Modelling the Effect of COVID-19 Lockdown on Air Pollution in Makkah Saudi Arabia with a Supervised Machine Learning Approach

Abstract

To reduce the spread of COVID-19, lockdowns were implemented in almost every single country in the world including Saudi Arabia. In this paper, the effect of COVID-19 lockdown on O3, NO2, and PM10 in Makkah was analysed using air quality and meteorology data from five sites. Two approaches were employed: (a) comparing raw measured concentrations for the lockdown period in 2019 and 2020; and (b) comparing weather-corrected concentrations estimated by the machine learning approach with observed concentrations during the lockdown period. According to the first approach, the average levels of PM10 and NO2 decreased by 12% and 58.66%, respectively, whereas the levels of O3 increased by 68.67%. According to the second approach, O3 levels increased by 21.96%, while the levels of NO2 and PM10 decreased by 13.40% and 9.66%, respectively. The machine learning approach after removing the effect of changes in weather conditions demonstrated relatively less reductions in the levels of NO2 and PM10 and a smaller increase in the levels of O3. This showed the importance of adjusting air pollutant levels for meteorological conditions. O3 levels increased due to its inverse correlation with NO2, which decreased during the lockdown period.