Retrieving PM10 Surface Concentration from AERONET Aerosol Optical Depth: The Cairo and Delhi Megacities Case Studies

Abstract

AbstractLarge concentrations of air-suspended particulate matter (PM) in megacities represent an important health risk for their populations, but PM time series are often missing or too short to quantify the associated burden of diseases. In this study, we propose a model for retrieving the surface PM in Cairo (Egypt) and Delhi (India) from the automated measurements of aerosol optical depth (AOD), precipitable water (PW), and Angström exponent (AE) performed by the sunphotometers of the Aerosol Robotics Network (AERONET). For this we exploit the (1) synchronous measurements performed from 2010 to 2015 at the headquarters of the Egyptian Meteorological Authority and in 2009 at the Gual Pahari station (25 km south of Delhi) and (2) the ERA5 estimate of the planetary boundary layer height (H). The correlation between the surface PM10 and the AOD is primarily controlled by the variations of PW and secondarily by those of H: for similar surface PM10 concentrations, the AOD tends to be the largest in summer because of the hygroscopic enhancement of the mass extinction efficiency (σ) of the particles and their dilution in the more developed mixing layer. The variations of composition also play a significant role in Cairo. This effect, particularly marked in spring (coinciding with the dust season), can be parameterized as a linear function of AE. Finally, we show that the variations of the surface PM10 concentration at the two sites can be retrieved simply from those of the AOD, PW, AE and H. At the weekly temporal resolution, the agreement between the model and the observations is very good at the two locations (correlation coefficient > 0.81, relative mean absolute error < 15%). This validates indirectly the assumption made in the development of the model, namely that the aerosols are mostly confined to the mixing layer of the two megacities and not transported in the free atmosphere. Provided a few years of surface PM measurements are available, the methodology proposed in this study could be easily applied to any other AERONET station.