Using the Geographically Weighted Regression to Modify the Residential Flood Damage Function

Flood damage functions are necessary to ensure comprehensive risk management. This study attempts to establish a residential flood damage function and explores residents living in the Keelung basin, where flood disasters occur frequently in Taiwan. Ordinary least squares (OLS) method is used to construct a flood damage function. Analytical results indicate that flood depth, is the significant variable, but the residual is non-stationary with spatial. The Geographically Weighted Regression (GWR) model is applied to modify the traditional regression model, which cannot capture spatial variations, and to solve the spatial non-stationary. Analytical results demonstrate significant spatial variation in the local parameter estimates for the variable flood depth and intercept. Therefore, a dummy variable, Zone, is added to the OLS model. The R-square value is found to increase from 0.15 to 0.24, and the residual is spatially stationary. In conclusion, the residential flood damage is determined by flood depth and zone, and the GWR model not only captures the spatial variations of the affecting factors, but also helps to discover the explanatory variable to modify the traditional regression model.