ПРОГРАММНО-АЛГОРИТМИЧЕСКИЙ КОМПЛЕКС ПРОГНОЗИРОВАНИЯ ДИНАМИКИ АРКТИЧЕСКИХ ОЗЕР РОССИИ НА ОСНОВЕ СПУТНИКОВЫХ СНИМКОВ И ЭНТРОПИЙНО-РАНДОМИЗИРОВАННОГО ПОДХОДА

Abstract

The article is devoted to the problem of predicting the evolution of thermokarst lakes in permafrost zones as intensive sources of natural emissions of greenhouse gases into the atmosphere in the Arctic territories. Goal of the work. The purpose of the work was to consider the issues of creating a software-algorithmic complex for predicting the spatio-temporal dynamics of lakes in the Russian Arctic based on methods and algorithms of randomized machine learning. Materials and methods. For forecasting, time series of satellite measurements of the areas of thermokarst lakes in the Arctic zone of Russia and data on the average annual temperature and annual precipitation obtained on the basis of the ERA-5, ERA-Interim, etc. reanalysis systems are used for forecasting. Methods of entropy-randomized dynamics modeling are used fields of thermokarst lakes, allowing to predict changes in the areas of lakes in the Arctic zone. For the software implementation of a complex for predicting the evolution of the area of thermokarst lakes, modern geographic information systems are used. Results. The architecture of the software-algorithmic complex has been developed, based on the use of entropy-randomized modeling algorithms. The software-algorithmic forecasting complex makes it possible to train and test the model based on available historical data on the dynamics of the area of thermokarst lakes and climate changes in the Russian Arctic. Conclusion. The implementation of a software package based on the NextGIS Web geographic information system allows you to include forecasting applications in Python. The developed software package can be used in assessing and forecasting the dynamics of greenhouse gas emissions from lakes, which influence changes in air temperature in the northern regions.