How storage uncertainty will drive CCS infrastructure

AbstractThis study focuses on the basin and site scales to identify physical constraints for CO2 injection and source-sink infrastructure. We analyze the sequestration of CO2 emissions associated with conventional and unconventional fossil fuel development in the Uinta-Piceance Basins, Colorado and Utah, USA. This area is underlain by abundant saline formations with excellent potential geologic sequestration. Land access, as a proxy for pore space access, is an important consideration in this area of rugged terrain, protected natural areas, and Indian and private lands. We integrate a model o f geologic CO2 sequestration (CO2-PENS) and a model of infrastructure optimization (SimCCS) to evaluate the design of CCS infrastructure under uncertainty.This research focuses on the effect of uncertainty in properties of sequestration reservoirs on CCS infrastructure, including the dedicated CO2 pipeline network and which sources should capture CO2 or where new CO2 emitting facilities should be located. CO2 source emissions and capture costs can be estimated with reasonable certainty based on current separation technologies. In contrast, the actual capacity and injectivity of saline aquifers (sinks) may vary over several orders of magnitude compared with estimated values, due to geologic heterogeneities that affect porosity, permeability, thickness, and extent of the saline aquifers.We report on modeling using CO2-PENS, an injectivity/capacity and risk assessment simulator package for geologic sequestration, and SimCCS, a geospatial decision optimization model for comprehensively designing CCS infrastructure. CO2-PENS uses statistically distributed input parameter values to characterize CO2 migration through the sequestration reservoir, caprock, and overlying freshwater aquifers, as well as potential leakage pathways like wellbores and faults. Representative parameter ranges were developed for the Cretaceous Castlegate and Jurassic Entrada sandstones, widespread permeable saline formations in the UintaPiceance basins. A GIS mask based on land use, land ownership, slope, and hydrology was developed to define land surface access for developing sequestration sites. Several sites of varying area and formation depth were defined for both the Castlegate and Entrada formations. Using the Monte Carlo modeling capability of CO2-PENS, multiple realizations were run for each site in order to develop probability density functions (PDFs) of reservoir capacity and cost for sequestration of various rates of CO2 delivery. These PDFs were provided to SimCCS for use in the calculation of optimal pipeline networks among CO2 sources and sinks. SimCCS plans CCS infrastructure with respect to objective functions that include information on financial budget, regional CO2 capture target, or a price on carbon (i.e., what scale of infrastructure is economically feasible). Because the model simultaneously examines source-network-sink components, it is ideal for analyzing the impact of sink uncertainty on overall CCS infrastructure. Model outputs include CCS costs, spatial network routing, and the scale at which deploying CCS infrastructure makes sense.Coupling CO2-PENS and SimCCS, both state-of-the-art models, allows us for the first time to examine how reservoir uncertainty propagates through, or even drives, the entire CCS infrastructure system. In our study this effect is evidenced by preferential selection of sinks with spatially proximal alternatives, robust pipeline networks that can respond to changes in CO2 flow, and sources chosen that make economical sense (though not necessarily financially optimal) and integrate well into a dynamic CCS system