Stochastic Fusion of Information: An application to Monitoring Water Movement in STVZ Infiltration Field Site

A stochastic information fusion technology that can assimilate different types of information (e.g., hydrologic, geologic, geophysical, geochemical, geostatistical information, etc.) to provide the best unbiased estimate and quantify its uncertainty is a promising tool to monitor and characterize the vadose zone. In this study, a stochastic fusion technology based on a combination of geostatistics, three-dimensional variably saturated flow model, and three-dimensional electric current flow model was developed. It was applied to the STVZ (Sandia-Tech Vadose Zone) field site at New Mexico to monitor the water movement in vadose zone during infiltration experiments. This fusion technology estimated water contents directly using data sets from surface and borehole dipole-dipole ERT surveys during different periods of the infiltration experiments. Water content measurements from neutron probe data of 13 boreholes were used to condition our estimates of water content. Spatial variability of the resistivity-moisture relation was also taken into account in our approach. To overcome difficulties associated with the large number of data sets, the fusion process was carried out sequentially by propagating conditional moments of variables. Results of our study illustrate the effectiveness of this fusion technology in real world problems.