Dynamic Assessment Of Building Vulnerability Under The Influence Of Debris Flow Hazards In The Wenchuan Earthquake Zone

Debris flows are a serious and prevalent geological hazard that can lead to significant economic losses,fatalities,and social instability.Researchers have focused on three typical debris flow gullies(Qipan,Yangling,and Cutou)in Wenchuan County to evaluate the dynamic vulnerability of buildings in the region under debris flow threats.These areas are located in the seismically active region of southwestern China,where geological structural activity is frequent,and the ecological environment is fragile.In particular,debris flows caused by earthquakes have become the main threat to the safety of buildings and residents in the region,especially after the devastating 2008 Wenchuan earthquake.Furthermore,as the region undergoes rapid post-earthquake recovery and reconstruction,the exposure and bearing capacity of buildings are continuously changing due to the impact of previous disasters and post-disaster recovery efforts.Therefore,it is essential to assess the dynamic vulnerability of buildings under the influence of debris flow in this region.This study integrates literature research,multi-temporal remote sensing interpretation and field investigations.By utilizing vulnerability index evaluation,Boruta algorithm,and random forest model,the vulnerability status of buildings during the predisaster,disaster,and post-disaster recovery periods is monitored and evaluated.This approach enables a quantitative assessment of the dynamic changes in building vulnerability under the influence of seismic area debris flows.The findings of this study provide practical guidance for reducing the risk of debris flow disasters and informing building planning in earthquake-prone areas.The specific research content and conclusions of this study are explained in detail below:(1)By combining drone imaging and field investigation results,data on 19 indicators were obtained in two dimensions: building resistance and debris flow intensity.These indicators include: area,number of floors,floor height,building orientation,geometric form,structure type,wall thickness,foundation depth,building type,existing status,opening ratio,opening height,opening quality,opening location,vegetation coverage,shielding effect,narrow passage effect,fence,and distance from the channel.(2)Using the Boruta algorithm,seven indicators were screened out: mainten ance status,opening quality,opening location,structure type,vegetation coverage,shielding effect,and fence.These selected indicators were used as the original dataset to train a random forest model using the Random Forest Classifier interface.By constructing learning curves and grid searches,the optimal values of four m odel hyperparameters were determined: n＿estimators=181,max＿depth=5,min＿sam ples＿split=2,min＿samples＿leaf=1.The accuracy and reliability of the results were verified by the model’s Accuracy value(0.759),error rate(21.68%)and AUC v alue(0.92).At the same time,according to the feature importance results,the w eights of the above 7 indicators were determined: maintenance status(0.239),op ening quality(0.167),opening position(0.157),structural type(0.144),vegetation coverage(0.138),shielding effect(0.099)and enclosure(0.056).(3)Based on the weighted index and scoring results,the RVI value of each building in the study area was calculated using a linear weighting formula.The study found that the proportion of buildings with high and very high fragility in the three debris flow channels was relatively low,indicating that overall,the buildings are relatively safe.Specifically,indicators related to building resistance,such as maintenance status,opening quality,opening location,and structure type,were found to be the main factors affecting their fragility.In addition,indicators related to debris flow intensity,such as vegetation coverage,shielding effect,and fence,can also provide some protection.(4)Based on the investigation data and Relative Vulnerability Index(RVI)in the pre-disaster,disaster,and recovery stages,the spatiotemporal characteristics and trends of building vulnerability were analyzed.The study showed that the average RVI of buildings in the study area increased after the occurrence of debris flow and then decreased to below pre-disaster levels during the recovery period.This indicates that after experiencing initial disaster damage,the vulnerability of surviving buildings inevitably increases.However,if well recovered after the disaster,their disaster resistance can be effectively improved and even exceed pre-disaster levels.In terms of spatial distribution,there are certain differences in building vulnerability among different regions.Among them,Cuitou Gully has the largest fluctuation in building vulnerability while Yangling Gully has the smallest fluctuation with Qipan Gully in between.(5)By comparing the evaluation indicators at the three stages of pre-disaster,during disaster and post-disaster recovery,the changes in the characteristics of each indicator at different stages were explored.The results show that the vulnerability of buildings generally increases during the disaster period,mainly due to the degradation of factors such as maintenance status,opening quality and vegetation coverage.In the post-disaster recovery period,for buildings with lower vulnerability than before the disaster,it is mainly due to the construction of walls,improvement of opening quality and recovery of maintenance status and vegetation coverage.The opening position and shielding effect are factors that have not been paid attention to in the reconstruction process of the disaster area.They are the main direction for further optimization and reduction of building vulnerability in future disaster prevention and mitigation work.This study has preliminarily developed a dynamic evaluation model for building vulnerability suitable for Wenchuan debris flow-prone areas,providing a new idea and method for building vulnerability evaluation.This helps to formulate more effective disaster risk management strategies and can provide decision-makers and stakeholders with disaster reduction plans.At the same time,it provides practical guidance for designing and improving building structures and improving the surrounding environment to reduce the vulnerability of buildings to debris flows.In addition,the social significance of this study lies in providing valuable insights into the spatial and temporal dynamics of building vulnerability in Wenchuan earthquake area.This information can be used to strengthen regional disaster preparedness work,reduce the risk of casualties and economic losses,and help develop more resilient communities.