated that surface displacements related to inelastic compaction of compressible units in the aquifer system are primarily vertical, which has been a widely used, albeit hitherto generally untested assumption in basin-scale studies of land subsidence. The observed displacement fields were temporally highly variable, reflecting the effects of both seasonal fluctuations and long-term trends of the stresses in the aquifer systems. By combining independent information on these stress variations with InSAR observations of the surface displacements I was able to estimate spatially variable storage parameters for the heterogeneous aquifer systems. Using a one-dimensional compaction model I interpreted InSAR surface displacement observations in Las Vegas Valley in conjunction with water-level observations to estimate spatially varying aquifer system elastic skeletal storage coefficients between 4.2 · 10−4 and 3.4 · 10−3 in the elastically deforming parts of the aquifer system. In the Antelope Valley aquifer system the drainage of thick low-conductivity units is delayed with respect to the drawdowns in the aquifers, causing continuing land subsidence for many years after the hydraulic head declines have ceased. I estimated inelastic skeletal storage coefficients up to 0.09 and compaction time constants for interbed compaction between 3 and 285 years in a three-dimensional groundwater flow and subsidence (MODFLOW) model. The parameter estimation was constrained both by InSAR subsidence observations and historical benchmark data. I investigated the sensitivity of the parameter inversion to the accuracy and frequency of the subsidence observations and the stress changes in the aquifer system in a set of numerical simulations. The results indicated that InSAR-derived displacement maps are well suited to provide the displacement observations necessary to estimate storage parameters. However the parameter estimation proved to be severely limited by the poor reliability of subsurface pore pressure change estimates in regional aquifer systems. This work is the first use of InSAR technology to investigate the time-dependent vi deformation processes in developed aquifer systems; here I employ InSAR-derived displacement data to estimate spatially variable aquifer system parameters. Where applicable, InSAR provides a powerful tool for characterizing and simulating aquifer systems, which often are an important resource to the local communities.