A GIS-based multivariate approach to identify flood damage affecting factors

Abstract. This paper investigates causal factors leading to pluvial flood damages, beside rainfall amount and intensity, in two Swedish cities. Observed flood damage data from a Swedish insurance database, collected under 13 years, and a set of spatial data, describing topography, demography, land cover and building type were analyzed through principal component analysis (PCA). The topographic wetness index (TWI) is the only investigated variable that indicates a significant relationship with to the number and amount of insurance damage. The Pearson correlation coefficient is 0.68 for the number of insurance damages and 0.63 for amount of insurance damages. With a linear regression model TWI explained 41 % of the variance of the number of insurance flood damages and 34 % of variance of amount of insurance flood damage.

Future studies on this topic should consider implementing TWI as a potential measure in urban flood risk analyses.