Estimation of Total and Biomass-Based BC at African Cities by Applying Image-Reflectance Method on Bam Tapes

Abhishek Anand, Richard Djarbeng, Evelyne Toure, Julien Bahino, Sylvain Gnamien, Allison Felix Hughes, Raphael E Arku, Victoria Owusu Tawiah, Araya Asfaw, Tesfaye Mamo, Sina Hasheminassab, Solomon Bililign, Daniel Westervelt, ALBERT A. PRESTO, Carnegie Mellon University

     Abstract Number: 498
     Working Group: Urban Aerosols

Abstract
There is a lack of continuous monitoring of air pollutants in the Global South due to high cost of regulatory monitors. We have developed a low-cost method that utilizes reflected visible light from particle deposits on filters to measure black carbon (BC). US embassies use Beta Attenuation Monitors (BAMs) to measure PM2.5. A BAM collects PM2.5 on filter tapes every hour as spots and uses beta ray attenuation through each spot to measure hourly PM2.5. We photograph particle spots on used filter tapes from these existing BAMs at US Embassies one at a time with a custom reference card and then apply an image processing algorithm to extract red, green, and blue color channels for the filter samples. The red channel on the photo is used to estimate total BC concentration for that hour. We present 2 months of hourly BC concentrations for Pittsburgh (USA) and three cities in Africa, including Abidjan (Côte d’Ivoire), Addis Ababa (Ethiopia), and Accra (Ghana). Our BC estimates from BAM tapes showed strong correlation (R2 ~0.96, slope = 0.97) with microAeth measurements at Addis Ababa, Ethiopia. The hourly BC data at the three sites show a clear diurnal pattern of BC with prominent peaks during the morning and evening rush hours on workdays. The average total BC for the measurement period at the Abidjan, Accra, Addis Ababa Central summer, Addis Ababa Central winter, Addis Ababa Jacros winter and Pittsburgh sites were 3.85 µg/m3, 5.33 µg/m3, 5.63 µg/m3, 3.89 µg/m3, 9.14 µg/m3 and 0.52 µg/m3, respectively. The average BC to PM2.5 fraction varied between ~13 – 20% at African locations compared to only 5% in Pittsburgh. We further calculated biomass-attributed BC at these cities by quantifying the difference of BC individually estimated from red and blue channels. BC composition data, thus acquired, can be crucial in identifying emission sources and help in effective policymaking.