A database of monthly mean air temperature data collected in Ukraine has been created. The database contains records of 514 stations, which were working at one time or another during the period 1812-2014. Metadata have also been collected using the available descriptions of the stations' histories. Regular observations at a sufficiently large number of stations contained in the database began in the early 1880s. However, due to the disruptions that occurred during World Wars I and II and resulted in a lot of missing data, a reliable homogenized data set can be obtained only for the period 1946-2014. For this period, we have homogenized the data from 178 stations by means of the HOMER software. The total number of breaks that we have identified is 287. However, only approximately 31% of them can be explained by the station relocations or other reasons. The shift magnitudes in annual series do not exceed ±1 ∘ C with the small negative mean value. Therefore, on average, the inhomogeneities of the time series have an effect of increasing the past temperature slightly over the territory of Ukraine and, consequently, decrease the rate of the temperature rise. The trend analysis of the raw and homogenized time series confirms this conclusion. The homogenized data show a more realistic pattern of the spatial distribution of the temperature tendencies, in contrast to many artificial spots of higher or lower trend values in the raw data. After the homogenization, the areas with insignificant trends were considerably reduced.

Digital databases of monthly means of daily extreme, minimum (Tn) and maximum (Tx) air temperature collected in Ukraine were created. The databases contain temperature records of 371/340 (Tn/Tx) climatological stations which recorded data on the territory of Ukraine anytime during 1881-2015. The significant part of the data sets (~48/47%) was obtained after conducting a data rescue process, recovering historical paper records. Due to a great deal of missing data during World War I and II, homogenization was possible to perform only for 178 time series covering the period of 1946-2015. A homogenization procedure, conducted by means of the HOMER software, revealed 323 break points for minimum temperature (~1.8 breaks per station) and 310 for maximum temperature (~1.7 breaks per station). Approximately 37/33% (Tn/Tx) of the breaks can be explained by metadata. Shift amplitudes in Tn time series ranged between −1.18 and 1.71 C, while for maximum temperature the range was slightly less from −0.99 to 1.15 C. For both temperatures, mean values of shift amplitudes were near zero (−0.04/−0.05 C). To validate the homogenization results, several statistical procedures were performed to compare inhomogeneous and homogeneous time series. The conducted statistical calculations provide the proofs that inhomogeneity of the Ukrainian extreme air temperature data should be taken into account when analysing regional climate change and variability. K E Y W O R D S extreme (minimum/maximum) air temperature, homogenization of time series, HOMER software, Ukraine

The widely used Global Historical Climatology Network (GHCN) monthly temperature dataset is available in two formats—non-homogenized and homogenized. Since 2011, this homogenized dataset has been updated almost daily by applying the “Pairwise Homogenization Algorithm” (PHA) to the non-homogenized datasets. Previous studies found that the PHA can perform well at correcting synthetic time series when certain artificial biases are introduced. However, its performance with real world data has been less well studied. Therefore, the homogenized GHCN datasets (Version 3 and 4) were downloaded almost daily over a 10-year period (2011-2021) yielding 3689 different updates to the datasets. The different breakpoints identified were analyzed for a set of stations from 24 European countries for which station history metadata were available. A remarkable inconsistency in the identified breakpoints (and hence adjustments applied) was revealed. Of the adjustments applied for GHCN Version 4, 64% (61% for Version 3) were identified on less than 25% of runs, while only 16% of the adjustments (21% for Version 3) were identified consistently for more than 75% of the runs. The consistency of PHA adjustments improved when the breakpoints corresponded to documented station history metadata events. However, only 19% of the breakpoints (18% for Version 3) were associated with a documented event within 1 year, and 67% (69% for Version 3) were not associated with any documented event. Therefore, while the PHA remains a useful tool in the community’s homogenization toolbox, many of the PHA adjustments applied to the homogenized GHCN dataset may have been spurious. Using station metadata to assess the reliability of PHA adjustments might potentially help to identify some of these spurious adjustments.

This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.