Integration of multi parameters geophysical models using PCA : application to geothermal exploration

Geophysical exploration of natural resources is challenging because of complex and/or narrow geological structures to image. Geophysical models should provide an image at a scale large enough to understand the complex geology but with the adequate resolution to resolve features like faults. One solution to overcome this difficulty is to integrate large multiphysics datasets to provide complementary insight of the geology. New approaches involve joint inversion of all datasets in a common process where models are coupled together. Geometrical or quantitative interpretation of the joint models image several physical properties shaping the same pattern of the target resources. In reality, models resulting from joint inversion are still challenging to interprete. Most of the joint inversion techniques are based on parameters relationship or geometrical constraint which imply common interfaces between models. This assumption may be wrong since geophysical methods have different sensitivity to the same geological object.

Geophysical integration cover a wide range of approach from the visual interpretation of model presented side by side to sophistical statistical analyses such as automatic clustering. We present here a geophysical models integration based on principal component analysis (PCA). PCA allow to gain insight on a multi-variable system with high level of interaction. PCA aims to reorganize the system by finding a new set of variables distributed along new orthogonal axis and keeping most of the variance from the data. Thus geophysical interaction are highlighted along components that can be interpreted in terms of patterns. We applied this integration method to gravity, ambient noise tomography and resistivity models obtained from joint inversion in the framework of unconventional geothermal exploration in Massif Central, France. PCA of the log-resistivity, the density contrast and the Vs velocity model has 3 independent components. The first one (PC1) representing 69% of the total variance of the system is highly influenced by the parameter coupling enforced in the joint inversion process. PC1 allows to point to geophysical structures that may be related to the geothermal system. The second component (PC2) represents 22% of the total variance and is strongly correlated to the resistivity distribution The correlation with the surface geology suggests that it may be a fault marker. The third component (PC1: 9% of the total variance) is still above the nul hypothesis and seems to describe the 3D geometry of the geological units. This statistical approach may help the geophysical interpretation into a possible geothermal conceptual model