Estimation of surface air temperature trends over the Russian Federation territory using the quantile regression method

Abstract

The results are presented of the estimation of surface air temperature variations in different climatically quasi-homogeneous regions of Russia using the nonparametric method of regression analysis (quantile regression). Daily observation records from 517 weather stations were used. The quantile regression technique used for analyzing the trends in long-term series allows obtaining information on trends for the whole range of quantile values from 0 to 1 of dependent variable distributions. Seasonal and regional features of daily minimum, mean, and maximum air temperature trends are considered in a wide range of quantile values. The proposed method that generalizes long-term trends obt ained for groups of stations by quantile regression, is applied to quasi-homogeneous climate regions identified on the territory of Russia.