Please use this identifier to cite or link to this item:
標題: Comparison of Landslide Probability Analysis Results Based on Different Mapping Units and Probability Models
作者: Chun-Yi Wu
Yen-Chu Yeh
關鍵字: landslide;mapping unit;landslide probability model;崩塌;分析單元;崩塌機率模型
Project: 水土保持學報, Volume 49, Issue 4, Page(s) 2249-2266.
In landslide probability analysis, choice of mapping units will affect the result of probability distribution. In the present study we compared different grid cells and slope unit, and concluded that landslide area and non-continuous phenomenon of probability distribution should be considered when choosing mapping units. The advantage of slope units is the reflection of terrain characteristics. However, it is impossible to show different landslide prone areas in a slope unit. Oppositely, the 20 × 20 m grid cells are more appropriate than slope units to present landslide probability on a slope. In this study we combined Poisson and Binomial probability models with mapping units, and obtained probability models with spatial and temporal capabilities. Next, we compared these two models and got the same probability distribution. One difference is that Binomial model appeared higher probability value than that of the Poisson model. This can be considered as more serious situation of landslide probability distribution than the distribution of Poisson model. The Binomial probability model can be used in the initial analysis of landslide probability distribution in a region.

在進行崩塌機率分析時,分析單元之選擇將影響機率分佈之結果,本研究針 對不同網格及邊坡單元進行比較,歸納分析單元之選擇需考量崩塌地面積和機率 分佈的不連續程度,其中邊坡單元之優點是能反映地形特徵,然而當需了解邊坡 單元內崩塌較易發生在上下邊坡等不同位置時,則以 20 × 20 m 網格較能呈現同 一個坡面之崩塌機率。本研究將 Poisson 及 Binomial 機率模型與分析單元結合, 得出同時具空間、時間機率內涵之機率模型,並比較二者之差異。發現兩種模型 得出之結果有著相同的分佈,唯 Binomial 模型得出之機率值較高,可視為較嚴重 情形之分佈結果,可應用在一地區崩塌機率分佈之初步分析之用。
Appears in Collections:第49卷 第04期

Files in This Item:
File Description SizeFormat
98630-4.pdf1.07 MBAdobe PDFView/Open
Show full item record

Google ScholarTM


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.