Can we monitor shallow groundwater using ambient seismic noise?

Nowadays, the majority of detailed information about groundwater is acquired by wells that provide limited insight in time and especially space. Therefore, it would be interesting to monitor groundwater by continuously measuring seismic velocity changes in the subsurface. The shallow soil is affected by environmental influences like temperature, rainfall or drought, which in turn changes the seismic velocity in the subsurface.

In this study, we use three-component seismometers, which are placed next to an in-situ measurement station of soil conditions (moisture and temperature at different depths) and a meteorological station in the city of Hamburg, Germany. We investigate the sensitivity of high-frequency (> 1 Hz) seismic waves with an anthropogenic origin to ground moisture changes in the uppermost layers of soil. To monitor velocity changes, Passive Image Interferometry is applied. Using the three-component data, we are able to retrieve Rayleigh and Love waves. Relative velocity changes are retrieved using the stretching method. A comparison of seasonal seismic velocity changes and environmental changes shows a positive correlation between velocity and temperature, as well as a negative correlation between velocity and groundwater content. Freezing events are exceptions, as they cause relative velocity increases twice as high as seasonal changes.

The aim of this work is to eliminate temperature effects to work towards inferring water content directly from seismic velocity changes. To eliminate the contribution of temperature, its relation to seismic velocity changes and water content is quantified using regression. Since the relative velocity change is influenced by both temperature and water content, a time period of stable water content is used to quantify the relation between velocity change and temperature. As a result, the residual relative velocity change reproduces the residual water content.