Geophysical Data Integration for Uncertainty Quantification in 3D Geologic Models of the Krafla Geothermal System

摘要

Geologic models relate the spatial distribution of different lithologies and the variability of rock properties. While such models constitute essential tools for high-value decisions in geothermal exploration and development, the underlying datasets used to construct the models are sparse, and quantification of the uncertainty of these models is usually not attempted. In recent years, new approaches have been developed that consider uncertainties in geological structure using stochastic simulation methods, and which can ensure that each simulated model is geologically reasonable and forms a valid representation of the subsurface in light of additional data (e.g. geophysical measurements). In this study, we use the GeoModeller software package to combine implicit geologic modelling methods with a statistical inference framework. The objective is to develop probabilistic geologic models of the Krafla geothermal area that are constrained by gravimetric data. Integration of geophysical data into the model construction process allows the uncertainty of the lithologic and density structure of the subsurface to be quantified. Results show that the most uncertain areas correspond to the locations where there is a low availability of direct primary observational data from drillholes.