Publications

Marj, AA; Mobasheri, MR; Matkan, AA (2017). Quantitative Assessment of Different Air Pollutants (QADAP) Using Daily MODIS Images. INTERNATIONAL JOURNAL OF ENVIRONMENTAL RESEARCH, 11(4), 523-534.

Abstract
Large-scale assessment of the amount of different air pollutants is routinely conducted using satellite images. This information is not useful in urban areas where more precise information is needed. In this work, a model called quantitative assessment of different air pollutants (QADAP) was prepared. The ability of QADAP is to produce images quantifying pollutant distribution throughout the city of Tehran at 500 m resolution. The model is based on surface reflectance changes due to air pollution. For this, the city was classified into various classes using GeoEye images. Then the reflectance of each class was calculated using this classified image along with a Hyperion image of the same region. Next, using this spectral information, a MODIS reflectance reference image in its first seven bands was prepared using MODIS image overlaid with the prepared classified image on a clear and clean sky. Then, some coefficients were calculated using surface reflectance differences between polluted days and the referenced clean day image. Required equations were then obtained by relating the pollution coefficients to the values of air pollutants measured in the pollution monitoring stations. Finally, given the equations and coefficients, pollutant concentration was calculated for each pixel and consequently daily images of different pollutants were produced. Data from ground stations were subsequently used to evaluate the model. The best results were for CO, PM2.5, NO2 and O-3, which had lower relative Root Mean Squared Errors (RMSE), and the worst result was for PM10 with a high relative RMSE. The relative error of the model was 13-25% for higher levels of pollution and 150-400% for lower values. Finally highly polluted areas were determined by accumulation of different pollution images acquired on consecutive days; this aggregated image was used to identify the most contaminated regions considerably in agreement with the ground measured values.

DOI:
10.1007/s41742-017-0046-y

ISSN:
1735-6865