Semi-empirical model of the spatiotemporal surface temperature distribution on the plain part of Ukraine

The spatial variation of temperature is found to depend linearly on climate continentality, morphology of the relief, the position of the site with respect to seas, in addition to the usual elevation, latitude and longitude predictors. There are other factors that can have an additional significant influence: big bodies of water, terrain attributes relief, atmospheric factors (local circulation), configuration and aspect of coasts and vegetation. Therefore, these multifactorial influences form the climatic field of temperature. In this study, the regional semi—empirical model of the spatiotemporal distribution of the average annual and monthly temperature for the plain part of Ukraine on the basis of the methodology for assessing the influence of height above sea level and geographic coordinates is proposed. Based on the method for determining the altitudinal, latitudinal, and longitudinal gradients of meteorological parameters, we calculated these gradients for annual and monthly air surface temperature for the periods 1961—1990 and 1991—2020. Thus, on the plain part of Ukraine, the annual surface air temperature decreases by an average on 0.60—0.63 °C with a shift of 100 m height above sea level, on 0.51—0.55 °C with a shift of one latitude degree to the north, on 0.067—0.071 °C with a shift of one longitude degree to the east. Also, the variations of these annual mean temperature gradients from year to year over the period 1991—2020 are characteristic. The seasonal variation of gradients has a pronounced non—monotonic character: highest values of altitudinal gradientare typical for July—August (from –0.63 to –0.73 °C per 100 m), and the lowest values — for April—May (from –0.45 to –0.55 °C per 100 m); highest values of latitudinal gradient are typical for August—September (from –0.60 to –0.70 °С per 1 °N), and the lowest values — for April—May (from –0.20 to –0.35 °С per 1° N); the longitudinal gradients have positive values in June—August (0.074—0.128 °C per 1° E), and negative values in November—March (from –0.228 to –0.154 °C per 1° E). We found that the altitudinal and latitudinal gradients of temperature have the most spatiotemporal variability and the longitudinal gradient has the smallest one. Greatest variabilities of temperature gradient values are typical for February—March and July—September, and the least variability — for April—May. The analysis of the dynamics of gradient changes in the period 1991—2020 compared to the period 1961—1991 showed the following: the altitudinal gradientvalues increased by 8—13 %. in January and March—May; the latitudinal gradient values increased by ~30 % in December—February and decreased by ~20 % in May—August. The proposed semi—empirical model contains a coefficient that takes into account influence of additional effects associated with pronounced orographic and other terrain features. This study presents the numerical values of this coefficient for some specific microclimate regions of the plain part of Ukraine. The model estimates of thirty—year monthly mean temperature in Ukraine for the periods 1961—1990 and 1991—2020was calculated. A comparison of the model estimates of of the average annual and monthly temperature for 72 meteostations in Ukraine with their actual values showed a statistically significant correlation (the reliability of the linear approximation is 0.89—0.98). Thus, the presented design of the semi-empirical model makes it possible to quite well restore the annual and monthly temperature on the territory of Ukraine