Application of New MODIS-Based Aerosol Index for Air Pollution Severity Assessment and Mapping in Upper Northern Thailand

Abstract

This paper reports capability of a newly-proposed index called the aerosol prediction index (API) in the determination and mapping of near-ground PM10 concentrations (at spatial resolution of 500 x 500 m) during the 2009 and 2010 burning seasons in upper northern Thailand. API is a normalized index defined based on the difference in the observed reflectance data at two spectral bands of the MODIS instrument aboard NASA�s Terra satellite; Band 3 (blue) and Band 7 (mid-infrared). Initial analysis suggested that API had strong correlation with the corresponding MODIS-AOD and AERONET-AOD with coefficient of determination (R2) about 0.62 in both cases, and also with the reference PM10 data with R2 of 0.66. In terms of predictive performance, it exhibited low bias at low PM10 condition and achieved impressive prediction accuracy with relative error of 10.78 %. The near-ground PM10 concentration map yielded from the proposed index was proved very useful in the comprehensive assessment of aerosol pollution situation over entire area at fine spatial detail. This task could not be fulfilled from sole use of the ground-based measured data or standard MODIS-AOD product. These findings indicate that API should be a promising tool for the regular monitoring of air pollution severity over the concerned area.

Keywords

• particulate matter
• aerosol
• air pollution
• optical depth
• MODIS