基于随机森林优化算法的城市群洪涝灾害韧性时空演变研究∗

Study on Spatiotemporal Evolution of Flood Disaster Resilience in Urban Agglomeration Based on Random Forest Optimization Algorithm

摘要:

为了提升城市应对灾害的能力，对城市洪涝灾害韧性进行研究。基于“ 驱动力-状态-响应（DSR）”模型建立城市洪涝灾害韧性评估指标体系，采用随机森林优化算法（RF）构建城市洪涝灾害韧性评估模型，通过均方根误差（RMSE）和决定系数（R2）评价模型性能。以关中平原城市群为例，分析西安、宝鸡、咸阳、商洛、铜川、渭南、临汾、运城、平凉、天水、庆阳等11 座城市2011~2020 年的城市洪涝灾害韧性时空演变规律。结果表明：驱动力和状态对城市洪涝灾害韧性水平影响较大，响应对城市洪涝灾害韧性水平影响较小；在时间序列上，关中平原城市群各城市洪涝灾害韧性水平发展存在较大差异，但整体处于平稳上升的状态，其中铜川的城市洪涝灾害韧性水平提升最为明显；在空间格局上，城市群洪涝灾害韧性水平呈现出由核心城市向外围递减的空间结构特征，各城市的洪涝灾害韧性等级的空间分布存在不均衡的现象。研究结果可为关中平原城市群韧性城市建设提供指导。

Abstract:

To improve the capacity of cities to cope with disasters, the study focuses on urban flood di-saster resilience. An evaluation index system for urban flood disaster resilience was established basedon the "driving force-state-response (DSR)" model. An assessment model was constructed using a ran-dom forest optimization algorithm (RF). Model performance was evaluated using the root mean squareerror (RMSE) and the coefficient of determination (R2). Taking the Guanzhong Plain urban agglomera-tion as an example, the spatiotemporal evolution of urban flood disaster resilience in Xi'an, Baoji,Xianyang, Shangluo, Tongchuan, Weinan, Linfen, Yuncheng, Pingliang, Tianshui, and Qingyangfrom 2011 to 2020 was analyzed. The results showed that driving force and state significantly impact-ed the resilience level of urban flood disaster, while response had a milder impact. Over the time se-ries, significant differences in the development of flood disaster resilience levels were observed amongthe cities in the Guanzhong Plain urban agglomeration, though the overall trend was rising steadily.Tongchuan exhibited the most significant improvement in flood disaster resilience. Spatially, resil-ience levels showed a declining gradient from core cities to the periphery, with an imbalance in spatialdistribution of flood disaster resilience levels among the cities. The research results provide guidancefor the construction of resilient cities in the Guanzhong Plain urban agglomeration.