Quantifying the Efficiency of Surveillance Strategies for Subsurface CO2 Storage Applications

In this paper we address the challenge of efficient and reliable monitoring of subsurface CO2 storage operations for conformance verification purposes. In order to acquire a storage license, the operator has to demonstrate convincingly to the regulator that the operation of the site will conform to safety regulations, and that CO2 will remain securely stored in the subsurface. The level of certainty to which site conformance can be assessed will depend on the uncertainty in the geological structure and the specifics of the monitoring strategy. The assessment of alternative monitoring plans should take into account the locations and accuracy of monitoring data, as well as the geological uncertainty, which will translate into uncertainty in behavior of the CO2 plume as predicted by models. The design of cost-efficient and effective monitoring systems therefore aims at producing, with minimal costs, sufficient evidence of the behavior of the storage sites. Due to the variety of available sensing technologies, it can be a challenge for operators to choose which measurements to gather. In this work, we present a practical workflow to assist operators (and regulators) in quantifying the efficiency of potential surveillance strategies, enabling the comparison of alternatives, or as a first step to optimize the design of monitoring systems. The workflow uses ensembles of model realizations and history matching techniques to statistically: (a) determine the degree of consistency between expected and actual behavior, and (b) assess the performance of monitoring strategies for various plausible behavior scenarios. This procedure allows us to derive a measure of ‘efficiency’ for monitoring strategies in terms of the contribution to correct assessment of conformance. We apply the proposed workflow to a hypothetical CO2 storage site containing a potential leak in the overburden with uncertain properties such as onset time, conductivity, and location. We compare different metrics to define the efficiency of the surveillance strategy. Thereafter, to illustrate how this approach could be used for optimizing monitoring configurations, we study the sensitivity of these metrics to different measurement noise levels. The methodology is relevant for the problem of designing monitoring systems to verify conformance of CO2 storage sites. Due to its modular nature, the proposed workflow is readily expandable to different measurement types and provides practitioners with a framework to quantify the expected contribution of a surveillance strategy to the assessment of storage site performance, something which is currently not available.