Assessing the Reproduction Quality of Meteorological Characteristics by Several Atmospheric Reanalysis Models on the Territory of Crimean Peninsula

The diversity of natural conditions of the Crimean Peninsula determines different regimes of the main meteorological characteristics that determine the water availability for the territory. The estimation of the spatiotemporal heterogeneity of these characteristics and the solution of the problem of gaps in the ground-based observation data can be based on the results of calculations by general circulation models of the Earth’s atmosphere with assimilation of ground-based observation data, also known as atmospheric reanalysis. Estimates of the quality of reproduction of the surface air temperature and the total precipitation by atmospheric reanalysis models EWEMBI, ERA5-Land, and MSWEP are given and compared with data from ground-based meteorological observations. The main characteristics of the data sets used (both observational and calculated), the main verification methods, the results of estimates and the conclusions regarding the applicability of the data used in simulation problems are given. The mean errors of the models in air temperature and the amount of precipitation over various averaging periods (day, month, year) are given. Thus, the mean coefficients of correlation over different averaging periods vary within 0.74–0.97 for temperature and 0.52–0.79 for precipitation. The results show that all model reproduce the values of the temperature and total precipitation over different averaging periods with an acceptable accuracy; however, all of them show a tendency toward underestimation of the daily sums of precipitation along with an overestimation of the number of days with precipitation.