The paper is carried out to the investigation of the possibility of retrieving absolute humidity profile of the atmosphere using an artificial neural network based on the modeling of radiometric data of the passive microwave complex MIRS, which is part of the scientific equipment of the space experiment "Convergence". The main approaches to the construction of artificial neural networks are considered. The process of modeling MIRSs radiometric data are described. Selection of optimal characteristics of the neural network is carried out. Necessity of the information about atmospheric temperature profile for the best accuracy in solving the inverse problem are shown. The advantages of using differential channels in the 22 GHz absorption band for the humidity profile retrieving are proved. The expected errors of the atmospheric humidity profile retrieving during the Convergence experiment at altitudes from 0 to 10 km are given.

In this work, the possibility of retrieving the profile of absolute humidity of the atmosphere by means of an artificial neural network is investigated based on modeling radiometric data of the MIRS passive microwave complex, which is a part of the scientific equipment of the Convergence space experiment. The process of modeling MIRS radiometric data is described. Optimum characteristics of the neural network are selected. The necessity of information about the atmospheric temperature profile for the best accuracy in solving the inverse problem is shown. The advantages of using “differential” channels in the 22-GHz absorption band for retrieving the humidity profile are demonstrated. The expected errors in retrieving the atmospheric humidity profile in the course of the Convergence space experiment at heights from 0 to 10 km are presented.

This paper presents the results of a comparison of the seasonal dynamics of carbon dioxide content in the atmosphere and the phenological phases of various water bodies located in the boreal and subarctic zones for 2012–2020. Six large freshwater areas were selected as research objects: Lake Baikal, Lake Ladoga, Gulf of Ob, Lake Huron, Great Bear Lake and Great Slave Lake. In addition, four areas of wetlands in Western Siberia were studied. The CO2 concentration in the air column over water bodies was obtained from CAMS global greenhouse gas reanalysis data. These data represent three-dimensional fields of aerosols and chemical components in the atmosphere, with complete coverage of the globe. The phenological phases of the water bodies were determined using data from the MIRAS microwave radiometer of the SMOS satellite. The comparison and analysis performed showed that the concentration of CO2 in the atmosphere over the studied water bodies has a seasonal cyclical nature. The minimum concentration corresponds to the summer period due to strong photosynthesis of phytoplankton and swamp vegetation, which results in the absorption of carbon dioxide. The maximum CO2 concentration over the water bodies corresponds to the end of the winter – the beginning of the spring period, when photosynthesis processes are minimal. During this period, CO2 is being accumulated in water bodies and subsequently emitted. In freezing lakes located in the boreal zone, in addition to a stable spring maximum of CO2, a strong short-term release of carbon dioxide is sometimes observed, also corresponding to the stage of destruction of the ice cover. It is associated with the release of CO2 accumulated over the winter period, which was “sealed” in the ice and in the water column under the ice. The absence of such a release in the waters of the subarctic zone is explained by the lower biological productivity of these reservoirs compared to the waters of the boreal zone.