The physical-geographical and socio-geographical conditions of the water area and the Ob Bay area as an integral part of the Arctic zone of the Russian Federation are analyzed. The directions of using geoinformation models and methods of representation and use of spatial data to support the management of territorial economic activity in the Arctic are determined. The methods of using artificial neural networks to assess the navigational-hydrographic, hydrometeorological and ecological situation

more »

... in areas adjacent to specially protected Arctic natural territories have been developed. A geodata model based on artificial neural networks containing the minimum set of territorial parameters necessary for spatial analysis is proposed. The set of parameters necessary to achieve the specified accuracy in the analysis of the situation is justified. The requirements for the training surface of such a network are formulated. The comparison of different architectures of neural networks with a different number of internal layers and neurons on them is carried out. The application of a neural network by the type of recurrent neural structure with two hidden layers consisting of N neurons is justified, as well as a procedure for training a network on ten thousand sets with an error propagation algorithm. The problem of the estimated analysis of the navigational-hydrographic and ecological situation in the area of the Gulf of Ob, which is a typical classification problem for an artificial neural network, has been solved. Based on the results, evaluation cartoids were compiled, which make it possible to analyze and predict the environmental situation in the region.