Towards the development of economic damage functions for weather and climate extremes

An important aspect of modeling the coupled human–Earth system is how weather and climate affects the economy. In particular, extreme weather and climate events can cause significant economic damages. Hence, there is a need to better understand the relationship between economic damages and extreme weather and climate events. Here, we analyze economic damage data using regression and probabilistic methods. We provide evidence for upward trends for inflation adjusted damages caused by extreme weather and climate events, while damages normalized by wealth have downward trends. We also find that relative to its wealth, Europe is the most affected continent followed by North America and Asia. We also identify covariates which affect economic damages. Multivariate regression models using either socio-economic or climate covariates explain only small amounts of the damage variation while the probabilistic Generalized Pareto Distribution (GPD) model with covariates fits the damage data well. This suggests that economic damages due to weather and climate extreme events should be modeled using probabilistic methods. Using non-stationary GPD models, we find that mainly population is a significant covariate while also Gross Domestic Product (GDP), global mean surface temperature and modes of climate variability are significant covariates too.