Research On Risk Exposure Data Extraction Methodology And Impact On Earthquake Insurance Rates

Earthquake disasters bring serious economic burdens to the country and the people,and there is a need to pool the efforts of society and the market to form a mechanism for sustainable social development to transfer and disperse risks reasonably and effectively.Earthquake insurance has emerged and is playing an increasingly important role in risk management and disaster response.At present,earthquake insurance is still in the development stage,and improving the earthquake insurance system has become the top priority of China’s earthquake disaster prevention and mitigation work.Accurate prediction of building losses in earthquakes is the core element of earthquake insurance premiums and rate setting.The earthquake loss of buildings is determined by three parts:earthquake hazard,building vulnerability,and building risk exposure,among which,buildings are the key disaster-bearing bodies that cause economic losses and casualties.Therefore,it is especially important to extract building information and establish a fine-grained building exposure database.However,at this stage,the collection of building information is still based on traditional methods such as field surveys and questionnaires.Although new technical have been involved in recent years,there are still problems such as high cost,low efficiency and slow update in general.How to obtain building information efficiently,accurately and scientifically has become the exploration direction for constructing a building risk exposure database.Therefore,this study aims to explore and improve the earthquake catastrophe insurance mechanism in China,and focuses on the two core issues of building exposure information extraction and insurance rate determination,and the main research work and results are as follows:(1)A building classification system for earthquake insurance is proposed based on the comprehensive consideration of the functional characteristics,engineering characteristics and regional characteristics of buildings.From the perspective of functional characteristics,the functional types of AOI and POI data are reclassified into four categories:residential,public service,commercial,and industrial;from the perspective of engineering characteristics,the number of floors,structure type,construction age,and seismic protection of buildings are classified in accordance with the building codes in China;from the perspective of regional characteristics,the multi-criteria weighting method is used according to population,economy,land use,and fortification intensity.From the perspective of regional characteristics based on factors such as population,economy,land use,and intensity of protection,a comprehensive weighting method with multiple indicators is used to classify rural areas and towns in China into eight categories of areas,respectively,by district and county.(2)A method of extracting urban building information by fusing multi-source spatio-temporal data is proposed.Based on high-definition remote sensing data,the U-net full convolutional network model is used for building contour recognition by deep learning.The spatial superposition method and kernel density estimation method are used to superimpose the AOI data and POI data on the building contours to obtain the building function information.The kriging interpolation method is used to extract the building layer information based on the deep network data.Combined with the multi-temporal surface coverage data,the spatial overlay analysis method was applied to estimate the building construction age range information of the buildings.Based on the deep-net data,the correspondence between the number of building stories,structure type,building function and construction age was established by statistical analysis to obtain the structure type of buildings with unknown structure type.The method of this paper is adopted in the urban area of Tangshan City to realize the extraction of information of building outline,function,number of floors,structure and construction age,etc.The verification results show the effectiveness and practicality of the method.(3)The data sources of rural building information are introduced,and the regional distribution characteristics and building structure types of rural buildings are summarized based on publicly available statistics such as the national population census.The extraction idea of sample estimation overall of rural building information is clarified,the sampling rate determination method is proposed,and two counties in eastern Henan Province are used as examples for example verification.The results show that the sampling results can well represent the overall rural building information under 30%random sampling condition.(4)Taking Tangshan City as an example,the differences in the pure rate determination of earthquake catastrophe insurance at three different spatial scales of city,district and county,and kilometer grid are compared and analyzed,and a suitable spatial scale unit for the pure rate determination in China is proposed.The effects of different levels of building exposure on the pure rates of earthquake catastrophe insurance are discussed.It is revealed that in areas with large differences in building densities and across multiple potential earthquake source zones within the same district and county,rate determination needs to be based on the true distribution of buildings.The selection of spatial scale units and the extraction of risk exposure data must be emphasized in rate setting,and the research on building information extraction technology and building exposure distribution identification technology is strengthened.(5)The impact of parameters such as the insured ratio(compulsory and voluntary degree),the upper limit of compensation,deductible,the proportion of insurance companies,and the construction time of capital pool on the value of compensation loss or rate determination of insurance companies is discussed,and relevant parameters suitable for earthquake catastrophe insurance in Tangshan City are suggested,taking into account the actual situation of Tangshan City,to provide reference for the refined determination of earthquake insurance premiums and rates.