چكيده لاتين
Snow cover is the largest single component of the cryosphere in terms of spatial extent and it affects the global surface energy balance due to its high albedo and the energy allocation involved in the snow melting. The amount of water stored in a snowpack, which is obtained as snow water equivalent (SWE), is of great importance for hydrological applications, numerical weather predictions, climate change research and land surface process simulations. Since passive microwave sensors can contribute to obtaining information about snowpack volume, microwave brightness temperatures (BT) have long been used to assess spatiotemporal variations in SWE. However, the coarse spatial resolution (typically at 25 km from passive microwave remote sensing images) of the existing SWE products cannot meet the needs of explicit hydrological modeling, and thus, the low spatial resolution of existing SWE products (i.e., the coarse resolution of AMSR2 based products) leads to less satisfactory results, especially in regions with complex terrain conditions, strong seasonal transitions and, great spatiotemporal heterogeneity. To improve the inversion accuracy and spatial resolution of BT difference (BTD) SWE in Zayandehroud River basin, a multifactor SWE downscaling model was developed by combining PMW SWE data from the AMSR2 sensor, optical snow cover extent data, and surface environmental parameters to produce fine scale (1 km × 1km) and high precision SWE data.Validations at 14 ground meteorological stations show that the developed model greatly improved the spatial resolution and inversion accuracy of the raw BTD SWE; its root-mean-square error (RMSE) reduced from 40.05 mm of the raw BTD SWE to 23.83 mm, and the correlation coefficient (R) increased from 0.5 to 0.85. Among the machine learning algorithms, the Random Forest (RF) algorithm has shown the best performance in downscaling and improving the estimation of SWE. Compared with the existing downscaling methods, the proposed model presented the best performance.In addition, in this research, a linear and exponential regression was used to estimate snow equivalent water using snow indices. NDSI index in exponential regression with RMSE, 38.24 mm and R 0.57 showed the best result in the estimation of SWE in the spatial resolution of 1 km × 1km. But in general, according to the results obtained from snow indices, saturation occurs faster in linear and power regression models than the BTD SWE resulting from PMW, but it can be used to estimate SWE in areas with low snow depth with reasonable accuracy. used snow indices to estimateSWE, even though they were not originally designed for this purpose
