Study On The Grid Of Housing Data Based On The Classification Of Earthquake Damage

Building structure reflects not only the economic development level of bothurban and rural areas but also their vulnerability of disaster. And it has importantreference significance on the regional disaster prevention policy and urban and ruraloverall macro-scale development planning. In recent years earthquake disasters hasbroken out more frequently and become a great threat to the development ofsociety, people’s life and property. It is shown that in many cases earthquakecasualties and direct economic losses are mainly caused by the collapse of thebuildings. Furthermore, the housing data is an important part of the earthquakeemergency database. Thus obtaining accurate housing data is significant for disasterprevention before earthquake, disaster relief and reconstruction after earthquake.In many cases, the housing data has played an important role, but at the sametime it also exhibits its limitations: a) The housing data is too old and difficult to beupdated, and the indicators of housing data in the statistical yearbooks do not agreewith those used by the department for seismic work. b) Data scale is too large andmost of the housing data was saved at the level of administrative districts which isconsidered as a homogeneous unit, ignoring the distribution of buildings. Thus, thatis one of the reasons that cause statistical bias to earthquake damage assessment.In recent years data gridding has become a focused research, and is the bestway to solve these problems mentioned above. Nowadays it has formed a relativelymature theory and methods, especially on the research of population data gridding,but sadly until now no relevant papers and reports about the housing data griddinghave come up.In view of this, with the support of basic geographic data and GIS software, thepopulation data grid model is established according to different region type, urban orrural, and different landforms, using the town population as a control factor andeventually the population grid data is obtained. Then this work used the km grid sampling method to obtain enough building samples and establish the housing datagrid model. Combined with the population grid data, the housing grid data of the fiveprovinces in the north-south seismic zone is obtained.Firstly, the research of the population data gridding has been carried out and itgives a solution to the problem of model scale transformation, urban and ruraldivision and model factor selection. Taking advantage of the method of estimatinggrid population based on residential areas, sampling on a km grid scale, the km gridpopulation density weight model is constructed on the basis of statistical analysis.There are6model factors which can be divided into4categories, closely related tothe cultivated land, which is closely related to population, land for urban and ruralresidents, roads, and residential areas. Considering town as the smallest unit, thetype of landform is divided from province to province, based on elevation, terrainruggedness and geological structure. The urban and rural division is done manuallyaccording to the RS data in different areas and then an urban model and a ruralmodel are established respectively.Secondly, the research of the housing data gridding has also be done on thebasis of population grid data which technically ought to solve the problem of theaverage housing area and the proportion of different housing structures. In this study,by using the method of km grid field survey, enough samples are obtained, on thebasis of which two housing gridding models are established, one based onpopulation density and the other based on area division. Later the housing grid dataof the five provinces, Yunnan, Sichuan, Qinghai, Gansu and Ningxia are obtainedbased on the two models.Finally, this study takes both the Yiliang earthquake and Lushan earthquake forexamples. Taking the advantage of GIS spatial analysis function, the seismic intensitydistribution data and the grid data gained above are overlaid to calculate the affectedpopulation and housing data, and then the casualties and economic loss areevaluated in order to check the science and reliability of the grid data. The resultshows that compared with the actual survey data, the error of the estimated direct economic loss of Yiliang earthquake in Yunnan province is6%. And the error is69.9%and63.8%when calculating the deaths in the Lushan earthquake using two sets ofmodel, but the error of the estimated direct economic loss of the Lushan earthquakeis11.7%and16.8%. The results reveal that the graph of housing density distributionbased on grid unit reasonably reflects the actual inner situation of the unit with ahigh accuracy. Rapid and reliable information on disaster situations, including directeconomic loss, casualties and the number and distribution of affected populationand houses can be obtained by means of rapid estimation based on the grid data.