| Heilongjiang Province is located in the northeast of China.Its total grain output accounts for oneninth of the country’s total grain output and ranks first in the country for 13 consecutive years.It is my country’s largest commodity grain base and grain strategic reserve base.The unique soil environment and favorable rain and heat characteristics in the growing season of crops in Heilongjiang Province give it unique advantages in agricultural production.However,in order to pursue the continuous improvement of grain production capacity,the irrational development and utilization of water and soil resources has resulted in serious local water and soil erosion,a large proportion of food crops,and a relatively simple industrial structure.Coupled with the increased risk of flood disasters induced by extreme climate changes in recent years,the economic and social development of Heilongjiang Province has been seriously affected.Aiming at the flood disaster risk problem faced by Heilongjiang Province,taking Heilongjiang Province and its 13 districts as an example,a support vector machine model based on Butterfly Optimization Algorithm(Improved Support Vector Machine Model Based on Butterfly Optimization Algorithm,BOA-SVM)was constructed and It has been applied to the risk assessment and spatiotemporal characteristics of flood disasters in Heilongjiang Province in the past 15 years,and then to identify the key driving factors and driving mechanisms that affect its risk changes.On this basis,by simulating the evolution trend of flood disaster risk under various scenarios in Heilongjiang Province in the future,some valuable information can be provided for effective regulation and reduction of regional flood disaster risk.The main research results are as follows:(1)Based on the disaster system theory,a flood disaster risk assessment index framework was constructed,and a total of 39 indicators were selected from the four levels of disaster-causing factors,disaster-forming environment,disaster-bearing bodies,and disaster prevention and mitigation capabilities to form a preliminary selection set of flood disaster risk assessment indicators in Heilongjiang Province.The information contribution rate eliminated 7 indicators with less information content,and then used the R clustering-coefficient of variation method to perform secondary screening on the remaining 32 indicators to eliminate a total of 13 indicators,and finally retained the 19 indicators to construct flood disaster risk assessment An optimal set of index systems.Through the index system reliability coefficient and comparative analysis with the existing research index screening methods,it is verified that the index screening method has obvious advantages,and the obtained flood disaster risk assessment index system is more reasonable and reliable.(2)Apply the constructed BOA-SVM model to assess the flood disaster risk in Heilongjiang Province from 2003 to 2017,and analyze the spatiotemporal variation characteristics according to the assessment results.The results show that during the study period,the overall flood disaster risk level in Heilongjiang Province fluctuated significantly in the early stage,and gradually stabilized in the later stage,showing a spatial distribution pattern of high in the northwest and low in the southeast.Among them,the flood disaster risk level in Daqing is the lowest,the risk level in Suihua is the highest,and the risk level in other areas has a clear downward trend with the inter-annual variation.In order to further verify the reliability of the model,the traditional support vector machine model(Support Vector Machine,SVM)and the Imperialist Competitive Algorithm-Support Vector Machine model(ICA-SVM)were introduced for comparative analysis with the BOA-SVM model,and it was found that the BOA-Compared with the SVM and ICA-SVM models,the MAE of the SVM model decreased by 38.15% and 9.18%,the MSE decreased by 58.5% and 21.56%,the MAPE decreased by 35.23% and 11.42%,and the R2 increased by0.62% and 0.12 respectively.%,indicating that the BOA-SVM model constructed in the study has more advantages in terms of fitting,adaptability,stability,reliability and evaluation accuracy.(3)Three methods of entropy weight method,Person correlation coefficient method and BOA-SVM model built-in weight coefficient calculation method were used respectively to calculate the weight of 19 indicators in the optimal set of evaluation index system and apply the serial number sum theory for rational sorting,and found that the water production Modulus,GDP per capita,monthly strongest precipitation,proportion of total output value of agriculture,forestry and fishery,natural population growth rate,total storage capacity of reservoirs per 10,000 hectares,and number of health care beds per10,000 people have a greater impact on flood disaster risk,which was identified as a key driver affecting flood disaster risk.In order to explore the impact of each key driving factor in the process of flood disaster risk change,the impact effect analysis was carried out from the scales of Heilongjiang Province and each region,and it was concluded that the monthly strongest precipitation and per capita GDP affect the flood disaster risk main driving force.(4)Apply the scenario analysis method to set four scenario models: trend stabilization scenario(S1),economic development scenario(S2),economic decline scenario(S3)and extreme climate scenario(S4),and use data forecasting methods to analyze the key driving forces.The factors are used to predict the trend,and then the predicted value is brought into the BOA-SVM model to calculate the flood disaster risk index of each year and draw the simulation curve of the flood disaster risk change trend in 2025,2030 and 2035.Through comparative analysis,it is found that the decline rate of flood disaster risk is different under different levels of economic development.In the face of extreme precipitation,the risk of flood disasters in the context of stable economic development still rises significantly.It can be seen that extreme climate change in the context of global warming has increased the threat of flood disaster risks.high.Considering the spatiotemporal variation characteristics of flood disaster risk in Heilongjiang Province,the driving mechanism and its future evolution trend,a flood disaster risk control strategy is initially proposed,which provides an important reference for flood control and disaster reduction and coordinated development in Heilongjiang Province and other urban areas. |
PDF Full Text Request