摘要
区域预警是提高防灾减灾意识、有效减轻地质灾害造成损失的重要手段。地质灾害显式统计预警理论考虑了地质环境变化与降雨参数等多因素的耦合作用,克服了传统单一临界雨量判据方法的局限。笔者应用地质灾害显式统计预警的基本原理,以中国的东南区为例开展应用研究。选取岩土体类型、地形起伏等12个基础地质环境因素.通过确定性系数模型(CF)综合分析了地质灾害分布与地质环境基础因素的关系.选取地质灾害"潜势度"作为地质环境优劣的指标,并进行了定量计算。选取当日雨量和一个降雨过程的前期累计雨量作为降雨激发因素的指标.采用多元回归的统计分析方法,分析了地质环境因素、降雨激发因素的耦合作用与地质灾害发育情况之间的关系.建立了显式统计的地质灾害预警预报模型。以2006年5月18日台风"珍珠"登陆期间的实况预警情况对模型进行了应用校验。验证了显式统计预警原理及模型方法的可行性和实用性。
Regional early warning is an important means for improving the consciousness of geohazard prevention and reduction and effectively mitigating the loss caused by geohazards. In the early-warning method of explicit statistics, the effect of coupling of multiple factors such as geo-environmental change and precipitation is considered, thus overcoming the limitation of the single critical rainfall criterion. Based on the explicit statistic theory on geohazard early warning, applied studies have been carried out in the southeast of China. Twelve geoenvironmental factors such as rock and soil types and topographic relief are selected to analyze the relationship between the distribution of geohazards and geo-environmental factors by using the determinacy coefficient function (CF), and the geohazard potentiality parameter is selected as an index of the geo-environment and quantitative calculation is performed. The very day rainfall and previous cumulative rainfall in one rain process are selected as an index of rainfall excitation conditions. Finally, the multi-regression analysis is used to analyze the relationships of the geo-environmental factor, rain excitation conditions and geohazards. Then an explicit statistic early-warning model for geohazards is constructed. Take for example the early warning during the land striking of the typhoon "Pearl" on May 18, 2006, the authors check the forecasting model and prove that the principle and model of explicit statistics early warning are feasible and practicable.
出处
《中国地质》
CAS
CSCD
2008年第2期344-350,共7页
Geology in China
基金
中国地质调查局地质大调查项目(1212010640333)资助
关键词
地质灾害
潜势度
区域预警
显式统计
预警模型
结果检验
geohazard
potentiality parameter
regional early warning
explicit statistics
early -warning model
multi-regression analysis
result checkout