GIS-based rare events logistic regression for landslide-susceptibility mapping of Lianyungang, China
GIS-based rare events logistic regression for landslide-susceptibility mapping of Lianyungang, China
复制标题
基于 GIS 的连云港滑坡敏感性逻辑回归稀有事件测绘
DOI：
10.1007/s12665-010-0509-3
复制
发表时间：
2011-01-01
影响因子：
2.8
通讯作者：
Ding, Liang
中科院分区：
文献类型：
--
作者：
Bai, Shibiao;Lue, Guonian;Ding, Liang
滑坡对人类生活、环境和当地经济发展产生了巨大影响，因此需要充分了解滑坡。在这项研究中，我们提出了一种方法来分析和建模的滑坡数据，使用罕见事件逻辑回归和应用的方法，在中国连云港地区。收集了该地区的数字正射影像图、数字高程模型、地质图和不同的地理信息系统图层，包括定居点、道路网和河流，并将其应用于分析。通过单视场人工判读识别滑坡，并在现场调查过程中进行验证。为了验证映射的质量，将研究区域的数据分为训练集和验证集。结果表明，4.26%的研究区被确定为具有非常高的滑坡敏感性，而其他分类为具有非常低的敏感性（47.2%），低敏感性（22.21%），中等敏感性（14.39%）和高敏感性（11.93%）。验证了本文所编制的滑坡易发性图的质量，可用于滑坡防治规划。滑坡易发性图是连云港市滑坡危险性评价的基础。
Landslides have had a huge effect on human life, the environment and local economic development, and therefore they need to be well understood. In this study, we presented an approach for the analysis and modeling of landslide data using rare events logistic regression and applied the approach to an area in Lianyungang, China. Digital orthophotomaps, digital elevation models of the region, geological maps and different GIS layers including settlement, road net and rivers were collected and applied in the analysis. Landslides were identified by monoscopic manual interpretation and validated during the field investigation. To validate the quality of mapping, the data from the study area were divided into a training set and validation set. The result map showed that 4.26% of the study area was identified as having very high susceptibility to landslides, whereas the others were classified as having very low susceptibility (47.2%), low susceptibility (22.21%), medium susceptibility (14.39%) and high susceptibility (11.93%). The quality of the landslide-susceptibility map produced in this paper was validated, and it can be used for planning protective and mitigation measures. The landslide-susceptibility map is a fundamental part of the Lianyungang city landslide risk assessment.