AN INVERSE AND DECOMPOSITIONAL ANALYSIS OF CHAIN TRIGGER FACTORS FOR SLOPE FAILURE HAZARD ZONATION

摘要

This paper presents an inverse- and decompositional-analysis of unobserved "chain-trigger factors" according to slope failure, based on Structural Equation Modeling (SEM). Quantitative prediction models for slope failures generally elucidate the relationship between past slope failures and causal factors (e.g. geology, soil, slope, aspect, etc.). Due to the difficulties of obtaining pixel-based observations on the trigger factors (e.g. rainfall, earthquake, weathering, etc.), the trigger factors as explanatory variables are substituted for some of the causal factors in constructing prediction models, on the assumption that there are some correlations between causal and trigger factors. As a measure, we had tackled to construct a Trigger Factor Inverse analysis model (TFI model) in which the relationship between past slope failures (i.e. endogenous variables), causal factors (i.e. explanatory variables), and trigger factors (i.e. unobserved variables) are delineated on the path diagram in SEM approach. In the TFI model, through the "measurement equation" defined between the causal factors (i.e. observed variables) and the trigger factors (i.e. unobserved latent variable), the trigger factor can be inversely estimated. As the subsequent subjects for the previous studies, in this contribution, we have tried to decompose trigger factors into the "1st trigger factor" and the "2nd trigger factor" with respect to slope failures, which had been induced by Niigata Heavy Rainfall (Jul. 13, 2004:Case1) and Niigata Chuetsu Earthquake (Oct. 23, 2004:Case 2). The analytical procedure consists of the following steps. (1) Step 1: The 1st and the 2nd Trigger Factor Influence maps (TFI map) are produced according to Case1 and Case 2, respectively. (2) Step 2: The differences in these TFI maps are delineated on a "difference (DIF) maps," which are also summarized on the "pair-wise comparative table." (3) Step 3: Through the Hayashi's quantification method of the fourth type, the scatter-diagram is delineated with respect to items corresponding to each TFI map. By using those scatter-diagram jointly with the pair-wise comparative table, the effective and efficient analysis on the "chaintrigger factors" can be achieved with respect to slope failures, simultaneously.