Climate change has been a global issue since the 19th century. The increase in rainfall variability, which covers the increase in the earth’s total precipitation, will definitely lead to frequent and more severe flood disasters. As the damage increases year after year with floods as the most chronic and costly disaster among these hazards, Korea has to improve its technological responses and countermeasures to better visualize the hazards brought about by such disasters. Gunsan City ranked number eight in the country’s most susceptible region to floods. From 2004 to 2013, Korea has experienced a total of 174 flood disasters which were estimated to cost USD 7.32 billion. But reports showed that the total expenditure of the government amounted to 1.4 times the estimated losses and damages and the private companies have spent twice the said estimated amount. To summarize, the post-disaster loss and damage reports showed underestimated values. This study aims to develop a semi-parametric geographically weighted regression which can implement a flood damage estimation model of Gunsan City. The model building process include parameters like flood depth, flood duration, inundated area, family income and land price. The datasets are composed of both untransformed and transformed data (using Box-Cox Method). Both Ordinary Least Squares (OLS) Regression and Geographically Weighted Regression (GWR) were evaluated in this study, but the search for best fit resulted to the use of GWR.

From 2004 to 2013, Korea has experienced a total of 174 flood disasters and has a total estimated cost of USD 7.32 billion. However, reports showed that the total expenditure of the government amounted to 1.4 times the estimated losses and damages and the private companies have spent twice the said estimated amount. To summarize, the post-disaster loss and damage reports showed underestimated values. In this regard, the National Emergency Management Agency (NEMA), the government institution designated to assess and analyze the damages and losses as well as evaluate the disaster risks of the said areas in accordance to their disaster risk management plans, are now developing a new estimating method for damages and losses. This study aims to develop flood damage functions that will estimate the flood damages of Gunsan City based on the building type: residential, commercial and agricultural facilities, by utilizing the Ordinary Least Squares Regression and later on, the Geographically Weighted Regression. The model building process includes flood depth, flood duration, inundated area, family income and land price as the parameter variables. Due to normality issue, the datasets were transformed through Box-Cox Method. Both Ordinary Least Squares (OLS) Regression and Geographically Weighted Regression (GWR) were evaluated in this study, but the search for ‘best fit’ resulted to the use of GWR.

Drought and flooding are just some of the ways the planet responds to climate change. As years pass, with increasing population and level of urbanization, these events become more frequent and severe. In South Korea, a disaster risk assessment system was developed to mitigate the flood risks in small streams. But since there is worldwide severity of water-related disasters, investigation and development of such methods of mitigation should not stop. In this study, the watershed was subdivided to reduce the time needed for a real-time flood simulation. The subdivision of the watershed was performed in three different cases with respect to the hydraulic characteristics of the watershed. The selection of an appropriate grid size to produce optimum flood simulation and efficient simulation time was also performed. This research aims to provide guidelines for watershed subdivision for real-time risk analysis systems as well as contribute to the existing knowledge of disaster risk assessment system in South Korea