Analysis of the probability of windfall occurrence on the territory of the Republic of Belarus

Abstract

Purpose. Systems analysis and development of an integrated approach to assessing the likelihood of windfalls to support management decision-making in the field of emergency prevention.

Methods. The study is based on the method of systems analysis to structure the multifactorial nature of the probability of windfall occurrence. Fuzzy logic (the Mamdani model) was used for the integrated assessment, allowing for processing of linguistic variables. A Python software module was developed that utilizes computer vision and machine learning methods to recognize tree species and determine their biometric parameters from photographs.

Findings. Three groups of factors influencing the probability of windfalls have been identified and classified: climatic (speed and frequency of wind gusts, precipitation), geographical (terrain exposure, soil type, position in the massif) and forest vegetation (stability of the species, height and diameter of trees). Term sets were developed for each group. A fuzzy logic model was proposed that integrates these factors into a single probability estimate. A software prototype with a graphical interface has been created that automatically extracts forest vegetation parameters from images and calculates the probability of windfalls.

Application field of research. The results have practical significance for the Ministry of Emergency Situations of the Republic of Belarus. The methodology and software module can be used to forecast storm impacts, plan preventive measures, and create probability maps of windfalls occurrence. Integration with monitoring systems will enable the creation of an effective early warning system and minimization of damage from windfalls.