On the issue of forecasting catastrophic floods in the territory of Crimea

The catastrophic situations of recent years – in June 2021 in the Yalta region and in January 2024 in Sevastopol – associated with heavy precipitation, rising water levels in rivers and the formation of mudflows, once again demonstrated the need for an early forecast of events with extreme precipitation in Crimea for a timely response and minimization of economic losses. The region of mountainous Crimea with its complex terrain and large slopes is especially susceptible to the emergence of dangerous situations after heavy (often multi-day) rains. Based on daily precipitation data from the Ai-Petri meteorological station, cases with a precipitation amount of ≥40 mm over three consecutive days were calculated and analyzed. Such conditions were used in the analysis as a threshold for extreme precipitation leading to erosion of river beds in mountainous Crimea and the formation of mudflows. The situation of a catastrophic flood on the river Chernaya in January 2024 is considered, caused by extreme precipitation that fell over three days in the Sevastopol region. Then, for such a situation, a study was conducted on the possibility of forecasting them with a lead time of 3 months using the developed artificial neural network model. The results showed satisfactory quality of the developed neural network for forecasting with a lead time of 3 months of 2–3-day extreme precipitation that intensifies erosion processes in the mountainous Crimea.