The full project name is Monitoring Geological CO₂ Storage: Quantitative CO₂ Prediction with Uncertainty from Physical Modeling and Multiple Time-Lapse Data Types. The project is funded by the Norwegian Research Council and the industrial partners Statoil and ExxonMobil. We collaborate with the University of Bergen and Stanford University to develop improved methods for quantitative prediction of the distributions of CO₂ in subsurface storage sites. Statistical models are used as the key tool for integrating repeated measurements of multiple data types within physical models describing the CO₂ saturation changes.

Main goal

Improved quantitative prediction of the spatial distributions of CO₂ in subsurface storage.

Contribution to  CO₂-handling:

• A rock physics library provides  the basis  for a systematic approach to CO₂ monitoring
• Joint inversion of multiple data sources  constrained by   rock physics  models  assures consistent use of data
• Quantitative interpretation of inversion provides the spatial distribution of CO₂  in subsurface storage
• Reliable prediction of CO₂  ensures public safety and support for continued CO₂  storage
• Quantitative assessment of  CO₂  saturation  provides the limits for storage rates and capacity  essential  for the  economy  of  a  storage site in emission trading

Large amounts of data

An essential part of the project is the integration of data from the Sleipner capture and storage gas processing facility.  This facility is the world's first fully operational offshore gas field with CO₂ injection. The storage is in a geological layer 1,000 meters below the sea floor. Multiple seismic and gravimetric monitoring surveys have been performed in the area starting prior to the first injection in 1996.

These repeated geophysical measurements are used to monitor the CO₂ plume, as they are indirect observations of the CO₂ saturation changes. There are several challenges in the interpretation of the geophysical measurements in terms of changes in CO₂ saturation. Firstly, the data amount is massive and thus there is a need for efficient algorithms to process the data. Secondly, there is uncertainty both in measured data, and related to how the rock properties will change when the original brine filling is substituted with CO₂. These uncertainties must be accounted for by the methodology.