Small-scale seismic monitoring of varying water levels in granular media
Water in the unsaturated vadose zone has an important function for many aspects of life. It is particularly important for plants’ growth and serves as a buffer for movement of pollutants from the land surface to the aquifers. The amount of water present in the vadose zone also determines the partitioning of rainfall at the land surface into infiltration and runoff (greater vadose zone water implies less infiltration and more runoff).
The characterization of the water in the vadose zone can be done using many methods. In-situ monitoring provides accurate data but is expensive and cannot characterize the variability over large spatial scales. Satellite remote sensing using microwave sensors is a well-established technique that has been used to quantify soil moisture in the shallow surface layer (0-5cm) over regional, continental and global scales using a combination of aircraft and satellite platforms. If we desire information in depth, we are however still restricted to either in-situ sensors such as time domain reflectometry (TDR) probes and/or ground penetrating radar (GPR) which use electromagnetic signals converted to water content.
Unfortunately, methods like GPR are not efficient in electrically conductive media such as fine textured soils (e.g. loess, clays), as the signal gets attenuated and restricts the depth of investigation. To overcome these limitations, a geophysical method using the properties of seismic waves – pressure (P) and shear (S) waves – in the subsurface, as modulated by the presence of water, has emerged as a good option. The ratio between the P and S wave velocities for a given geological formation has been shown to reflect the amount of water in the subsurface as measured by in-situ observations. However, water content in the subsurface varies on a continuous manner while the inversion of seismic data yields water content at discrete layers, bringing the need to improve inversion models.
In order to answer those questions, a small-scale seismic experiment was set up in the laboratory, expanding on previous efforts to use laser-doppler vibrometer and unconsolidated granular physical models. Seismograms were recorded for different water levels in the granular medium, and compared to those generated with an elastic three-dimensional finite difference simulation of the experimental setup to ensure a correct identification of the seismic events. The velocity models derived from the dry experiment match very well with previous experimental and theoretical results. The presence of water in the granular medium has a clear impact on the recorded seismogram, both on the amplitudes and the velocities of the seismic waves. However, the wet data could not be inverted due to the lack of a theoretical model for such partially saturated granular medium. To overcome these drawbacks, travel time and phase velocity differences between the wet and dry models were studied. Only travel time differences were found to be significant, showing inflection points at the water and capillary fringe levels.
This study, published in the July 2016 issue of Vadose Zone Journal, offers attractive perspectives for studying soil water content variations on the field, following on recent studies showing that water content changes in the vadose zone could result in significant P and S wave velocity variations. This setup can be employed in the field in addition to TDR. The resulting characterization of water content variability will be useful in a wide range of hydrological and geophysical applications.
Read the full article in Vadose Zone Journal. Free preview available Sept 7 - Sept 14.