Open AccessAir Quality Impacts at an E‐Waste Site in Ghana Using Flexible, Moderate‐Cost and Quality‐Assured Measurements
Author(s) -
Kwarteng Lawrencia,
Baiden Emmanuel Acquah,
Fobil Julius,
ArkoMensah John,
Robins Thomas,
Batterman Stuart
Publication year - 2020
Publication title -
geohealth
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.889
H-Index - 12
ISSN - 2471-1403
DOI - 10.1029/2020gh000247
Subject(s) - environmental science , air quality index , combustion , particulates , aerosol , spatial variability , sampling (signal processing) , atmospheric dispersion modeling , environmental engineering , biomass burning , meteorology , air pollution , hydrology (agriculture) , atmospheric sciences , geography , engineering , geology , ecology , chemistry , statistics , mathematics , organic chemistry , filter (signal processing) , geotechnical engineering , electrical engineering , biology
Air quality information is scarce in low‐ and middle‐income countries. This study describes the application of moderate cost approaches that can provide spatial and temporal information on concentrations of particulate matter (PM) needed to assess community and occupational exposures. We evaluated PM levels at the Agbogbloshie e‐waste and scrap yard site in Accra, Ghana, and at upwind and downwind locations, obtaining both optical and gravimetric measurements, local meteorological data and satellite aerosol optical depth. Due to overload issues, the gravimetric 24‐hr samplers were modified for periodic sampling and some optical data were screened for quality assurance. Exceptionally high concentrations (e.g., 1‐hr average PM 10 exceeding 2000 μg/m 3 ) were sometimes encountered near combustion sources, including open fires at the e‐waste site and spoil piles. 24‐hr PM 2.5 levels averaged 31, 88 and 57 μg/m 3 at upwind, e‐waste and downwind sites, respectively, and PM 10 averaged 145, 214 and 190 μg/m 3 , considerably exceeding air quality standards. Upwind levels likely reflected biomass burning that is prevalent in the surrounding informal settlements; levels at the e‐waste and downwind sites also reflected contributions from biomass combustion and traffic. The highest PM levels occurred in evenings, influenced by diurnal changes in emission rates, atmospheric dispersion and wind direction shifts. We demonstrate that moderate cost instrumentation, with some modifications, appropriate data cleaning protocols, and attention to understanding local sources and background levels, can be used to characterize spatial and temporal variation in PM levels in urban and industrial areas.