逻辑回归模型在堰塞湖应急风险评估中的应用
|
摘要
堰塞湖是一种由地震、降水等原因引发的山崩、滑坡堵截河道,使河道贮水形成湖泊的常见地质现象。大部分堰塞湖会在形成后短时间内溃决,所以快速准确地评价堰塞湖溃决的风险就显得相当重要。从应急评估的要求出发,综合考虑影响堰塞湖溃决的关键因素,借助主成分分析确定评价指标。在此基础上利用日本的36例堰塞湖及汶川地震形成的8例堰塞湖建立了2种评估堰塞湖稳定性的逻辑回归模型,其预测结果的准确率均为86.4%,证明其有较好的预测能力。同时文章进一步讨论了如何利用逻辑函数值进行堰塞湖风险评估,并总结了逻辑回归模型的优势与不足。
Dammed Lake is a common geological phenomenon which happens when landslides caused by earthquakes,rainfalls block the river course.After the formation of a dammed lake,it usually fails in a short period of time.Thus,it's vital to assess the probability of dammed lakes' breaching quickly and precisely.This paper focuses on the demand of emergent risk assessment,takes chief factors that influence the stability of dammed lakes into consideration to select several significant variables,and utilizing principal components analysis(PCA)to determine the combination of these variables.Based on a dataset consisting of 36 Japanese dammed lakes and 8Chinese dammed lakes showed after the Wenchuan earthquake,this study establishes two logistic regression models.The accuracy of two models' forecast is 86.4%,showing that they have rather high predictive abilities.Then this paper discusses how to use the results of two models to assess the risk of dammed lakes,and discusses advantages and limitations of the logistic regression models.
Dammed Lake is a common geological phenomenon which happens when landslides caused by earthquakes,rainfalls block the river course.After the formation of a dammed lake,it usually fails in a short period of time.Thus,it's vital to assess the probability of dammed lakes' breaching quickly and precisely.This paper focuses on the demand of emergent risk assessment,takes chief factors that influence the stability of dammed lakes into consideration to select several significant variables,and utilizing principal components analysis(PCA)to determine the combination of these variables.Based on a dataset consisting of 36 Japanese dammed lakes and 8Chinese dammed lakes showed after the Wenchuan earthquake,this study establishes two logistic regression models.The accuracy of two models' forecast is 86.4%,showing that they have rather high predictive abilities.Then this paper discusses how to use the results of two models to assess the risk of dammed lakes,and discusses advantages and limitations of the logistic regression models.
引文
[1]Schuster,Robert L.Perspective on Landslide Dams[J].Landslide Dams,Processes,Risk and Mitigation,1986,(3):1-20.
[2]Ermini L,Casagli N.Prediction of the behaviour of landslide dams using ageomorphological dimensionless index[J].Earth Surface Processes and Landforms,2003,28(1):31-47.
[3]崔鹏,韩用顺,陈晓清.汶川地震堰塞湖分布规律与风险评估[J].四川大学学报(工程科学版),2009,41(3):35-42.
[4]陈风光,姚海林,史卫国.模糊综合评价法在堰塞湖风险评估中的应用[J].上海交通大学学报,2011,(8):67-70.
[5]孙研,王绍玉.基于自然和社会属性的堰塞湖风险评估[J].四川大学学报(工程科学版),2011,43(1):24-28.
[6]童煜翔.山崩引致之堰塞湖天然坝体稳定性之量化分析[D].台湾:中央大学应用地质研究所,2008.
[7]Dong J J,Tung Y H,Chen C C,et al.Logistic regression model for predicting the failure probability of a landslide dam[J].Engineering Geology,2011,117(1):52-61.
[8]田畑茂清,水山高九,井上公夫.天然ダ.と災害[M].东京:古今书院,2002.
[9]崔伟才,肖泉,乔路,等.四川省绵竹市绵远河上游二岗桥至滴水岩河段堰塞湖安全性评价报告[R].成都:四川大学工程设计研究院,2009.
[10]冯利华.环境质量的主成分分析[J].数学的实践与认识,2003,33(8):32-35.
[11]马国庆.管理统计-数据获取、统计原理SPSS工具与应用研究[M].北京:科学出版社,2006.
[2]Ermini L,Casagli N.Prediction of the behaviour of landslide dams using ageomorphological dimensionless index[J].Earth Surface Processes and Landforms,2003,28(1):31-47.
[3]崔鹏,韩用顺,陈晓清.汶川地震堰塞湖分布规律与风险评估[J].四川大学学报(工程科学版),2009,41(3):35-42.
[4]陈风光,姚海林,史卫国.模糊综合评价法在堰塞湖风险评估中的应用[J].上海交通大学学报,2011,(8):67-70.
[5]孙研,王绍玉.基于自然和社会属性的堰塞湖风险评估[J].四川大学学报(工程科学版),2011,43(1):24-28.
[6]童煜翔.山崩引致之堰塞湖天然坝体稳定性之量化分析[D].台湾:中央大学应用地质研究所,2008.
[7]Dong J J,Tung Y H,Chen C C,et al.Logistic regression model for predicting the failure probability of a landslide dam[J].Engineering Geology,2011,117(1):52-61.
[8]田畑茂清,水山高九,井上公夫.天然ダ.と災害[M].东京:古今书院,2002.
[9]崔伟才,肖泉,乔路,等.四川省绵竹市绵远河上游二岗桥至滴水岩河段堰塞湖安全性评价报告[R].成都:四川大学工程设计研究院,2009.
[10]冯利华.环境质量的主成分分析[J].数学的实践与认识,2003,33(8):32-35.
[11]马国庆.管理统计-数据获取、统计原理SPSS工具与应用研究[M].北京:科学出版社,2006.
版权所有：© 2023 中国地质图书馆 中国地质调查局地学文献中心