Computing societal dynamics in response to climate change

Abstract:

The growing demand for models of coupled human-natural systems requires behaviorally-rich and dynamic representations of socio-economic actors. For example, behavioral climate adaptation and mitigation calls for integrating diverse actors, behavioral change, learning, social networks and institutions in a computerized code to be compatible with natural systems models. The adequate representation of these agents of change in computational models is crucial. In this presentation I will discuss how agent-based models (ABMs) help advancing modeling of coupled human-natural systems while embracing the insights from social and behavioral sciences regarding the theoretical and empirical microfoundations of agents decisions.

Firstly, I briefly outline the state-of-the-art in computational models with heterogeneous behaviorally-rich adaptive agents and its potential for coupling with natural systems models. Secondly, I provide an example of a spatial ABM developed to explore how relocation and housing prices in flood-prone urban areas evolve as behavioral heuristics, preferences or risk perceptions of households change. This empirical ABM captures socio-economic and behavioral heterogeneity of boundedly-rational actors, their interactions, learning and adaptive behavioral change, and traces the out-of-equilibrium dynamics in urban housing markets prone to increasing climate change risks. We find that when households agents switch behavioral heuristics after facing a flood, as indicated by our survey data, they start avoiding flood-prone locations. Cumulatively it drives a decline of housing prices in hazard-prone areas, which after repetitive floods becomes permanent. Due to market sorting low-income households are either outpriced out of safe areas or get trapped in flood-prone urban areas, leading to climate gentrification. I conclude with open challenges, including modeling of institutions, endogenous societal transformations, and scaling up of behaviorally-rich ABMs.