Analysis on Snow Accumulation and Groundwater Resources by a Distributed Hydrological Model

Background: Japanese water resources rely on snow water in mountainous regions and groundwater contributing to base flow in rivers. Especially paddy fields use snowmelt water for rice planting in spring season. Although a lot of studies have been made on storage of snow water resources or groundwater, little attention has been given to the point of the interaction of the both storages in a mountainous area. Therefore we applied a distributed model in a Japanese heavy snow region and evaluated the interaction in a view of water resources.

Study Area: The study area is the Yoneshiro River Basin in the northern Japan, which has 136km length and 4,100km2 catchment area. This area is famous for the heavy snow region from humid air from Japan Sea to the central mountain range in Japan.

Distributed hydrological model: The hydrological model is composed of a SWE model based on water balance in snowpack involving a consolidation process and an assimilation method by observed data, a snowmelt model based on energy balance model on the snow surface and below the snow bottom, and runoff model based on kinematic wave theory on channels and storage function theory on hill slopes. Almost meteorological data for input to the simulation are available at 3 to 10 stations in the basin but vapor pressure is only available at Akita city, out of the basin.

Results and discussion: The model applied to the Yoneshiro River Basin from September, 2015 to May, 2016. The Nash Efficiency Coefficients are 0.70 for discharge at Futatsui point, the most downstream observing station, and 0.93 for snow depth at Kazuno point, the second highest station. Main errors for both data occur in the end of snowmelt season.

Storages of snow and groundwater simulated in time series, show that the snow storage changes greatly larger than groundwater one, that snow storage trend has negative correlation to groundwater trend, and that the maximum groundwater storage arises in the beginning of the spring season. Also the simulation and single regression analysis are carried out for daily variation correlation for both storages. The analysis indicates the strong negative correlation on more than 5 days period and time difference about 1 week for groundwater level changing after snowmelt.