Assessment On Disaster Risk Reduction Capacity With TOPSIS And Machine Learning And Analysis Of Regional Disaster Characteristics Of Jinshan District In Shanghai

China is one of the countries which are hardest hit by natural disasters.The suddenness and complexity of disasters are exacerbated by global climate change and human activities.Frequent natural disasters will cause huge economic losses and human casualties.Shanghai has a dense population,buildings and all kinds of infrastructure as an important mega-coastal city in China.As a coastal city with flat terrain,it is frequently affected by typhoons and storm surges.Besides,the geological environment of Shanghai is relatively fragile with mollisol covering,which causes land subsidence.Jinshan District is a coastal suburb of Shanghai that is rapidly promoting urban construction and renewal.Not only numerous people live in coastal areas,but also a large number of hazardous chemical enterprises locate in coastal areas.In order to prevent,resist and reduce the impact of disasters,On one hand,it is necessary to have an objective understanding of regional disaster reduction risk(DRR)capacity.On the other hand,analyzing the characteristics of regional disasters is also important.Combining the two aspects,exploring strategies to improve regional disaster reduction capacity is of great significance for Jinshan District to prevent and defuse disaster risk and reduce disaster losses.This study assesses the DRR capacity of Jinshan District based on Technique for Order Preference by Similarity to an Ideal Solution(TOPSIS).In order to build more efficient and intelligent evaluation models,machine learning is used to assess disaster reduction capacity.We derive the whole-zone surface deformation velocity from 2016to 2022 and seawall surface deformation velocity from 2020 to 2022 based on Small Baseline Subset(SBAS)and Sentinel-1A remote sensing data.Considering comprehensively the impact of seawall failure and extreme water levels in two scenarios,we simulate the possible flood inundation range after the seawall failure based on LISFLOOD-FP simulation model in Jinshan District.For towns with weak DRR capacity,we analyze the sensitivity of evaluation indicators to explore key indicator which affects the improvement of DRR capacity.Finally,we propose some strategies which could improve DRR capacity of Jinshan District based on the assessment results of DRR capacity and regional disaster characteristics.In this work,we evaluate DRR capacity of town(street)and community(village)in Jinshan District based on TOPSIS and the evaluation indicator system of the First National Survey on Integrated Risk of Natural Disasters with equal weight.Eleven streets and towns in Jinshan District are divided into five levels from strong to weak.Then,the influence of change of evaluation indicator weight on the assessment results is also analyzed based on extremum method.It is found that change of indicator weight has no significant influence on the assessment results.The disaster reduction risk capacity levels of town(street)and community(village)of most streets and towns do not change,and the changes also float between adjacent levels.The DRR capacity assessment data will constantly change with the use and input of human,material,and financial resources.Therefore,machine learning was introduced into the assessmentof DRR capacity of all streets and towns based on the simulation data in Shanghai.It is found that the results evaluated by Bagging and Random Forest are mostly consistent with the results evaluated by TOPSIS and the inconsistent results float between adjacent grades.Assessment of DRR capacity is an assessment that does not take into account the level of regional comprehensive disaster risk or disaster intensity.Therefore,we analyzed the characteristics of regional disasters in Jinshan District.Firstly,the Line of Sight(LOS)deformationtime series of Jinshan District and seawall were retrieved by using SBAS method based on 119 Sentinel-1A images from 2016 to 2022 and 87Sentinel-1A images from 2020-2022 with ascending orbit.It was found that the land subsidence of Jinshan District mainly occurred in the border area between Caojing Town and Fengxian District and the north of Tinglin Town.The subsidence velocity of seawalls in the south of Shanyang Town is the highest,which reaches 20 mm/year.In order to further explore the impact of seawall failure,the seawall was divided into24 units with a length of 1 km and 12 units with a length of 2 km.Based on LISFLOOD-FP two-dimensional hydrodynamic model and the Global Tide and Surge Reanalysis(GTSR)dataset and the extreme water level data caused by storm surge in Jinshan District,the flood inundation area was simulated in the case of failure of every seawall unit.The results show that the flood inundates an area of more than 30km~2when the seawalls in the south of Caojing Town and Shanyang Town are damaged.Based on the assessment results of DRR capacity,we focused on the streets whose DRR capacity of town(street)and community(village)are weak.Using a simple variable method,sensitivity analysis is made on the indicators participating in the assessment to explore the key indicators influencing DRR capacity.The results show that the number of disaster management personnel has the most significant impact on the DRR capacity of town(street).The number of medical and health institutions,the total amount of reserve materials and equipment invested,and the funds for disaster prevention and reduction can improve the DRR capacity of the community(village).Based on this,combined with the characteristics of regional disaster in Jinshan District,strategies to improve DRR capacity are proposed for areas with weak DRR capacity,possible flooding,and land subsidence,providing disaster risk information and scientific decision-making basis for improving comprehensive regional DRR capacity.