Abstract
Surface soil thickness is important for assessing slope failure hazards and estimating debris
flow impacts. Here, we study the estimation of surface soil thickness using multivariate analyses
considering a combination of various data types. Two patterns corresponding to the objective
and explanatory variables were used for multivariate analyses to evaluate which combination
best estimated surface soil thickness. Moreover, because the data included a significant amount
of shallow surface soil thickness, we also assessed the estimation restricting the amount of these
data. As a result, the grid size did not affect the estimation accuracy, and the estimated values
tended to be smaller than the measured values at sites with thick soils for all patterns. Using the
exposed rock and surveyed soil thicknesses improved the accuracy of the estimation, and
restricting the number of shallow surface soil thicknesses improved the accuracy of the soil
thickness estimation. Thus, these parameters were concluded to be useful for improving the
accuracy of the surface soil thickness estimation using multivariate analysis.
