Simulation of spring snowmelt runoff by considering micro-topography and phase changes in soil layer

Abstract. The NICE model was extended to include the effect of the micro-topography in slope and shading characteristics and the phase changes in soil moisture on snow/frost depths and snowmelt runoff by combining the land-surface, the multi-layer runoff, and the groundwater flow models (NICE-SNOW). The model was applied to the upstream regions of shrinking Kushiro Mire in the invasion of alder, where the spring runoff affects greatly the annual sediment and nutrient transports because the spring flood continues in longer time than that in typhoon seasons. The simulation reproduced excellently the observed values of annual river discharge including snowmelt runoff with the greater time-to-peak of runoff than in snow-free period, in addition to snow depth, frost depth, soil temperature, soil moisture, and groundwater level, by conducting the quantitative assessment of goodness-of-fit and parameter sensitivity analysis. We quantified that the mechanism of spring snowmelt runoff is related to changes in micro-topography, soil structure, soil temperature, soil moisture, and groundwater flow. The model shows that the local effect of snow depth and the frost depth disappears in the snowmelt runoff discharge of catchment in the same way as some previous researches though they are very important as water resources of catchment. After the frozen soil restricts the infiltration in the coldest part of winter, the thawed soil increases the pore size in the early spring. The NICE-SNOW could explain the snowmelt flood continues a longer time than that in the typhoon period because some part of meltwater flows as an intermediate flow in the partially-thawed hillslope soil layer. This is also related to the simulation result that more than half of total soil moisture stays unfrozen at some places even in winter periods, which indicates that there is a high degree of spatial heterogeneity of frozen ground.