Forecasting air pollution rates in industrial centres: a case study for Kocaeli-Turkey

Abstract

Around the world, air pollution is a leading social problem despite all the precautions. Of all the reasons for this, human-induced ones are at the top of the list. These can be prevented or at least reduced by some measures. Thus, forecasting air pollution is an important issue for all related agencies. In this paper, in order to forecast future values of air pollution rates, two different model approaches were considered. The first one is the Box-Jenkins time series model and the second one is a trigonometric Y-t = mu + Acos(w(k)t) + Bsin(w(k)t) + e(t) time series model. As a main air pollution measurement index, monthly average amounts of particulate matter, PM10 were used. Kocaeli was selected as a representative city because it is the most important industrial city in Turkey. Hence, data were collected monthly from different air monitoring stations in the city of Kocaeli between 2007 and 2018 and averaged to produce a single time series. Both models yielded high forecast values of PM10.