Abstract View
Evaluating an Indoor Air Quality Model Using Simultaneous Measurements of Cookstove PM2.5 Emissions and Indoor Concentrations
MOHAMMAD MAKSIMUL ISLAM, Roshan Wathore, Hisham Zerriffi, Julian Marshall, Rob Bailis, Andrew Grieshop, North Carolina State University
Abstract Number: 522
Working Group: Indoor Aerosols
Abstract
Combustion of biomass in residential cookstoves is a major source of household air pollution (HAP), an acknowledged threat to human health. Multiple studies explore the effect of cookstove use on HAP, but few had simultaneous measurements of both emission and indoor air quality in different seasons and locations. Measurements of air exchange rate (AER) in houses in developing countries are also limited. Thus, the quantitative linkage between cookstove emission and indoor air quality is still poorly constrained. Here, we aim to improve links between estimates of cookstove emissions and indoor PM2.5 using data collected during a cookstove intervention trial in two rural areas in India (Kullu in Himachal Pradesh State; Koppal in Karnataka State). We measured real-time and gravimetric indoor PM2.5 concentrations during ~5000 cooking events of traditional and alternate biomass and modern-fuel stoves. We also conducted simultaneous emission measurements for a subset of those cooking events. We use these data to evaluate a Monte-Carlo single box model for HAP, developed by the World Health Organization (WHO), and used to establish emission rate targets (performance tiers) for clean stoves.
In general, Kullu households had ~50% lower PM2.5 concentration than those in Koppal, consistent with the observed higher estimated AERs and shorter cooking times in Kullu compared to Koppal. We applied multilinear regression modeling, which showed that ventilation and cooking characteristics have a large influence on indoor PM. We used the WHO model with measured emissions rates and other household characteristics (volume, AER) and compared the results to observed indoor PM2.5 concentrations. We find that model greatly overestimates (by a factor of ~10) average kitchen concentration. Further analysis explores various factors (e.g. stove types, presence of chimney, AER, monitoring height and kitchen volume) that may affect model estimations.