Forecasting of air pollution with time series and multiple regression models in Sofia, Bulgaria

• Published in: Journal of environmental engineering and landscape management Vol. 31; no. 3; pp. 176 - 185
• Main Authors: Stoyanov, Nikolay, Pandelova, Antonia, Georgiev, Tzanko, Kalapchiiska, Julia, Dzhudzhev, Bozhidar
• Format: Journal Article
• Language: English
• Published: Vilnius Vilnius Gediminas Technical University 02.08.2023
• Subjects: Adequacy; Air monitoring; Air pollution; Air pollution forecasting; Air quality; Air temperature; Analysis; Atmospheric models; Autoregressive models; Case studies; Environmental engineering; Environmental impact; Forecasting; Integrated Autoregressive Moving Average (ARIMA); meteorological variables; multiple linear regression; Multiple regression models; Neural networks; Outdoor air quality; Particulate emissions; Particulate matter; PM10; Radiation; Regression analysis; Software; Solar radiation; Statistical analysis; Suspended particulate matter; Time series; Variables; Weather; Wind direction; Wind speed; Bulgaria; Sofia Bulgaria; PM10; air pollution; Integrated Autoregressive Moving Average (ARIMA); meteorological variables; multiple linear regression
• ISSN: 1648-6897; 1822-4199
• DOI: 10.3846/jeelm.2023.19467

Air pollution is one of the serious environmental problems. The high concentrations of particulate matter can have a serious impact over human health and ecosystems, especially in highly urbanized areas. In this regard, the present study employs a combined ARIMA-Multiple Linear Regression modelling approach for forecasting particulate matter content. The capital city of Bulgaria is used as case study. A regression analysis techniques are used to study the relationship between particulate matter concentration and basic meteorological variables – air temperature, solar radiation, wind speed, wind direction, atmospheric pressure. The adequacy of the models has been proven by examining the behavior of the residues. The synthesized time series model can be used for forecasting, monitoring and controlling the air quality conditions. All analyzes and calculations were performed with statistical software STATGRAPHICS.