European Climatology of Day-to-Day Surface Air Temperature Difference in Multiple Data Sets

Abstract

Substantial progress has been made in analyses of climate means and extremes, while much less has been done in understanding short-term (intraseasonal and synoptic-scale) variability. One particular aspect of short-term atmospheric variability is day-to-day temperature difference (DTD). Large DTDs negatively affect human health and also impact animals and plants, hence constituting one of the many weather-related risks to society. The main objectives of this study are to provide the climatology of statistical characteristics (distribution width, skewness and kurtosis) of DTD in Europe and compare DTD climatologies between five data sets (station observations, ECA&D; gridded observations, E-OBS; and reanalyses, NCEP/NCAR, JRA-55 and 20CR). DTD is largest in winter and smallest in summer; it tends to grow from the coast towards inland. Skewness is mostly negative in summer, whereas positive skewness prevails in winter with the exception of the British Isles and southeastern Europe. DTD distributions are leptokurtic (i.e., having kurtosis higher than Gaussian), with lower values of kurtosis prevailing in western Europe. Comparisons among data sets reveal their specific deficiencies, such as the existence of outlying unrealistic temperature values in station data that were not detected by quality check, the underestimation of the magnitude of DTD, particularly by the 20CR reanalysis, and overestimated skewness in the British Isles in winter by the reanalyses. The data set differing most from the others is the 20CR reanalysis, which exhibits substantial biases in skewness and kurtosis over large parts of Europe. On the other hand, the reanalysis most consistent with observations is JRA-55.