Login / Signup

Low-Cost Hourly Ambient Black Carbon Measurements at Multiple Cities in Africa.

Abhishek AnandN'Datchoh Evelyne TouréJulien BahinoSylvain GnamienAllison Felix HughesRaphael E ArkuVictoria Owusu TawiahAraya AsfawTesfaye MamoSina HasheminassabSolomon BililignVaios MoschosDaniel M WesterveltAlbert A Presto
Published in: Environmental science & technology (2024)
There is a notable lack of continuous monitoring of air pollutants in the Global South, especially for measuring chemical composition, due to the high cost of regulatory monitors. Using our previously developed low-cost method to quantify black carbon (BC) in fine particulate matter (PM 2.5 ) by analyzing reflected red light from ambient particle deposits on glass fiber filters, we estimated hourly ambient BC concentrations with filter tapes from beta attenuation monitors (BAMs). BC measurements obtained through this method were validated against a reference aethalometer between August 2 and 23, 2023 in Addis Ababa, Ethiopia, demonstrating a very strong agreement ( R 2 = 0.95 and slope = 0.97). We present hourly BC for three cities in sub-Saharan Africa (SSA) and one in North America: Abidjan (Côte d'Ivoire), Accra (Ghana), Addis Ababa (Ethiopia), and Pittsburgh (USA). The average BC concentrations 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/m 3 , 5.33 μg/m 3 , 5.63 μg/m 3 , 3.89 μg/m 3 , 9.14 μg/m 3 , and 0.52 μg/m 3 , respectively. BC made up 14-20% of PM 2.5 mass in the SSA cities compared to only 5.6% in Pittsburgh. The hourly BC data at all sites (SSA and North America) show a pronounced diurnal pattern with prominent peaks during the morning and evening rush hours on workdays. A comparison between our measurements and the Goddard Earth Observing System Composition Forecast (GEOS-CF) estimates shows that the model performs well in predicting PM 2.5 for most sites but struggles to predict BC at an hourly resolution. Adding more ground measurements could help evaluate and improve the performance of chemical transport models. Our method can potentially use existing BAM networks, such as BAMs at U.S. Embassies around the globe, to measure hourly BC concentrations. The PM 2.5 composition data, thus acquired, can be crucial in identifying emission sources and help in effective policymaking in SSA.
Keyphrases
  • particulate matter
  • air pollution
  • low cost
  • cystic fibrosis
  • electronic health record
  • machine learning
  • transcription factor
  • big data
  • artificial intelligence