The Action titled "Development of a Method for the Exploration of the Joint Geology-Geophysics Model-Space" (GeoMos) aimed to develop modeling techniques that integrate geophysical data, such as gravity field anomalies, with geological measurements, such as the orientation of the contacts between rock units and their locations. This development was driven by the need to capture a snapshot of the range of models allowed by the data and the plausible features that the data can accurately resolve. Indeed, the necessity to quantitatively and reproducibly integrate geological and geophysical modeling has been recognized by the geoscientific community as an efficient means to reduce uncertainty in exploration, monitoring, and risk assessment related to natural resources and subsurface management.
In addition, given the current challenges in modeling increasingly complex areas, it has become necessary not only to produce a single 'best guess' model but a series of models representing a range of plausible scenarios to provide insights into the 'unknown unknowns', to reduce risks and plan future exploration efforts. To address this, this project introduced methods for the robust integration of geological and geophysical modeling through the development of a modeling algorithm capable of jointly considering geological and geophysical measurements. It relies on new techniques to generate collections of models that fit the data and represent different geological scenarios, using modeling algorithms:
1. For both geological and geophysical data, where geology and geophysics are linked by converting information from a geophysical model into inputs for geological modeling and vice versa.
2. By calculating series of intermediate models between differing scenarios that all fit the geophysical data with a similar level of fidelity.
3. By calculating a large number of models that fit the data within uncertainty using an automated process for the addition or removal of possibly unknown rock units to better explore the full range of possibilities.
These algorithms were initially tested with idealized data and successfully applied to real-world measurements in the Western Pyrenees (spanning France and Spain) and the Boulia region in Northeastern Australia (Queensland). This allowed for the generation of plausible models that deviate from commonly accepted interpretations, showcasing the new methods and contributing to discussions about the studied areas.
The results are to be made open source upon acceptance of the journal publication presenting the final version of the code developped during the project.
