The results of the development of a simulation model for assessing the environmental risk of changes in the state of the forest fund in the Khanty-Mansiysk Autonomous Okrug using heterogeneous multidimensional data on accidents in oil pipelines in 2010–2018 are presented. The developed model is based on machine learning methods that allow to determine the degree of risk of repeated emergencies in the license areas, and geoinformation analysis
methods that are used to build a digital map of the risks of negative impacts on forest lands. Presented as an interactive digital map of the region, the simulation model allows to identify the territories with the highest environmental
risk. The reliability of the results discussed is confirmed by numerical experiments.
Materials and methods
Machine learning methods, spatial analysis methods, geoinformation technologies, risk-based approach.
risk-based approach, environmental risks, neural networks, machine learning, geoinformation analysis, GIS technologies