Repeatability indicators in time lapse seismology and their application to the Sleipner CO₂ storage project

Brian H. Russell

I compare three repeatability measures using the time lapse data from Sleipner CO₂ storage project in offshore Norway. The three repeatability are NRMS, predictability, and cross-correlation coefficient. I first review the work of Kragh and Christie (2002) who used NRMS and predictability and created a random noise model to explain their relationship. Using the Sleipner dataset, I show an excellent fit to their theory. I then review the work of Coléou et al. (2013), who used NRMS and cross-correlation measures and introduced two new attributes: quality Indicator (Q) and anomaly indicator (A). After discussing the relationship between predictability and cross-correlation I apply the Q and A attributes to the Sleipner dataset, showing how well the CO₂ plume can be identified.