Abstract—One of the options for the practical application of GNSS technology, in addition to geodetic and navigation needs, is remote sensing of the atmosphere by radio signals from navigation satellites in order to improve the quality and detail of weather forecasts. The propagation of a radio signal from GNSS satellites to a ground receiving device (GNSS receiver) through a neutral atmosphere is accompanied by a decrease in the phase velocity of the radio waves (additional atmospheric delays). This is due to the presence of nitrogen, oxygen, carbon dioxide, and water vapor molecules in the atmosphere. Therefore, measurements of the additional delay of the radio signal in the atmosphere (tropospheric delay) provide information on the integral properties of the atmosphere along the propagation path of the radio signal. As a result of the primary processing of the GNSS measurement results, the distances from the observation station to GNSS satellites are determined. Secondary processing of GNSS measurements consists in solving a navigation problem and provides information on the location of the station. In order to obtain meteorological information, it is necessary to develop special methods of secondary data processing based on solving inverse problems. The combination of primary and secondary data (processing) along with meteorological information makes it possible to obtain a global model of the atmosphere in near-real time. The efficiency of this approach, the complete automation, and the absence of consumables during remote sensing provide opportunities for the widespread implementation of operational monitoring of the state of the atmosphere in order to improve the data’s detail and accuracy of regional short-term weather forecasts. Currently, due to cross-border cooperation with European countries in conducting joint GNSS observations in the UA-EUPOS/ZAKPOS network of stations, we are able to have an accurate, dense, and continuous sampling of tropospheric delay values, which allows us to determine and predict the dynamics of atmospheric changes in real time. The main goal of the work is to study the spatio-temporal instability of the atmosphere over an area covered by active reference stations. The results of the study can be used to improve the quality of weather prediction.