Forecast of urban air pollution level by expertise

Introduction. A climate pattern with marine features is typical for St. Petersburg. Vagaries of weather and climate conditions in the last decade specify the timeliness of this work, the purpose of which is to outline the expected level of the open air pollution in St. Petersburg by the “decision tree” method. Materials and Methods. Current data of weather observations carried out at station 26063 (St. Petersburg) from 2006 to 2014 are studied and processed. Within the framework of the study, the data were considered on the vertical profile of the atmosphere obtained through radiosounding the atmosphere of St. Petersburg at 00.00 and 12.00 UTC (Universal Time Coordinated) at Voeykovo station. Research Results. In the course of the investigation, the dependence of the atmospheric air pollution level on the synoptic process and the inertial factor was established which made it possible to figure a scheme for forecasting the air pollution level in the form of the decision tree by expertise. Accuracy of the predictive determination of the expected air pollution group in St. Petersburg was calculated on the dependent material and topped 90% (nighttime hours) and 91% (daytime hours) for a cold period; and - 84% (nighttime hours) and 87% (daylight hours) for a warm period of the year. This suggests that the proposed schemes allow obtaining a more efficient prediction of the atmospheric air pollution level in a cold period of the year. Discussion and Conclusions. In conclusion, basic outcomes and inferences are summarized. - Archives of baseline standard meteorological data and data of the atmosphere radiosounding, as well as synoptic situations and information on the level of atmospheric air pollution in St. Petersburg for the period from 2006 to 2014, are formed. - Groups of synoptic processes typical for St. Petersburg from 2006 to 2014 are established. - Schemes for forecasting the atmospheric air pollution level are developed using the “decision tree” method with accuracy of 84-91%. The research results are applicable for forecasting the urban air pollution level.