FORECASTING EMERGENCY SITUATIONS AND EVENTS DURING MARTIAL LAW

Abstract

Purpose. The processing of statistical data and forecasting the risks of military emergencies and their consequences (using the Poltava region as an example) to increase the efficiency of civil protection forces' preparation.

Methodology. The study uses the mathematical apparatus of queuing systems, specifically the simplest Poisson process, to determine the probabilities of air raids, explosions, and human losses. The general risk assessment method - Event Tree Analysis (ETA) - based on inductive reasoning, was also applied to quantify possible scenarios.

Results. Based on the analysis of statistics in the Poltava region for the period from September 2024 to July 2025, the intensity of key events was determined: air raids (3.49 per day), explosions (1.04), casualties (1.39), and fatalities (0.26). The probabilities of these events occurring within a single day were calculated: the probability of at least one air raid is 97%, an explosion - 65%, casualties - 75%, and fatalities - 23%. A graphical model (event tree) was constructed to identify emergency development scenarios from the initiating event (alarm) to the final consequences.

Novelty. The novelty lies in applying mathematical methods of queuing theory to specific events of martial law (missile and drone strikes) to transition from simple danger notification to full-scale forecasting of emergency probabilities. For the first time, using a specific rear region as an example, the probabilistic characteristics of military risks have been quantitatively assessed.

Practical Significance. The obtained predictive models allow authorities and risk management entities to make timely decisions regarding the preparation of civil protection forces, transferring educational institutions to online learning or remote work, and preparing medical facilities to receive casualties. The methodology can be scaled to other regions of Ukraine to create evidence-based response plans for wartime threats.