Seasonal Changes in Soil Moisture Content in Northern Chile and Southern California Inferred from SAR data

Abstract:

InSAR-based studies of the seismic cycle have focused primarily on the interferometric phase observations, which place constraints on the amount of uplift or subsidence of the ground surface. Recently, coseismic InSAR coherence has also been used to rapidly identify urban damage, surface ruptures, cracking, and soil liquefaction. Here we demonstrate that time-variable correlation and amplitude data contain additional information about surficial processes and material properties that may affect ground deformation and seismic hazard. In the use of correlation for hazard response, distinguishing the coseismic signal from other changes in surface properties associated with variations in soil moisture content, vegetation and snow cover, and wind is critical. Building SAR-based catalogues of ground properties will therefore improve the reliability of rapid response and aid in the designing of future SAR missions to better map surface ruptures, off-fault deformation, and coseismic damage.

In this project, we characterize the seasonal variations in the soil moisture content in the Northern Chilean Coastal Cordillera and Southern California. The extreme climate of the Atacama Desert characterized by hyperaridity and coastal fog during the non-summer months creates an ideal landscape for exploring surface properties. We produce interferograms using L-band ALOS data (λ = 23.6 cm) that span 46 days to three years and have perpendicular baselines less than 1500 m. We observe a strong seasonal dependence on correlation that extends to the maximum elevation of the fog penetration. Interferograms with only austral summer acquisitions are more correlated than interferograms with one or both acquisitions in the autumn, winter or spring, even when the summer interferograms span multiple years. We propose that the seasonal dependence is due to small changes in the radar path length caused by variable soil moisture content in the very shallow subsurface. We further consider local variations in correlation surrounding aeolian dunes, quebradas or ravines, cities, and salars. We extend our work to include the Owens Valley and Death Valley in California.