Optimization of Carbon Dioxide Removal Pathways: Quantifying the Effects of Foresight, Technology Inertia, and Demand Shocks

Abstract

Carbon dioxide removal (CDR) is essential across all pathways that limit global warming to 1.5°C, yet a gap remains between required CDR levels and current policy support. Model-based scenarios are widely used to inform CDR deployment strategies, but commonly rely on assumptions of perfect foresight and unconstrained technology deployment. These assumptions fail to capture real-world constraints such as political short-sightedness and technological inertia, which can lead to path dependencies and technology lock-ins. A cost-minimization optimization model for CDR deployment in a Canadian context is developed, incorporating six technologies across three demand scenarios and varying levels of foresight and technological inertia. Limited foresight is represented through three decision-making approaches and compared against a perfect conditions baseline: a rolling horizon approach that allows previous decisions to be revisited, a myopic approach where deployments become irreversible, and a novel blind approach where commitments are made without any future knowledge. Technological inertia is incorporated through literature-based growth and decline rate constraints. Additionally, disruptions are modeled as sudden shifts in CDR demand occurring at a given year. The results show that reduced foresight and constrained deployment flexibility jointly produce the largest cost penalties and deployment deviations, ranging from a 5% cost premium under adaptive planning to five times the perfect conditions cost under blind commitment at high inertia levels. In undisrupted scenarios, the two constraints can partially compensate for one another, but once disruptions are introduced, both low inertia and long foresight are required to limit deviations. The two disruptions examined are symmetric in deployment terms but asymmetric in cost: underdeployment penalties exceed stranded asset costs, which are not penalized, establishing that failing to deploy enough is considerably more costly than deploying too much. The portfolio mix further shapes outcomes, as delaying investments in long-lead technologies constrains the system’s ability to meet demand later. These findings suggest that adaptive review cycles and strategic portfolio development are essential to CDR governance, particularly in high inertia systems operating under high demand uncertainty, and should be considered in Canadian CDR strategy.