论文部分内容阅读
云南省是我国各种地质灾害的多发区。该文以2000-2014年云南省发生的500个典型地质灾害事件为基础,选取与地质灾害相关的诱发因素因子和潜在形成条件因子,构建了相应的评价指标体系。然后采用地理加权回归模型对其进行易发性评价,并对结果进行了等级划分。最后利用引入空间准确测量思想的成功率曲线和ROC曲线,对评价结果进行了精度验证和分析比较。结果表明:(1)云南省地质灾害较高和极高易发区主要位于西北部地区,且以泸水县、中甸县和维西傈僳族自治县为最高;(2)对评价结果进行了分级,从低到高不同易发性等级的面积百分比分别为6.14%,24.80%,18.41%,24.68%,20.78%和5.18%;(3)在未引入空间准确测量思想情况前,成功率曲线和ROC曲线的AUC值分别为0.503 8和0.965 1;引入空间准确测量思想后,AUC值分别为0.681 3和0.992 7,证明所提基于地理加权回归模型的云南省地质灾害易发性评价具有较高的可靠性。
Yunnan Province is a frequent area with various geological disasters in our country. Based on the 500 typical geological disasters in Yunnan Province from 2000 to 2014, this paper selects the predisposing factor factors and potential formation condition factors related to geological disasters, and constructs the corresponding evaluation index system. Then it uses geo-weighted regression model to evaluate its vulnerability and classifies the results. Finally, the success rate curve and ROC curve of the thought of introducing space accurate measurement are used to verify and analyze the evaluation results. The results show that: (1) Yunnan and Yunnan provinces with high and extremely high levels of geological disasters are mainly located in the northwestern region, with Lushui County, Zhongdian County and Weisi Lisu Autonomous County as the highest; (2) Grading the evaluation results, The percentages of the areas with different levels of vulnerability from low to high were 6.14%, 24.80%, 18.41%, 24.68%, 20.78% and 5.18%, respectively. (3) The success rate curve and ROC The AUC values of the curves are 0.503 8 and 0.965 1, respectively. After introducing the idea of accurate measurement of space, the AUC values are 0.681 3 and 0.992 7, respectively. It proves that the evaluation of the vulnerability of geological disasters in Yunnan Province based on the geo-weighted regression model has a high reliability.