Hydrofacies simulation based on transition probability geostatistics using electrical resistivity tomography and borehole data

Hydrofacies distribution and heterogeneity influence groundwater flow and solute transport. The characterization of hydrofacies distribution is challenging because of the scarcity of observation data, and because of the various types of uncertainties in whatever observation data exist and in the geostatistical methods used. This paper presents a method to integrate geophysical data and borehole data to improve the characterization of hydrofacies distribution and to reduce uncertainty in the simulations of porous media, in particular, unconsolidated sedimentary clay/sand environments in unsaturated states. A sandbox experimental investigation was conducted by the combined use of drilling and electrical resistivity tomography (ERT). Borehole data reveal the vertical variation in hydrofacies in the model, and ERT data can well represent the continuity of hydrofacies in the profiles. A histogram probability matching method was used to assimilate the two kinds of data. The ERT profile was converted into a hydrofacies section based on a cutoff value of resistivity, and then a series of highly credible virtual boreholes were extracted from the hydrofacies profiles. Finally, the transition probability geostatistics (TPG) method, based on the Markov chain, was employed to simulate the hydrofacies distribution. The results show that the ERT data, as a kind of high-density soft data, can provide more lithological information for TPG simulation. The presented method with TPG can construct a relatively high-accuracy hydrofacies model by the combined use of ERT geophysical data and borehole data.

Automated summary

This paper presents a method that combines geophysical data and borehole data to improve the characterization of hydrofacies distribution and reduce uncertainty in porous media simulations. Through a sandbox experimental investigation, the method assimilates borehole data and electrical resistivity tomography (ERT) data using a histogram probability matching method. The hydrofacies distribution is then simulated using the transition probability geostatistics (TPG) method based on the Markov chain. The results show that ERT data can provide more lithological information for TPG simulation and the combined use of ERT data and borehole data can construct a relatively high-accuracy hydrofacies model.