Short-term estimations of PM10 concentration in the Middle Black Sea region based on grey prediction models

Abstract

The Middle Black Sea region has experienced severe air pollution, with a significant increase in particulate matter (PM) concentration due to a growth in population, financial activity, and an expansion of transportation in recent years. Therefore, the prediction of PM concentration has become a topic of great significance to reduce air pollution and assess the effects on public health. In this study, the grey prediction model (GM (1,1)), the discrete grey model (DGM (1,1)), and the grey Verhulst model (GVM (1,1)) were used to estimate the PM10 concentration of the cities Amasya, Çorum, Ordu, and Samsun in the Middle Black Sea region, for the period from 2022 to 2026. The accuracy of the GM (1,1), DGM (1,1), and GVM (1,1) models in fitting data was calculated using the mean absolute percentage error (MAPE) value. Since three types of prediction models of MAPEs were less than 20%, they were considered a good value for prediction performance. Furthermore, the results showed that the PM10 concentrations of Amasya, Çorum, and Ordu showed a downward trend over the next 5 years. However, the GVM (1,1) model showed an upward trend in the yearly average PM10 concentration in Samsun. In conclusion, these models could be considered a reliable approach in early warning systems for emissions reduction and as a long-term policy for managing air quality in the Middle Black Sea region.