Ground motion measurement in the Lake Mead area, Nevada, by differential synthetic aperture radar interferometry time series analysis: Probing the lithosphere rheological structure

Doin, Marie-Pierre; Cavalié, O.; Lasserre, C.
Ecole Normale Supérieure, Paris, doin@geologie.ens.fr

SAR interferometry has proven to be a reliable method for detecting small displacements due to ground subsidence. In this study, we measure ground motion around the lake Mead (Nevada, USA) using InSAR. This artificial lake has been filled with water in 1935. An earlier study, based on leveling measurements, has shown that the load associated with lake impoundement induced a subsidence of 17 centimeters (Kaufmann and Amelung, 2000). This relaxation process has been argued as analogous to the postglacial rebound, but at a smaller spatial scale and with a much lower viscous relaxation scale. To quantify the deformation and thus constrain the crust and mantle rheological parameters in the lake area, we analyze multiple small baseline interferograms (~500) based on 62 ERS images acquired between 1992 and 2009 and on 40 ENVISAT images acquired between 2003 and 2010. All interferograms show a very good coherence due to arid conditions. On most interferograms, strong atmospheric patterns mask the deformation signal. The atmospheric delays are partly due to the variation of water vapor vertical stratification between two satellite passes. This tropospheric delay is correlated with elevation on the interferometric scene. It can thus be estimated and corrected from the interferometric phase. This correction is validated using dry and wet delays computed from global atmospheric models (ECMWF). Corrected interferograms are then inverted to solve for the time series of ground motion in the lake Mead area. The linear inversion treats each pixel independently from its neighbours and uses the data redundancy to reduce errors such as local decorrelations or a few local unwrapping errors. Additional constraints such as temporal smoothing allow to reduce the turbulent atmospheric delays. We obtain spatio-temporal series of the deformation in the lake Mead area from 1992 to 2010. The deformation is non linear in time and spreads over a large spatial scale. A few areas connected with the lake show ground displacements interpreted as poro-elastic deformation in a connected aquifer in sedimentary layers. We also observe a broad subsidence pattern of up to 16 mm between 1995 and 1998 due to a 10 meters water level increase, followed by a stronger uplift due to the drop of the water level after 2000. We model the deformation, taking into account the water and sediment loading history of the lake since 1935. A simple elastic model with parameters constrained by seismic wave velocities does not explain the amplitude of the observed motion. The two-layer visco-elastic model proposed by Kaufmann and Amelung (2000), with a mantle viscosity of 10**18 Pa s, adjusts well the data amplitude and its spatio-temporal shape. We will discuss the models that could explain the broad deformation pattern taking into account errors derived from atmospheric perturbations.