Fig. 2
From: Increasing occurrence of cold and warm extremes during the recent global warming slowdown

Linear regressions of wintertime cold and summertime warm extremes. Time series of wintertime cold (a) and summertime warm (b) extreme temperature occurrences (d season−1) over Northern Hemisphere land from 1951–2014 in observations (dark blue and red) and for linear regression models with four predictors (light blue and orange). Shading indicates the 95% confidence interval for the regression. c, d Contribution of each individual predictor for the frequency of cold extreme (c) and warm extremes (d), calculated by multiplying the predictors by their corresponding regression coefficients. Red lines indicate linear trend lines for the hiatus period (2002–2014). Major volcanic eruptions are labeled in the AOD contribution plots. The top regressions in (c) and (d) indicate the influence of the z500 and SST extremes patterns, respectively