چكيده به لاتين
The variable of snow and soil moisture in hydrological models is one of the hydrological variables with significant effects on the water cycle and the need to use an efficient method to estimate the resulting runoff is raised. In this study, in order to estimate the jointed state variables and parameters, from the assimilation of snow cover observations, soil moisture and runoff using particle filter, the catchment basin limited to the upstream of the Safakhaneh hydrometery station on Zarrineh River has been modeled with regard to snow considerations. In this study, by considering a number of sensitive parameters and also considering the state variable snow water equivalent and soil surface moisture in SWAT model and by assimilating satellite observations of soil moisture by AMSR-E sensor and snow cover by MODIS sensor and groundwater observations by particle filter Three separate scenarios and the output runoff of the model are evaluated by the method of estimating the connected state-parameter variable. For this purpose, the necessary changes have been made to edit and extract the values of the state variables in the SWAT reference codes. The results show that the assimilation of observations of these variables in the scenarios in which the observations of the variable itself are assimilated, improves the prediction accuracy of the same variable and its RMSE value is reduced compared to the cases where the observations of other variables are assimilated. For example, in this study, with the assimilation of snow cover observations compared to the assimilation mode of runoff observations, the RMSE value of snow cover estimation has decreased from 0.15 to 0.07. Also, assimilation of runoff observations and considering snow water equivalent and soil moisture as variables, compared to calibration with SUFI-2 algorithm, reduced the RMSE value from 4.12 to 0.93 in runoff estimation and the results improved the estimation accuracy using particle filter. it shows that the results suggest the idea that hydrological models go in the direction that each variable is calibrated according to the observations of the same variable, and promise that by considering both absorbable observations, in hydrological models simulating a wetland, including Groundwater, vegetation, evapotranspiration and calibration of these variables in hydrological models in accordance with the unique observations of the same variable, the end-of-life problem will be significantly solved. Other satellite observations from various sensors, such as the AMSR-E snow depth as well as snow temperature, can also be captured and research will be ongoing.
