Income-Specific Vulnerability in Event-Based Models for the Impact Assessment of Tropical Cyclones

Tropical cyclones are among the meteorological and climatological extreme events with the largest economic impact worldwide, although they only affect coastal areas in certain geographical latitudes. At the same time, anthropogenic climate change affects the characteristics of tropical cyclones: they move more slowly, achieve greater intensities and reach higher latitudes more frequently.

In view of these complex changes, there is great demand for models assessing the economic impact of tropical cyclones that are not based on simple macroscopic indicators. In a so-called event-based approach, a spatially and temporally resolved representation of a single tropical cyclone is blended with spatially resolved data on the distribution of economic capital. Thus, even complex changes in storm characteristics can be included in the impact assessment.

The vulnerability of affected regions is expressed through impact functions that relate loss of economic capital to a storm’s local wind speed. These impact functions are calibrated against historic damage databases and then applied to hypothetical storm scenarios for risk analysis or climate change forecasting.

This research aims at the calibration of impact functions that explicitly depend on the per capita income of the affected region. For this purpose, a United States damage database identifying impacts at the county level is introduced, allowing the investigation of the statistical relationship between vulnerability and per capita income of affected counties.

A quantitative formulation of the income dependence of vulnerability will not only improve the overall predictive quality of event-based models, but also enhance the understanding of the impact of tropical cyclones on social inequality.

We find that vulnerability is highest in the poorest affected regions, indicating the extraordinary relevance of tropical cyclones on low-income coastal communities. Furthermore, we can show that vulnerability mostly declines with per capita income.