A new method to analyze the driving mechanism of flood disaster resilience and its management decision-making

Abstract

• Research on Flood disaster resilience based on Bayesian belief network. • Using fuzzy mathematical theory to make decision from the manager's position. • Realizing the prediction of resilience level with missing indicators. • Using diagnostic and sensitivity analysis to explore the driving mechanism. Flood disaster resilience management and decision-making are comprehensive prevention and control strategies that integrate pre-disaster prevention, in-disaster scheduling and post-disaster reconstruction. In view of the many deficiencies in current flood disaster resilience research in terms of uncertainty measurement, dynamic regulation and management decision-making, a new method for the analysis of the driving mechanism of flood disaster resilience and management decision-making based on the fuzzy mathematical theory improved Bayesian belief network (BBN) is proposed. A case study was conducted in 15 state-owned farms under the jurisdiction of the Jiansanjiang Branch in Heilongjiang Province, China. The model predicts the resilience level with missing indicators. A diagnostic analysis of the correlation between indicators and resilience improvement and a sensitivity analysis of the dynamic evaluation of the importance of indicators are performed, and fuzzy decisions are made to quantify different decision-making schemes. The results show that when a single indicator is missing, the prediction accuracy exceeds 80%. However, the lack of economic indicators has a substantial impact on the results, and the accuracy is reduced to 63.3%, indicating that economic indicators are the key dimension affecting the evaluation of flood disaster resilience levels in the study area, which also provides a focus for previous data collection efforts. The diagnostic analysis constructs an index sequence to guide the improvement in resilience, in which the water surface rate and rainfall are the most direct driving forces affecting resilience. On this basis, the differences in the driving forces of each farm are further analyzed in detail. Sensitivity analysis shows that rainfall, per capita GDP, and the energy consumption of agricultural machinery are the indicators that occur most frequently among the top three sensitivity indicators of all farms. The 859 farm, Chuangye farm and Nongjiang farm are typical farms with a high sensitivity to rainfall and economic indicators. Fuzzy decision-making takes the comprehensive grade score of implementation motivation as a quantitative evaluation criterion, which provides guidance for decision-making based on the two dimensions of time and space.