Modeling of daily PM2.5 concentration based on the principal components regression in South and Central Jakarta

Abstract

Abstract Jakarta as a megacity, has many roles in the distribution of PM2.5 pollutant concentrations, especially from the activities of vehicle users. This research estimates daily PM2.5 concentration from January to June 2018 at two points, each in the South Jakarta and Central Jakarta based on the principal components regression from daily averaged air temperature, rainfall, relative humidity, and surface wind speed. We retained the first three in principal components to account for over 90% of the variance in the climatological variables. The results show the daily concentration of PM2.5 in South Jakarta has the highest correlation coefficient of 0.32 in March 2018 with a p-value of 0.079 and MAE value of 10.28. Meanwhile, Central Jakarta has the highest correlation coefficient of 0.46 calculated from daily PM2.5 estimation in January 2018 with a p-value of 0.013 and MAE value of 8.27. Based on verification results at the two observation points from January to June 2018, it can be concluded that the estimation of the daily PM2.5 concentration based on the principal component regression method in the South Jakarta showed slightly better than in South Jakarta.