北京大学学报自然科学版 ›› 2024, Vol. 60 ›› Issue (3): 513-520.DOI: 10.13209/j.0479-8023.2024.022

上一篇    下一篇

滑坡位移预测模型补充评价方法

田原1,2, 常啸寅1,2, 赵文祎3,†, 程楚云1,2, Bronte Scheuer1,2, 邓杨兰朵1,2, 马睿平1,2, 张建学1,2
  

  1. 1. 北京大学遥感与地理信息系统研究所, 北京 100871 2. 空间信息集成与3S工程应用北京市重点实验室, 北京 100871 3. 中国地质环境监测院, 北京 100081
  • 收稿日期:2023-05-14 修回日期:2023-06-02 出版日期:2024-05-20 发布日期:2024-05-20
  • 通讯作者: 赵文祎, E-mail: 395447712(at)qq.com
  • 基金资助:
    中国地质调查局地质调查项目(DD20211364)资助

Supplemental Evaluation of Landslide Displacement Prediction Models

TIAN Yuan1,2, CHANG Xiaoyin1,2, ZHAO Wenyi3,†, CHENG Chuyun1,2, Bronte Scheuer1,2, DENG Yanglanduo1,2, MA Ruiping1,2, ZHANG Jianxue1,2
  

  1. 1. Institute of Remote Sensing and Geographical Information Systems, Peking University, Beijing 100871 2. Beijing Key Laboratory of Spatial Information Integration and Its Applications, Beijing 100871 3. China Institute for Geo-Environmental Monitoring, Beijing 100081
  • Received:2023-05-14 Revised:2023-06-02 Online:2024-05-20 Published:2024-05-20
  • Contact: ZHAO Wenyi, E-mail: 395447712(at)qq.com

摘要:

针对目前滑坡位移预测模型评价方法中总体性指标区分能力不足, 以及对滑坡高风险期预测效果评价不准确等问题, 设计滑坡位移预测模型快速位移期综合误差(CERDP)和峰值预测误差(PPE)两个补充评价指标以及对应的补充评价技术流程。基于滑坡监测数据的实例研究表明, 所提补充评价指标和应用流程合理可行, 可以有效地应对均方根误差(RMSE)等指标差异不显著带来的评价不准确问题, 遴选出总体性能可靠, 在风险时段和最高风险时点表现更好的模型, 为滑坡风险管理工作提供有效的技术支持。

关键词: 滑坡位移预测模型, 评价指标, 补充评价, 快速位移期, 位移峰值

Abstract:

The distinguishing ability of typical overall indicators is becoming increasingly insufficient in the evaluation tasks of landslide displacement prediction models while the predicting ability of the models in the highrisk deformation periods has unexpectedly not been paid enough attention. Therefore, two supplemental evaluation indicators for landslide displacement prediction models, combined error of rapid displacement period (CERDP) and peak prediction error (PPE), along with their application process, are accordingly designed. To verify the feasibility and performance of the proposed supplemental evaluation indicators, a case study based on actual landslide monitoring data has been carried out. It can be concluded that the proposed supplemental evaluation indicators and application process are reasonable, feasible, and practical for effectively addressing the evaluation issues caused by insignificant differences in indicators, such as root mean squared error (RMSE), and are able to finally select an overall reliable model which performs best during rapid displacement periods and around peak deformation times, thus providing effective technical support for landslide risk management tasks.

Key words: