Influencing Factors Analysis And Assessment Of Storm And Flood Susceptibility

With climate change,extreme weather such as heavy rain frequently occurs,storm and flood disasters have become one of the most common and most harmful natural disasters to human society.Storm and flood susceptibility refers to the possibility of flood disasters when the regional climate,topography,landform,vegetation,hydrology,soil and other environments are subjected to heavy rain,which is the natural attribute of the region.This paper takes Anyi County of Nanchang City as the research object to carry out the analysis and assessment of the influencing factors of storm and flood susceptibility.On the one hand,according to the characteristics of the natural environment of Anyi County,more comprehensive influencing factors were selected,and then a new method for optimizing the selection of influencing factors was proposed,which could enrich the technical methods of flood susceptibility assessment;on the other hand,the flood susceptibility assessment result not only could provide scientific guidance for the formulation of disaster prevention and mitigation measures and the land use planning in Anyi County,but also could provide a reference for the countylevel flood susceptibility assessment of the geographical environment similar to Anyi County in China.The main research contents and conclusions are as follows:(1)The radar image of Sentinel-1 before and after the disaster was used to extract the inundation range of storm and flood disasters in Anyi County,Nanchang City from June 30 to July 5,2016,fifteen influencing factors were selected,including elevation,precipitation,land use,distance from the rivers,convergence index,terrain ruggedness index,slope,plan curvature,topographical wetness index etc,and the frequency ratio model was used to analyze the correlation between the inundation range and the influencing factors of storm and flood susceptibility.The correlation analysis between the inundation range and the influencing factors shows that: precipitation is the direct influencing factor for inducing storm and flood,and the probability of flood occurrence is negatively related to the distance from the rivers,and the probability of flood occurrence is positively related to topographical wetness index and stream power index.(2)The random forest model was used to obtain the importance ranking of the influencing factors,starting with the least important influencing factor,the unimportant influencing factors were gradually streamlined,and the neural network model was designed for optimizing the selection of influencing factors,and the area under the receiver operating characteristic curve(AUC value)was used to measure the performance of the neural network model.The order of importance of the influencing factors shows that the seven most important influencing factors are elevation,distance from the rivers,land use,slope,precipitation,length slope and topographical wetness index.The sum importance of these seven influencing factors exceeds 0.8;the optimal selection of the influencing factors shows that the AUC value of the neural network model improves from 0.99228 to 0.99704 after reducing the five least important influencing factors of convergence index,aspect,profile curvature,topographical position index and stream power index.(3)Based on the optimally selected influencing factors,the neural network model was used to assess the flood susceptibility of Anyi County,and the reliability of the flood susceptibility assessment result was verified by examples.The flood susceptibility assessment shows that: areas with moderate or higher flood susceptibility levels are mainly distributed on both banks of the Liaohe River,occupying approximately one third Anyi County;example verification shows that nearly 70% of the floods are distributed in areas with moderate or higher flood susceptibility levels,and the result of flood susceptibility assessment is consistent with the actual situation in Anyi County.