Analysis of the Regional Variability of Surface Air Temperature in the Northern Hemisphere Using Quantile Regression Based on the ERA5 Reanalysis and SLAV072L96 and INM-CM6 Hydrodynamic Models

The paper analyzes the regional long-term temperature variability in the Northern Hemisphere using quantile regression. This method allows assessing not only changes in average values but also those in the entire distribution, including extreme low and high temperatures. The study is based on ERA5 reanalysis data and historical forecasts of the SLAV072L96 and INM-CM6 hydrodynamic models. It is found that warming is distributed unevenly: in some regions (for example, in southern European Russia), extreme low temperatures are increasing most significantly, reducing overall climate extremity, while in others (for example, in northern Asia), the most significant increase is associated with extreme high temperatures rise, increasing climate contrasts and risks. A comparison with reanalysis data showed that both models systematically underestimate warming in key Arctic regions. The INM-CM6 has demonstrated a higher quality of simulating changes in seasonal averages, but both models struggle with simulating extreme high/low temperatures in summer/winter. The results of the study can be used in forecasting activities by the Hydrometeorological Research Center of the Russian Federation/North Eurasia Climate Center.