Document Type : Research Article
Authors
1
M.Sc. student/Department of Space physics/ Institute of Geophysics/University of Tehran/Tehran/ Iran
2
Associate professor/Department of Space physics/Institute of Geophysics/University of Tehran
3
Associate professor/Department of Space physics/ Institute of Geophysics/University of Tehran
Abstract
Snow cover is the most widely distributed and most dynamic component of the cryosphere, and its significant seasonal and annual variability have notable influences on the global water circulation and surface energy balance. Ground observations represent the most direct and reliable snow depth data source. Direct measurements are generally not used to monitor snow depth at hemispheric scales due to the inadequate number of measurements and the sparse distribution of measurements in remote regions, mountains, and high elevation areas. Instead, the method examined in this study is the use of reanalysis data. The aim of the current study is to evaluate the performance of the ERA5 and MERRA2 reanalysis datasets in estimating snow depth in Northwestern Iran. The evaluation has been done using snow depth data from meteorological stations throughout the study area.
The area examined in this study includes the provinces of Ardabil, East Azerbaijan and West Azerbaijan, which are located in the range of 44 to 51 degrees in longitude and 35.7 to 40 degrees in latitude. This area was selected for two primary reasons: firstly, the mountainous terrain of the region that results in a higher amount of snowfall, and secondly, there is an abundance of observational data available for estimation purposes. To conduct this study, monthly mean snow depth data from the ERA5 and MERRA2 reanalysis databases for the period 1981 to 2020 were used. The spatial resolution of the ERA5 data is 0.25 x 0.25 degrees. The MERRA2 data, on the other hand, has a spatial resolution of 0.625 x 0.5. Since the data from MERRA2 and ERA5 have different spatial resolutions, to compare these two databases, a regridding method is used to equalize their spatial resolutions. The ERA5 data is regridded to MERRA2 spatial resolution using nearest-neighbor interpolation.
The results of this study showed that by examining the spatial and temporal distribution, the ERA5 database underestimates the mean value of snow depth over the entire study area. On the other hand, the MERRA2 database overestimated the average snow depth at most stations and had more errors than the ERA5 database. According to the ERA5 reanalysis database, the highest snow depth over a 40-year period has occured in February, while for the MERRA2 database, it has occurred in January. In the ERA5 database, the estimated snow depth values were consistent with the observations only in February, while in other months there was a discrepancy between the snow depth estimated by ERA5 and the observations. As the station's height and latitude increase, the error increases and the underestimation of snow depth from ERA5 also increases, but in the MERRA2 dataset, there is no significant relationship between the height of the station and latitude with bias. The analysis of the results of the present study shows that the ERA5 database is more accurate than MERRA2 for studying the spatial and temporal distribution of snow depth in the northwestern region of Iran. Of course, in the areas with high snow cover, the MERRA2 data estimates values closer to observations.
Keywords
Main Subjects
)