Evolutionary Learning in Spatial Agent-Based Models for Physical Climate Risk Assessment

Climate risk assessment requires modelling complex interactions between spatially heterogeneous hazards and adaptive economic systems. We present a novel geospatial agent-based model that integrates climate hazard data with evolutionary learning for economic agents. Our framework combines geospatial agent-based modelling with asset-level damage functions, featuring an illustrative three-sector economy (commodity, manufacturing, retail) with adaptive learning behaviours that allow firms to evolve strategies for budget allocation, pricing, wages, and risk adaptation through fitness-based selection and mutation. We demonstrate the framework using riverine flood projections under RCP8.5 until 2100, comparing four scenarios: baseline and hazard conditions with and without evolutionary learning. Our results show that increasingly frequent and intense acute hazards lower firm production levels, liquidity, and capital, while increasing the prices of goods and unemployment. The framework reveals systemic risks where even agents not directly exposed to floods face impacts through supply chain disruptions. Importantly, evolutionary adaptation enables firms to maintain higher production, capital, liquidity, wages and employment levels while keeping prices lower compared to non-learning counterparts. This open-source framework provides financial institutions and companies with tools to quantify both direct and cascading climate risks while evaluating cost-effective adaptation strategies.