High-quality and yet cost-effective monitoring of the ground- and surface water above Geological CO2 storage is needed to lower long-term monitoring cost and mitigate the public concern for effects on ground- and surface water. This can be obtained by combining a) modelled leakage from the geological CO2 storage with b) release rates of specific parameters from local geological formations in contact with CO2 in reactive transport models for the ground-and surface waters, to enable the optimization of the monitoring network and the monitoring frequencies in monitoring wells and possibly volume integrating water supply wells and surface water. Further cost reductions and long-term monitoring consistency can be obtained by applying Machine Learning techniques.
A numerical model for CO2 and brine leakage via faults and legacy wells is constructed. Kinetic models for parameters release from site sediments are derived from lab tests and upscaled through field tests. A detailed hydrostratigraphic model of the ground and surface water is set up for modelling reactive transport of the release parameters. Leakage from the CO2 storage is introduced into the ground and surface water model at the modelled locations and with the calculated fluxes. The model results are then used to optimize the positioning of the monitoring network for cost and detection efficiency. Existing data, new monitoring data and model data with anomalies are used to train an ever-alert Machine Learning model for consistently flagging anomalies far into the future.
A methodology for cost-efficient long-term monitoring of ground- and surface water that could become a regulatory standard replacing the current imprecise requirements. Standard or not, a methodology that can be applied by consultancies enabling them to offer consistent, high quality, low cost, long-term monitoring operations.
