Seasonal monitoring and estimation of regional aerosol distribution over Po valley, northern Italy, using a high-resolution MAIAC product

Abstract

In this work, the new 1 km-resolved Multi-Angle Implementation of Atmospheric Correction (MAIAC) algorithm is employed to characterize seasonal PM10 – AOD correlations over northern Italy. The accuracy of the new dataset is assessed compared to the widely used Moderate Resolution Imaging Spectroradiometer (MODIS) Collection 5.1 Aerosol Optical Depth (AOD) data, retrieved at 0.55 μm with spatial resolution of 10 km (MYD04_L2). We focused on evaluating the ability of these two products to characterize both temporal and spatial distributions of aerosols within urban and suburban areas. Ground PM10 measurements were obtained from 73 of the Italian Regional Agency for Environmental Protection (ARPA) monitoring stations, spread across northern Italy, during a three-year period from 2010 to 2012. The Po Valley area (northern Italy) was chosen as the study domain because of its severe urban air pollution, resulting from it having the highest population and industrial manufacturing density in the country, being located in a valley where two surrounding mountain chains favor the stagnation of pollutants. We found that the global correlations between the bin-averaged PM10 and AOD are R2 = 0.83 and R2 = 0.44 for MYD04_L2 and for MAIAC, respectively, suggesting a greater sensitivity of the high-resolution product to small-scale deviations. However, the introduction of Relative Humidity (RH) and Planetary Boundary Layer (PBL) depth corrections allowed for a significant improvement to the bin-averaged PM – AOD correlation, which led to a similar performance: R2 = 0.96 for MODIS and R2 = 0.95 for MAIAC. Furthermore, the introduction of the PBL information in the corrected AOD values was found to be crucial in order to capture the clear seasonal cycle shown by measured PM10 values. The study allowed us to define four seasonal linear correlations that estimate PM10 concentrations satisfactorily from the remotely sensed MAIAC AOD retrieval. Overall, the results show that the high resolution provided by MAIAC retrieval data is much more relevant than the 10 km MODIS data to characterize PM10 in this region of Italy which has a pretty limited geographical domain but a broad variety of land usages and consequent particulate concentrations.