Main content area

Exposures to and origins of carbonaceous PM2.5 in a cookstove intervention in Northern Ghana

Piedrahita, Ricardo, Kanyomse, Ernest, Coffey, Evan, Xie, Mingjie, Hagar, Yolanda, Alirigia, Rex, Agyei, Felix, Wiedinmyer, Christine, Dickinson, Katherine L., Oduro, Abraham, Hannigan, Michael
The Science of the total environment 2017 v.576 pp. 178-192
adults, air quality, biomass, burning, carbon, children, climate, combustion, cooking, cooking stoves, data collection, emissions, households, models, particulates, polycyclic aromatic hydrocarbons, Ghana
REACCTING (Research on Emissions Air Quality, Climate, and Cooking Technologies in Northern Ghana) was a 200-home cookstove intervention study from 2013 to 2015. Study households were divided into four groups: a control group, a group given two locally made rocket stoves, a group given two Philips forced draft stoves, and a group given a locally made rocket stove and a Philips stove. In a subset of study households, 48-hour PM2.5 exposure samples were collected for adults and children, as well as in the primary cooking area. Further, weekly ambient background PM2.5 samples were collected for the first nine months of the study. All PM2.5 samples were analyzed for elemental and organic carbon (EC/OC), and a subset was also analyzed for organics. Mixed effects modeling was applied to quantify differences in PM exposures between the groups and to assess relationships between exposures and cooking area measurements. Results showed that personal OC exposure for the intervention groups was 56.6% lower than the control group (p≤0.01). Both intervention groups given Philips stoves had significantly lower EC exposure than the control group (60.6% reduction, p≤0.02). Only weak relationships were found between personal and cooking area EC or OC. Source apportionment modeling was performed on both the personal/microenvironment and the ambient organics PM2.5 data sets to assess the sources of the observed PM. We identified six PM sources. The identified source factors were similar among the data sets, as well as with previous work in Navrongo. Two sources, one characterized by the presence of methoxyphenols, and one by the presence of polyaromatic hydrocarbons and EC, were associated with biomass burning, and accounted for a median of 9.2% of OC and 15.3% of EC personal exposure. Here, we demonstrate the utility of using the cooking-related source apportionment factors within a mixed effects model for more precise estimation of exposures due to cooking, rather than other combustion sources unrelated to the intervention.