نمذجة الأتربة العالقة في مكة المكرمة

Abstract

Atmospheric particulate matters, especially those with smaller sizes such as PM10 are associated with numerous adverse health and environmental impacts. PM10 concentrations are high in Makkah and exceed air quality standards. High PM10 levels might pose threat to human health, therefore for effective management it is vital to model and investigate the effect of different factors on PM10 concentrations in Makkah. In this paper for the first time in Makkah a quantile regression model is developed using hourly PM10 concentrations (µg/m3) as dependent variable and several air pollutants and meteorological variables as independent variables for year 2012. Quantile regression model addresses the problem of non-normal distributions of air quality data, and non-linearity in the association of PM10 with the covariates. All covariates show significant effect at least at one or more quantiles, however, wind speed, carbon monoxide (CO), nitric oxide (NO) and previous day PM10 (lag_PM10) concentrations have significant effect at all quantiles and hence are considered the most important factors for controlling PM10 concentrations. Furthermore, CO has negative impact, whereas wind speed, NO and lag_PM10 have positive impact on PM10 concentrations. The strength and nature (positive or negative) of coefficients vary at different quantiles of the PM10 distribution. The performance of the model was assessed using several statistical metrics, including correlation coefficients (R, 0.82), factor of 2 (FACT2, 0.96), Root Mean Square Error (RMSE, 129), Normalised Mean Bias (NMB, 0.12) and Normalised Mean Gross Error (NMGE, 0.34). The values of these metrics and graphical presentations show that QRM performs better and explains significantly more variations in PM10 concentrations than the multiple linear regression model. This is the first study that uses a quantile regression approach for modelling PM10 levels in Makkah and probably elsewhere, and may help characterise and manage PM10 concentration in Makkah.