Abstract
Although the effective rainfall method has been widely used for analyzing variations in the groundwater level (GL) in a landslide site, few previous studies have applied the method to snow-cover season. This study estimated the snowmelt water without using large-scale observation facilities and calculated the meltwater and/or rainwater (MR), which can affect the GL variation. Then, the half-life of the effective MR (EMR), delay days, and coefficients of a regression line were optimized to reproduce the GL variation the best. The results showed that EMR is a significant explanatory variable of the observed GL, suggesting that the method used in this study is very effective to analyze the GL variation in a landslide site in snow-cover season, where large-scale observation facilities are difficult to install. Properties of the GL response were different between two periods probably due to the effects of pore air, snow distribution property in the last stage of snowmelt period, horizontal drainage boring, and evapotranspiration. Individual analyses for the two periods improved the accuracy of the GL reproducibility. The parameter values reported in previous studies and this study showed no clear relationship with landslide properties such as geology and the depths of GL variation, suggesting that more cases are needed to predict parameter values from landslide properties.
