The possibility of increasing the reliability of assessing air pollution through the use of probabilistic methods is shown. Evaluation of quality indicators on a quantitative basis made it possible to find an error of the first and second kind in assessing the quality of the atmosphere. Based on the Bayesian approach, dependencies were obtained for estimating and predicting atmospheric pollution in the area of ​​a city that does not have a stationary post, based on data from stationary posts in other areas.

Keywords: atmosphere quality, reliability of the estimate, probabilistic methods, quantitative characteristic, error of the first and second kind, Bayesian approach, a priori probability, posterior probability, likelihood of a quantity