Abstract
While long-term changes in measures of central tendency of climate elements, that is, mean temperature, are well acknowledged, studies of trends in measures of their variability are much less common. This is despite the fact that trends in variability can have higher effect on climate extremes than trends in mean.
Here, four measures of intraseasonal variability are examined: (a) standard deviation of mean daily temperature, (b) the range between the 90th and 10th quantile of mean daily temperature, (c) mean absolute value of day-to-day temperature change, and (d) one-day lagged temporal autocorrelation. ECA&D daily data from 168 stations and linear regression method are utilized to calculate trends of these characteristics in period from 1961 to 2018.
Significant trends (positive and negative) are revealed with substantial differences between seasons, regions and measures. The most considerable decreases in temperature variability were recorded in winter, for temporal autocorrelation in eastern Europe and for variance-based measures in northern Europe.
For example, the standard deviation has decreased by more than 10% in the Arctic Ocean. This can indicate a decrease in the frequency of cold extremes in Scandinavia.
On the contrary, increasing persistence may suggest a greater likelihood of cold extremes in the East European Plain. Increases in variability prevail only in summer, but not for all measures and not as clearly as decreases in winter.
Trends in temporal autocorrelation and day-to-day change appear to be sensitive to data issues, such as inhomogeneities and changes in observational procedures.