High-resolution projections of vulnerable populations under the Shared Socioeconomic Pathways

Vulnerable populations are disproportionately affected by climate change and face limited adaptive resources, yet remain largely overlooked in regional planning and decision-making. Future population composition and the spatial distribution of vulnerability shape both mitigation strategies, which reduce the magnitude of climate change, and adaptation strategies, which reduce its impacts, with direct implications for policy interventions and planning. This study provides high-resolution (1 km) population projections for Canada until 2100 across age, sex, ethnicity, income, and education under the Shared Socioeconomic Pathways. Using a novel deep learning–based population dynamics model that simultaneously predicts full population compositions and incorporates macroeconomic, governance, and land surface variables, we examine demographic shifts at national, regional, and local scales. Results reveal substantial variability across scenarios: the proportion of seniors rises most in low-mitigation-challenge scenarios, particularly along lower-emission pathways, while the share of children and youth increases most in scenarios with high mitigation challenges. Ethnic diversity expands when mitigation challenges are low but decreases or stagnates when they are high. Income and education distributions are most sensitive to adaptation challenges, with the strongest polarization of socioeconomic groups into a three-class society and the largest increase in the share of individuals without a diploma occurring when both adaptation and mitigation challenges are high, highlighting the need for high-resolution, disaggregated demographic projections to guide targeted and equitable interventions for vulnerable communities.