Abstract
The Snow and Ice Research Center (SIRC) of the National Research Institute for Earth Science and Disaster Prevention (NIED) of Japan has been developing a snow disaster forecasting system. This system consists of an atmospheric mesoscale model NHM, the numerical snow cover model SNOWPACK, and three diagnostic models of snow disasters. In this paper, the performance of NHM and SNOWPACK is investigated in the Niigata area, selecting the events of December 2005 as the case for the experiment. NHM reproduces precipitation events but tends to underestimate the amount of precipitation. This is because the positions of snow clouds are not precisely reproduced in the simulation. The predicted snow depth by SNOWPACK is in a good agreement with the observation. However, the snow type has a large prediction error. A snow cover forecasting experiment with a combination of NHM and SNOWPACK reproduces the observed snow depth and snow density well for 24 hours with slight prediction errors. The snow depth tends to be underestimated when NHM does not suitably reproduce the precipitation. To apply the system to a wet snow region, such as Niigata, accurate predictions of temperature and precipitation are needed.
