Chen Jianping, Wang Zepeng, Chen Wei, Wan Changyuan, Liu Yunyan, Huang Junjie
College of Mining, Liaoning Technical University, FuXin, 123000, China.
College of Environment, Liaoning Technical University, FuXin, 123000, China.
Environ Sci Pollut Res Int. 2023 Mar;30(15):44756-44772. doi: 10.1007/s11356-023-25454-2. Epub 2023 Jan 26.
The reasonable selection of non-geological disaster samples is of great significance to improve the accuracy of geological disaster assessment, reduce the cost of disaster management, and maintain the sustainable development of ecological environment. Liulin County was selected as the study area. This paper creatively divided non-geological disaster sampling areas by macro-geomorphology, and carried out susceptibility mapping based on random forest (RF) and frequency ratio-random forest (FR-RF) models. The accuracy of each model was evaluated by receiver operating characteristic curve (ROC) combined with the distribution characteristics of geological disasters and the actual urban construction in the study area. The results show that the FR-RF model constructed by selecting non-geological disaster samples in hilly area is most suitable for the susceptibility mapping of this study area. The different results in different sampling areas are mainly due to the great changes in the representativeness of non-geological disaster samples. The distance from the roads is the most important factor affecting the occurrence of disasters in the study area. The statistical results of disaster management cost estimation and gross domestic product (GDP) value show that the disaster management cost of HFR-RF model decreases by 13.45% on average compared with other models, and the ratio of GDP to disaster management cost is relatively high. These research results promote the progress of geological disaster prevention technology, maintain the stability of geological environment, and are of great significance to the stable and sustainable development of local economy.
合理选取非地质灾害样本对于提高地质灾害评估精度、降低灾害管理成本以及维护生态环境可持续发展具有重要意义。选取柳林县作为研究区域。本文创新性地依据宏观地貌划分非地质灾害采样区，并基于随机森林（RF）模型和频率比 - 随机森林（FR - RF）模型开展易发性制图。通过结合研究区域地质灾害分布特征和实际城市建设情况，利用接收者操作特征曲线（ROC）对各模型精度进行评估。结果表明，在丘陵区选取非地质灾害样本构建的FR - RF模型最适用于本研究区域的易发性制图。不同采样区结果不同主要是由于非地质灾害样本代表性变化较大。道路距离是影响研究区域灾害发生的最重要因素。灾害管理成本估算和国内生产总值（GDP）值的统计结果表明，与其他模型相比，HFR - RF模型的灾害管理成本平均降低了13.45%，且GDP与灾害管理成本之比相对较高。这些研究成果推动了地质灾害防治技术的进步，维护了地质环境稳定，对地方经济的稳定可持续发展具有重要意义。
