In Central Asia, since most of water resource is supplied in mountainous region in winter as snowfall, snow process is key factor to reproduce seasonal water resources. However, in the previous works with land surface model, river discharge was earlier estimated. Therefore in this research, snow process in land surface model was validated in Zarafshan river basin which is the third biggest river basin in Aral Sea Basin. Upstream region is located in Pamir mountains in which altitude is more than 5000m with several glaciers.
In this research, satellite analysis was firstly conducted to validate land surface results with comparing snow cover estimated by MOD10 product in MODIS. And it was found that snow melt is overestimated in high altitude mesh where snow cover disappeared immediately in spring. In this basin, even though there is no big dam, peak discharge is in summer due to snow melt, however, analyzed peak is on April and discharge in summer is much lower. Therefore in this research, sub-grid scale snow existence and altitude dependency of meteorological forcing were studied.
To clarify sub-grid scale impacts on snow melting process, four analyses at deferent spatial resolutions which are 20km, 10km, 5km and 1km were compared and it was found that glacier melting is much better in higher resolution since glaciers exist only in limited region in high mountain which spatial scale is much smaller than mesh size. To consider this mixture of several altitude bands in mesh for lower resolution analysis, altitude mosaic scheme was developed which is applied instead of land use mosaic when the mesh is higher than 1000m. In this scheme, vertical water balance is analyzed in each altitude band, and accumulated with weight of each altitude bands’ area. With this scheme, 20km resolution analysis was not different much from 1km resolution analysis.
Secondary, altitude dependency of meteorological forcing for land surface analysis was studied. In addition to air temperature and air pressure, downward short wave radiation and air humidity was changed in height. As for long wave radiation, its altitude dependency was calculated by differential form of Stefan-Boltzmann law and one of specific humidity was modified to assume that relative humidity is stable in height. As a result, by considering altitude dependency especially long wave radiation, analyzed peak discharge was improved.
Through these trials, land surface model was improved, but there are still early trends on snow. In the future study, short wave radiation will be modified depending on sun duration time, slope rate and its direction. After accuracy of snow melt results improved, climate change impacts to monthly river discharge will be analyzed.