REGRESSION MODELLING OF AIR QUALITY BASED ON METEOROLOGICAL PARAMETERS AND SATELLITE DATA

Although field monitoring can provide an accurate measurement of pollution, these measurements are of a limited spatial coverage. On the contrary, satellite-based observations can provide Aerosol Optical Depth (AOD) products with higher spatial resolution and continuous spatial coverage; however these products cannot directly measure the pollution concentration. In this study, the potential of a Moderate-Resolution Imaging Spectroradiometer (MODIS) sensors was investigated to evaluate the air quality parameters, after which water consumption in the studied area was considered. For this purpose, linear regression analysis was used in order to develop a relationship among MODIS-AOD, metrological data (relative humidity, temperature, precipitation, and wind speed) and air pollution data (CO, O-3, NO2, SO2, PM2.5) gathered 22 monitoring stations from 2012 to 2016. Among the 5 years of pollution data collection, the period of 2012 to 2014 was used for the model calibration and the period of 2015 to 2016 was used for the validation of the model. The results indicated that the regression models were of the best performance during spring (R-2 = 0.901 for CO), moderate performance during winter (R-2 = 0.674 for CO) and autumn (R-2 = 0.694 for CO), and weak performance during summer (R-2 = 0.181 for SO2). The results of the validation process also showed that the maximum determination factor (R-2 = 0.83) was obtained during spring season and for PM2.5 and the least (R-2 = 0.18) was obtained during summer and for SO2. Meanwhile, the assessment of water consumption demonstrated that there is significant relationship between water consumption and the concentration of pollution parameters.