Fine-scale Seismic Disaster Risk Assessment Considering Building Function Types

Earthquakes are one of the most destructive natural disasters in the world.They occur almost instantaneously,causing massive casualties and economic losses.China is located between the Pacific seismic belt and the Eurasian seismic belt,facing a high risk of destructive earthquakes.With the significant improvement in urbanization,population density and the number of buildings continue to increase,amplifying the potential threat of earthquakes to cities.However,limited by the current state of science and technology,humans do not yet possess the ability to accurately predict earthquake occurrences.Therefore,scientific and accurate regional seismic disaster risk assessment is particularly crucial for prioritizing post-earthquake rescue efforts and minimizing earthquake losses.The collection of basic data on vulnerable elements is an essential component of seismic disaster risk assessment.The accelerated urbanization process has led to rapid population aggregation and rapid urban development,posing significant challenges to the timeliness of seismic disaster risk assessment.Traditional field survey methods are no longer suitable for assessing seismic risks in the context of rapid socio-economic development.Therefore,it is of great importance to study and establish a large-scale,high-precision,and rapid method for extracting building information to ensure the timeliness and accuracy of assessment results.Fine-scale seismic disaster risk assessment can provide more accurate basic data and guidance for urban emergency response,enhancing the ability and efficiency of urban emergency response.It also helps urban residents and relevant departments gain a clearer understanding of potential risks and response measures,thereby improving urban disaster prevention and mitigation awareness.In order to scientifically and reasonably consider the functional characteristics of buildings and the distribution of populations in urban areas,this study,based on population dynamics within different types of buildings,established a set of methods applicable to large-scale,high-precision,and rapid extraction of building function types.Then,this study comprehensively considered factors such as seismic hazards and building vulnerability,and developed a fine-scale seismic disaster casualty assessment model suitable for urban areas.The rationality of the assessment results has been validated through a case study in Beijing.The main research work of this paper are as follows:(1)In response to the requirements of fine-scale seismic disaster risk assessment,this study developed a method for identifying building function types using mobile signaling time series data.Considering the different temporal variations of mobile signaling data for different building function types on weekdays and weekends,we categorizes buildings into five types: residential,working,entertainment,visiting,and hospital.Using time-series mobile signaling(MS)data as input and rasterized building function types obtained from Open Street Map combined with manual labeling as output,The MS-RF(Mobile Signaling Random Forest Classification Model)building function type extraction model based on random forest(RF)machine learning method is constructed.Evaluated the accuracy by Holdout test,the overall classification accuracy is 84.89% and the Kappa coefficient is 0.78.Compared to the S2-RF(Sentinel-2 Random Forest Classification Model)building function classification model based on Sentinel-2 remote sensing data,which achieves a classification accuracy of 73.33%,the MS-RF building function extraction method demonstrates higher classification accuracy.(2)Introducing real-time population distribution data and building function classification information,a refined estimation of seismic mortality risk in Beijing area was conducted.Based on the traditional vulnerability-based model for seismic casualty assessment,firstly,we identify building function types of Beijing by the MSRF model.Then,multiple data sources,including building footprint data,building function data,real-time population distribution data,GHSL(The Global Human Settlement Layer)building height data,and building age data,were integrated using the GIS platform.By considering the occupancy rates of different building function types,the time series of mortality risk distribution maps for 48 hours(weekday and weekend)under seismic intensity scenarios of VI,VII,VIII,IX,and X were subsequently generated for Beijing city.Furthermore,using the 8.0 magnitude Sanhe-Pinggu earthquake in 1679 as a designated earthquake event,the seismic mortality assessment results for nighttime at 3 a.m.and daytime at 10 a.m.in Beijing were calculated under the current level of economic and social development.(3)By distinguishing building function types,the accuracy and spatiotemporal refinement of the traditional seismic casualty assessment model have been improved.Taking 8 moments of the Beijing X intensity earthquake scenario as an example,a comparative analysis was conducted between the assessment methods that distinguish building function types and those that do not.The following conclusions were drawn: 1)Compared to the assessment method that distinguishes building function types,the traditional method tends to overestimate the number of fatalities during weekday nights and weekends by approximately 9.85% to 47.38%,while it underestimates the number of fatalities during weekday daytime by about 17.66%to 18.74%.2)From the perspective of mortality rate(excluding the influence of population spatiotemporal distribution),the traditional method tends to overestimate mortality during weekday nights and weekends by approximately 0.36‰ to 2.5‰,compared to the method that distinguishes building function types.Conversely,during weekday daytime,the traditional method underestimates mortality by approximately0.66‰ to 0.71‰.3)Residential buildings have higher seismic mortality risk during nighttime than during daytime,and the mortality risk is slightly higher on weekends compared to weekdays.In contrast,the other four types of buildings show the opposite trend,with higher seismic mortality risk during daytime than during nighttime.Specifically,for working and hospital buildings,the mortality risk is higher on weekdays,while for entertainment and visiting buildings,the mortality risk is higher on weekends.