编目图数据形式对区域崩塌易发性评价结果的影响

IMPACT OF INVENTORY FORMS ON RESULT OF REGIONAL ROCKFALL SUSCEPTIBILITY ASSESSMENT

本文以北京市怀柔区为例,通过现场调查,对688处崩塌灾害分别以面数据和点数据的形式获取了两套编目图。根据现场调查和资料分析,选取岩性、地形、断裂和道路建设作为该区崩塌灾害的主控因素,采用频率比(FR)模型对崩塌灾害的易发性进行了评价。为了对评价结果的预测性进行检验,采用随机分割法,选取了415处崩塌用于频率比模型的计算,剩余的273处崩塌用于评价结果预测性的验证。预测曲线表明,基于崩塌面数据的评价结果比基于点数据的评价结果具有明显的优越性。根据基于面数据的频率比模型评价结果,可以将研究区的崩塌灾害易发性划分为5个等级:较低易发(占全区14%)、低易发(占全区20%)、中等易发(占全区27%)、高易发(占全区22%)和极高易发(占全区17%)。相关工作和结论可以为区域地质灾害易发性评价中编目图的编制提供参考,并为怀柔区区域国土利用和防灾减灾提供指导。

Regional geo-hazard susceptibility assessment is the basis for early warning of geo-hazards. Huairou district is located in the north of Beijing and has favorable geological conditions to rockfall. The widely developed rockfalls threat to the lives and property of local people. Therefore,the region rockfall susceptibility assessment is significantly meaningful. According to investigation,two forms of rockfall inventory maps including point distribution and area distribution for the same 688 rockfalls in Huairou district of Beijing are obtained on site. The rockfall inventory maps provide the necessary background information for building the susceptibility model,which affects the reliability of evaluation. In order to quantitatively assess rockfall susceptibility,frequency ratio( FR) model is applied using GIS.In general,to predict landslides,it is normally assumed that rockfall occurrence is determined byrockfall-related factors,and that future rockfalls will occur under the same conditions as past rockfalls. For the modeling,lithology,terrain,faults and roads construction are considered as significant rockfall-related factors in rockfall occurrences based on site investigation and statistical analysis. After assessing the susceptibility maps,validation is essential to determine the predictive value of susceptibility maps. It also helps to better select the most suitable function and significant variables,thus improving the efficiency of the mapping process. Validation can be performed using a rockfall population independent from the one used to construct the map. Therefore,FR model is run including only 415 rockfalls that randomly selected from the total rockfalls and then validated with respect to the left 273 rockfalls. The prediction-rate curve based on rockfall area inventory shows a better performance in prediction than that based on rockfall point inventory,which reveals that the uncertainty exists in rockfall point inventory. Based on the FR model using rockfall area inventory map,Huairou district is classified into five classes of rockfall susceptibility: very low susceptibility zone( 14% of the whole area),low susceptibility zone( 20%),medium susceptibility zone( 27%),high susceptibility zone( 22%),and very high susceptibility zone( 17%). This study can provide reference for inventory mapping in regional rockfall susceptibility assessment. The susceptibility map can be a helpful tool for regional land use planning and disaster prevention and mitigation in Huairou district.