Investigating the early 21st century droughts in Central California using seismic interferometry

Quantitatively monitoring groundwater storage in space and time can be difficult due to the lack of - or only discrete - direct measurements. This difficulty increases if the temporal and/or spatial scales increase. Groundwater monitoring at a basin scale over several decades is still technically challenging if one wants to achieve a high spatial and temporal resolution. In this presentation, we show that a seismic-based approach might help to fill the gaps to accurately monitor in space and time groundwater storage at regional scale. To do so, we use about 15 years of vertical continuous records from permanent seismic stations in Central California to compute yearly cross-correlation functions. We use these correlation functions to monitor the yearly seismic velocity variations within the Central Valley. Instead of performing Coda Wave Interferometry, we measure the velocity variations directly on the ballistic Rayleigh waves in the frequency/calendar-time domain using a newly developed approach based on the wavelet transform. This new method allows us to better spatially locate the velocity changes thanks to the deterministic path traveled by the wave. Moreover, by taking advantage of the different depth sensitivities of the fundamental and higher modes of the Rayleigh waves, the depth localization is also improved. This approach can resolve relative variations as small as 0.1%. Our results show a clear increase of velocities during droughts, compensated by the decrease of velocities during wetter periods, with sometimes reversed behavior depending on the analyzed frequency. Our methodology brings new possibilities for purely passive 4D seismic tomography and could be applied to various type of geological targets.