ИСТИНА

The Arctic is the region most sensitive to climate change on the globe, this applies both to physical feedbacks in the climate system and to ecological systems. However, the spatial distribution of climate changes in the region is unclear, and there are different estimates of trends in changes in certain meteorological parameters in different water areas and coastal zones of the Russian and foreign Arctic regions. Taking into account the growing number of dangerous phenomena and the prospects for the development of the Arctic coast and the Northern Sea Route, the task of providing the region with detailed hydrometeorological and climate information with a horizontal resolution of several kilometers becomes especially necessary. There are poor observational and satellite data for a long period in the Arctic, and the existing model datasets and reanalyses has quite coarse resolution. Therefore, it is important to use high-resolution modeling data to get appropriate climate change assessments in the region. This work uses a new long-term model COSMO-CLM Russian Arctic hindcast hydrometeorological data covering the period 1980-2016. The model domain is represented by a grid resolution of ≈12 km (≈0.108°) and covers most of the Russian Arctic, including the Barents, Kara, and Laptev Seas. All long-term calculations were carried out at the supercomputer "Lomonosov-2”. The objective of this work is to evaluate for the first time the quality of surface temperature reproduction by the new high-resolution COSMO-CLM Russian Arctic hindcast. To achieve this goal, the following tasks will be performed: - creation of a database of average monthly temperatures of weather stations in the studied region for the period 1980-2016; - statistical evaluation of the model's reproduction of the thermal regime of the Russian Arctic on a multiyear, seasonal and monthly scales; - assessment of the quality of the archive based on a comparison with observational data of both average and extreme characteristics; - assessment of detailed regional changes in the thermal regime in the Russian Arctic based on the COSMO-CLM Russian Arctic hindcast. 145 weather stations were selected with a small number of gaps in the observations for the period 1980-2016, 65 European and 80 Russian Arctic stations, which geographically fit the study area of COSMO-CLM modeling. At the first stage of the study, the data on the 2-m temperature were averaged to monthly values. Then, the nearest model grid points were selected for each weather station according to its coordinates, to perform comparisons. The study of the distribution of differences between the average monthly values of the model temperature and the station data showed that the distribution function is close to normal. This indicates the uniformity of errors sample. However, the function is biased a few towards negative values, which indicates a slight underestimation of the real values by the model, the error distribution mode is -0.5 0С. The distributions of the 1-% and 5-% quantiles differences between the model and the actual air temperatures show accuracy of the model reproduction of extreme air temperatures. In general, the model copes with this task, but significant errors are noticeable (in the ranges from -30С to -50С) at more continental stations in the region, as well as at high-altitude stations. The calculated statistics of the COSMO-CLM model archive indicate that the model reproduces the surface temperature with quiet high quality. In the future work, the quality of the archive will be evaluated in more detail, based on comparison with observational data of both mean and extreme characteristics for different seasons of the year, months, as well as on a more detailed timescale. Based on the obtained results, an assessment of regional changes in the thermal regime over the research period will be carried out.