Coupling TRIGRS and TOPMODEL in Shallow Landslide Prediction

碩士 === 國立中央大學 === 應用地質研究所 === 99 === Water infiltration can cause an increase in unit weight of soil and a decrease in strength of soil. In the past, the Transient Rainfall Infiltration and Grid-based Regional Slope-stability (TRIGRS) model uses an infinite-slope stability analysis with only one-dim...

Full description

Bibliographic Details
Main Authors: Hao-wei Li, 李浩瑋
Other Authors: Chyi-Tyi Lee
Format: Others
Language:zh-TW
Published: 2011
Online Access:http://ndltd.ncl.edu.tw/handle/86507871965703795249
Description
Summary:碩士 === 國立中央大學 === 應用地質研究所 === 99 === Water infiltration can cause an increase in unit weight of soil and a decrease in strength of soil. In the past, the Transient Rainfall Infiltration and Grid-based Regional Slope-stability (TRIGRS) model uses an infinite-slope stability analysis with only one-dimensional vertical infiltration. The lateral flow, however, was not considered in the model, whereas the Topography-based hydrological MODEL (TOPMODEL) can describe the tendency of lateral water accumulation. This model has the merit of simplicity and the lateral flow processes using a limited amount of watershed topographic information. In the present study we use a conceptual framework of coupling TRIGRS and TOPMODEL to estimate groundwater table in a drainage basin, and conduct an infinite slope analysis to determine the instability of grid points. The study area was selected at Piya creek watershed of the Than River and the study aimed at predicting shallow landslides during the Typhoon Aere. The result indicates that the overall accuracy and the area under the success rate curve (AUC) of coupling TRIGRS and TOPMODEL are 89.1% and 0.822, respectively, whereas the overall accuracy and the AUC of TRIGRS only are 87.4% and 0.787, respectively. Therefore, considering lateral flow in the present proposed model does help increase accuracy for shallow landslide. The model is validated by the data set from the Matsa typhoon event. The result of validation shows that AUC of the prediction rate curves are 0.813, showing a good accuracy in predicting shallow landslides.