10th International Aerosol Conference September 2 - September 7, 2018 America's Center Convention Complex St. Louis, Missouri, USA
Abstract View
Field Measurements of Solid-Fuel Cookstove Emissions from Uncontrolled Cooking in China, Honduras, Uganda, and India
ROSE EILENBERG, Kelsey Bilsback, Michael Johnson, Jack Kodros, Eric Lipsky, Christian L'Orange, Jeffrey R. Pierce, R. Subramanian, John Volckens, Allen Robinson, Carnegie Mellon University
Abstract Number: 1573 Working Group: Combustion
Abstract Roughly 40% of the global population relies on cookstoves for their daily cooking and heating. The impacts of cookstoves are wide-reaching, affecting human health, local air quality, and the climate; however, the magnitudes of these effects are poorly quantified. Emissions factors, which are used in both climate and health-effects models, are highly uncertain because current laboratory testing protocols do not capture the magnitude and variability of real-world emissions. Several field studies have reported in-use emissions; however, comparing the emissions between regions is difficult due to the range of reported pollutants, metrics, and study instrumentation.
To address this gap, we used a portable emissions sampler to measure in-use cookstoves in 40 households in China, Honduras, Uganda, and India. These regions covered a range of fuels, cooking practices, and stove technologies (including traditional and improved stoves). The same instruments and protocols were used at each field site. We report test-integrated fuel-based emission factors (EF), including carbon monoxide (CO), fine particulate matter (PM2.5), organic matter (OM), elemental carbon (EC), OC:EC ratios, and total particle number (PN), as well as average particle number size distributions.
Our results indicate that the characteristics of cookstoves emissions vary across regions. We found the largest PM2.5 mass EFs in China, which included some very high emitters (>100 g/kg-fuel). The smallest PM2.5 mass EFs were measured from charcoal stoves in Uganda. The magnitude of PM2.5 mass EFs in Honduras and India were similar, but their composition was distinct: the OC:EC ratios measured for stoves in India were smaller than those in Honduras, indicating that, in addition to fuel, stove technology and cooking practices impact emissions. Variations in CO EFs reflect differences in combustion conditions: smoldering operations emitting more CO emissions (as in China, Honduras, and Uganda) while flaming combustion produces less (as in India). While CO has been proposed as a proxy from PM, we did not find any correlation between average CO emissions and test-integrated PM2.5 mass overall, though there was some correlation (R2=0.67) between the average CO EFs and OM EFs for wood stoves. We did not find any statistically significant differences between the emissions of wood-burning improved and traditional stoves in Honduras and India.
A novel aspect of this study was the measurement of particle number size distributions from the field. Charcoal stoves in Uganda emitted the smallest particles with a mode near 35 nm, followed by Honduras (~40 nm), and India (~90 nm). There were a wide range in China, with three groups: one around 50 nm, one near 125 nm, and a third, broader group which appears bimodal. The total PN EFs varied by two orders of magnitude, on average. The PN EFs were lowest in India, an order of magnitude larger in Uganda, and two orders of magnitude larger in China.
Time-resolved data includes black carbon (BC), particle number, and gases (from which modified combustion efficiency, MCE, was calculated). We saw the widest and most skewed distribution of MCE and particle size in China, with stoves producing larger particles than were measured in the other countries. Stoves in India operated in the narrowest range of MCEs and emitted particle size. The distributions in Uganda and Honduras were similar. The distribution of BC and PN EFs were skewed, indicating that total emissions are driven by high emission events; elevated emissions of BC and particle number were associated with cooking events such as fire starting and fuel addition. CO emissions, however, were relatively evenly distributed across the entire test.