POSSIBILITIES OF THE LONG_TERM FORECASTING OF THE
APPEARANCE DATES OF ICE PHENOMENA AT THE DNIPRO CASCADE
RESERVOIRS BY HEAT FLOW COEFFICIENTS

The development of reliable and accurate long-term ice regime forecasts is one of the most difficult tasks in hydrological practice, which has not yet been satisfactorily solved. Reliable and lead time forecasts of ice phenomena appearance and the freeze-up appearance on reservoirs are essential for the rational use of water resources and for establishing reservoir operation regimes that take into account the requirements and interests of various economic sectors: hydropower, navigation, fisheries, communal services, etc. Currently, the Ukrainian Hydrometeorological Service has no methods for long-term forecasting of ice phenomena on the Dnipro River reservoirs, and those that are officially adopted and used were developed for long-term forecasts of phenomena only on rivers. Therefore, it is extremely relevant to develop methods for long-term forecasting of ice phenomena on the Dnipro reservoirs.

The purpose of research is to find correlation links between quantitative indicators of atmospheric processes, which expressed in terms of heat flux coefficient, and the dates of ice phenomena in the reservoirs of the Dnipro cascade and to analyze the results.

The paper also analyzes current climate change in the study region, since the ice regime is directly dependent on temperature. Changes in air temperature over the past decades have led to a reduction in the total number of days with ice phenomena on rivers, an increase in water temperature in the autumn and winter, a later of ice phenomena appearance, unstable ice cover, and difficulties in determining the appearance of ice regime phases.

Taking into account the change in the climate normal in recent years, the data from 2000 to 2020 were used for the study, since it is from this period intensification of the climate warming process is observed. A representative hydrological gauging station was selected for each of the 6 reservoirs of the Dnipro River and dependencies of ice phenomena on the thermal coefficient, which is defined as the ratio of the number of periods with cold air flows to the number of periods with warm air flows during natural synoptic periods (September 01 - October 10), were obtained. This takes into account the direction of movement of air masses and their intensity.

The results show that the correlation coefficients of the forecast dependencies range from 0.0005 to 0.449, and forecast error tolerance was 40-57%. This result indicates that forecasting dependencies cannot provide reliable forecasts. Therefore, the methodological approach used in the study, which was developed in the second half of the twentieth century under conditions of a more stable air temperature regime, cannot be used in modern climatic conditions. To obtain more reliable prognostic dependencies for long-term forecasting of the ice-appearance date, it is necessary to expand the search for predictors that characterize atmospheric processes both in space and time.

Keywords: Dnipro reservoirs; ice phenomena appearance; methods of long-term forecasting; prognostic dependencies.