Dokus A. Long-term forecasting of spring flood characteristics in the Pivdenny Buh river basin

In the thesis work the regional methodology of long-term forecasts of characteristics of spring flood of the rivers in the Pivdenny Buh river basin is developed. The methodology based on the preliminary typification of floods by water content in choosing the main factors of spring runoff formation, establishing regional parameters (within the basin zoning by a set of hydrometeorological factors) for long-term forecasting of runoff layers and maximum water consumption, determining their probabilistic values over many years. In the dissertation research using the methods of factor and cluster analyzes; hydrological zoning was performed under the conditions of spring flood of rivers in the basin of the Pivdenny Buh, which allowed to distinguish two hydrological regions (region I and region II with subregions IIa, IIb, IIc). The analyzed natural conditions have shown that within limits of the carried out hydrological zoning these regions have specific features of spring floods formation. Data from long-term meteorological, agrometeorological and hydrological observations in the Pivdenny Buh river basin (as of 2015) were used as input information for the development of the methodology, using data from observations in the Velykyi Kuyalnyk and Tiligul rivers. To typify the water content of spring floods, the method of discriminant analysis by the formed vector-predictor was used, to which the main factors of spring runoff formation of the considered territory were selected (by the method of factor analysis) – maximum water reserves in snow cover, precipitation and maximum soil freezing depth. Three groups of water content (high, medium and low) of floods were distinguished and the coefficients of discriminant equations for reference hydrological station within the homogeneous areas (subdistricts) under the conditions of spring floods formation were obtained. Regional dependences (in the form of modular coefficients) for homogeneous regions (subregions) of spring flood formation have been established for spring floods with different types of watercontent. This made it possible to predict the runoff layers and maximum water consumption of the spring flood for other rivers within the zoning of the Pivdenny Buh river basin, and to consider the developed methodology as regional. Representation of predicted values and their predicted sufficiency in a long-term context, established by the three-parameter gamma distribution of S.M. Kritsky and M.F. Menkel (taking into account the average-region ratios of coefficients of variation and asymmetry determined by the author) is carried out in cartographic form, which allows spatial monitoring of the size of the expected spring floods. The regional methodology of long-term forecasts of the characteristics of spring floods in the Pivdenny Buh river basin is also justified for the hydrologically unexplored rivers of the study area. To determine the average long-term maximum water consumption of spring floods of hydrologically unexplored rivers, the operator model of maximum runoff formation using meteorological values at its inlet (average long-term maximum snow reserves and precipitation during spring flooding) was used. Within the operator model all input parameters of the calculation methodology are defined and generalized. Parameters that are not measured by the hydrometeorological network were obtained numerically. The average long-term modules of the maximum water consumption of spring flood calculated by the method conforming with the initial data with an error of their determination ± 16.8%. To assess the impact of future climate change on the water regime of rivers in the spring flood period, climate change coefficients were calculated (according to IPCC climate models (CORDEX project) of RCP 4.5 scenario for the period 2021-2050). The results showed that in the Pivdenny Buh basin in the region I in the period 2021-2050 was expected to reduce the long-term average characteristics of runoff (runoff layers and maximum water flow) of spring floods on average by 23% ( kch = 0.77), and in region II: sub-district IIa ‒ by 37% ( kch = 0.63), sub-region IIб ‒ by 51% ( kch = 0.49), sub-region IIв ‒ by 75% ( kch = 0.25). The layers of spring water runoff modeled according to climatic models in the period of 2021-2050 were expected to be lower than their long-term average value (established taking into account), which was 23 mm. The performed assessment of the quality of the regional methodology of long-term forecasts (for the period 1966-2015) and its verification on the data of independent years (for the period 2016-2018) showed that the developed methodology mainly belongs to the category "good" and "satisfactory", which allows to recommend it for practical application in activity of hydroprognostic divisions, including the rivers of the Pivdenny Buh basin not studied hydrologically.