FORECASTING THE HAZARDOUS IMPACT OF TRANSPORT DURING EMERGENCY TROPOSPHERIC PHENOMENA USING ARTIFICIAL INTELLIGENCE

In the article, based on a critical review of positive practices for monitoring extreme pollution of the urban environment with transport pollutants, a comprehensive two-level instrumental and digital methodology for assessing and predicting the hazardous impact of exhaust gases on the population in the abnormal physical state of the troposphere caused by temperature inversion and weak wind is substantiated. At the first stage of the methodology implementation, zones of extreme pollution of the urban area are determined based on data processed by a neural network, concentration fields of substances measured by stationary and mobile stations.The concentration fields are approximated using the Gaussian model. At the second stage, extremely dangerous, in relation to the LTCMS, accumulations of substances in exhaust gases at the levelof breathing of city residents are calculated using atmospheric diffusion equations in relationto the conditions of intense traffic during rush hours. For priority pollutants, for places of joint influence of motor transport and ships, the probability of exceeding concentrations of NO2 and PM2.5-10, respectively, up to 4–7 LTCMS and 2–3 LTCMS is substantiated. Original authorial approachesto forecasting emergency situations are implemented in domestic and international methods.