Detailed long-term hydrometeorological hindcast for Russian Arctic was created using regional nonhydrostatic atmospheric model COSMO-CLM ver. 5.05 for 1980–2016 with ~12 km grid size and shared online partially on the figshare service (https://figshare.com/collections/Arctic_COSMO-CLM_reanalysis_all_years/5186714, [1]). The hindcast includes about a hundred hydrometeorological variables at both surface and 50 model levels, and covers the Barents, Kara and Laptev Seas [2].

Surface temperature and wind speed reproduction by the COSMO-CLM Russian Arctic hindcast was evaluated in this study according to 145 stations and satellite data.

The COSMO-CLM Russian Arctic hindcast is generally successful in reproducing the thermal regime of region according to monthly mean values, while a significant part of the errors is explained by a large distance of the nearest model grid from the station, inaccuracies in the coastline or surface altitude description. On the average, the model underestimates the mean monthly temperatures within one degree, the errors are in the range from +2 to -4 ^oC. The patterns of the temperature 5-% quantiles differences are within +-1 ^oC, but there are significant positive errors at more continental stations of the region (from +1 to +2 ^oC), as well at highlands stations (up to +2 ^oC). The ETCDDI indices FD and ID pattern according to the hindcast is climatologically adequate, however the model overestimates the indices in the western and maritime parts of the region, and underestimates them in severe climate continental regions.

The comparison with the station data showed that the monthly average wind speed is well reproduced by the COSMO-CLM hindcast, while the errors relate mainly to cases when the wind speed is overestimated by the model data up to 2 m/s. Unlike the average wind speed, the extreme values (for 95% percentiles) according to the hindcast are underestimated compared to the stations data with up to -5 m/s.

Model data were evaluated according to the SAR high-resolution satellite images including the FSS score. Its evaluation for specific extreme wind speeds cases near the Novaya Zemlya Island shown the hindcast could capture the spatial structure of wind speeds higher than 10 m/s and partially 15 m/s, however could not reproduce 20 m/s. This revealed the model capability to reproduce β-mesoscale processes, unlike the γ-scale processes. An analysis of the difference between the quantiles of model and satellite data showed that the more extreme is the observed wind speed, the greater is the probability of underestimation by the model.

Future perspectives include the hindcast prolongation to 2019, sharing more data online; focus on extreme and severe events statistics assessment (downslope windstorms, polar lows, MCAO climatologies using satellite data); quality estimation based on other datasets (e.g., ERA5, CARRA, satellites climatology).

References:

1. Data from the COSMO-CLM Russian Arctic Hindcast archive, figshare repository https://figshare.com/collections/Arctic_COSMO-CLM_reanalysis_all_years/5186714; https://doi.org/10.6084/m9.figshare.c.5186714

2. Platonov V, Varentsov M. Introducing a New Detailed Long-Term COSMO-CLM Hindcast for the Russian Arctic and the First Results of Its Evaluation. 2021 Atmos. 12(3) 350 http://dx.doi.org/10.3390/atmos12030350