Real-time indoor measurement of health and climate-relevant air pollution concentrations during a carbon-finance-approved cookstove intervention in rural India

Biomass combustion in residential cookstoves is a major source of air pollution and a large contributor to the global burden of disease. Carbon financing offers a potential funding source for health-relevant energy technologies in low-income countries. We conducted a randomized intervention study to evaluate air pollution impacts of a carbon-finance-approved cookstove in rural South India. Prior research on this topic often has used time-integrated measures of indoor air quality. Here, we employed real-time monitors (∼24 h measurement at ∼ minute temporal resolution), thereby allowing investigation of minutely and hourly temporal patterns. We measured indoor concentrations of fine particulate matter (PM_2.5), black carbon (BC) and carbon monoxide (CO) in intervention households (used newer, rocket-type stoves) and control households (“nonintervention”; continued using traditional open fire stoves). Some intervention households elected not to use only the new, intervention stoves (i.e., elected not to follow the study-design protocol); we therefore conducted analysis for “per protocol” versus “intent to treat.” We compared 24 h averages of air pollutants versus cooking hours only averages. Implementation of the per protocol intervention cookstove decreased median concentrations of CO (by 1.5 ppm (2.8 − 1.3; control − per protocol), p = 0.28) and PM_2.5 (by 148 μg/m³ (365 − 217), p = 0.46) but increased BC concentration (by 39 μg/m³ (26 − −12), p < 0.05) and the ratio of BC/PM_2.5 (by 0.25 (−0.28 − −0.03), p < 0.05) during cooking-relevant hours-of-day relative to controls. Calculated median effective air exchange rates based on decay in CO concentrations were stable between seasons (season 1: 2.5 h⁻¹, season 2: 2.8 h⁻¹). Finally, we discuss an analytical framework for evaluating real-time indoor datasets with limited sample sizes. For the present study, use of real-time (versus time-averaged) equipment substantially reduced the number of households we were able to monitor.