2020 Volume 57 Issue 1 Pages 12-18
In areas hit by landslide, the fluctuation characteristics of the groundwater level may change. This study used Bayesian optimization to estimate the parameters of a groundwater level reproduction model under the assumption that they are time-varying coefficients. Then, we used parameter fluctuations and functions to evaluate how the groundwater level fluctuation characteristics changed over a specific analysis period. We applied this method to a landslide area where a groundwater drainage works were constructed, reproduced the groundwater level fluctuation, and evaluated the changes in the characteristics of the groundwater level fluctuation. Changes in groundwater level are shown to be accurately reproduced by varying each parameter. Furthermore, the tendency of parameter fluctuation indicated that the drop amount in groundwater level during no rainfall was increased after groundwater drainage works.