Grantee Research Project Results
Final Report: Experimental Interventions to Facilitate Clean Cookstove Adoption, Promote Clean Indoor Air, and Mitigate Climate Change
EPA Grant Number: R835421Title: Experimental Interventions to Facilitate Clean Cookstove Adoption, Promote Clean Indoor Air, and Mitigate Climate Change
Investigators: Bailis, Robert , Grieshop, Andrew P , Unger, Nadine , Zerriffi, Hisham , Dwivedi, Puneet , Talashery, Pradeep , Marshall, Julian D. , Chandar, Mamta
Institution: Stockholm Environment Institute
EPA Project Officer: Keating, Terry
Project Period: September 1, 2015 through August 31, 2018 (Extended to September 30, 2019)
Project Amount: $1,499,985
RFA: Measurements and Modeling for Quantifying Air Quality and Climatic Impacts of Residential Biomass or Coal Combustion for Cooking, Heating, and Lighting (2012) RFA Text | Recipients Lists
Research Category: Climate Change , Air Quality and Air Toxics , Tribal Environmental Health Research , Air
Objective:
This project had four broad objectives linked to improvements in clean stove design and dissemination and impacts on health and climate: 1) assess the acceptability and availability of stove technologies and fuels, 2) experiment by offering stoves for free or at a subsidy and under varying social interactions to determine the impact of these factors on stove adoption rates and outcomes, 3) measure in situ impacts of stove adoption on air pollution, and climate-forcing, and 4) model the impacts of widespread stove adoption on regional and global climate through a range of scenarios directly informed by data from the field. Fieldwork will occur in Himachal Pradesh (HP) and Karnataka (KA). The experimental stove sales and giveaways are funded by a grant from the Global Alliance for Clean Cookstoves (GACC). EPA funding was used only for surveying, monitoring, and field measurements of baseline conditions, and interventions in households with non-vulnerable populations.
Summary/Accomplishments (Outputs/Outcomes):
To achieve the objectives described above, the team selected two locations: one in the northern Indian state of Himachal Pradesh, and one in the Southern Indian state of Karnataka. Four communities were selected at each study site in consultation with local partners Samuha and Jagriti. Approximately 60 households were recruited from each site and treated with a combination of interventions so that participating households received one or more stoves under different conditions. Air quality measurements and household surveys were implemented in all households at the start, middle, and end of the study. Stove emissions measurements, kitchen performance tests (KPTs) and stove use monitors (SUMs) were deployed in a subset of households.
Conclusions:
Objective 1) Assess the acceptability and availability of stove technologies and fuels
Fourteen stoves were evaluated in controlled cooking tests (CCTs), and 4 were omitted from the subsequent interventions due to poor performance or lack of availability when the intervention began.
Willingness to pay (WTP) exercises carried out soon after the CCTs demonstrated that a 75-80% subsidy would ensure that the majority of participants would be willing to pay for the stove technologies we were offering. At the end of the CCTs and WTP tests, the team settled on eight stove models to offer in both locations plus a hybrid cooking/heating stove offered only in HP, where there is seasonal heating demand.
Objective 2) Experiment by offering stoves for free or at a subsidy and under varying social interactions to determine the impact of these factors on stove adoption rates and outcomes,
During recruitment, it became apparent that there was already significant penetration of liquid petroleum gas (LPG) and electric stoves in HP communities but very limited penetration in KA communities. Despite the presence of aspirational stoves, 96% of households (HHs) enrolled in the study reported that wood was their primary cooking fuel prior to our intervention. When initially offered a range of stoves, we found the majority of participants at all sites choose either LPG or electric stoves. A considerable minority selected other types of stoves.
Studying variation in baseline stove ownership and preliminary stove preference revealed that wealth and higher caste were significant predictors of pre-intervention ownership of non-solid fuel cooking options as well as preference for cleaner technologies offered through the intervention. The experimental treatments also influence preferences in some communities.
When given the opportunity to exchange, households in one region were more likely to choose solid fuel stoves (p < 0.05). Giving free stoves had mixed results; households in one region were more likely to select clean options (p < 0.05), but households in the other region appeared to prefer solid fuels (p < 0.10). When offered to exchange stoves in subsequent stove bazaars, nearly every household that did not initially select LPG selected it. By the end of the study, 86% of households in Karnataka had LPG. Among the households that were allowed to exchange, 98% had selected it.
Selecting LPG does not mean that HHs forego solid fuels. Using self-reported data, we find families used 2.0 cylinders during the first year and 3.4 cylinders during the second year of the project. HHs cooking exclusively with LPG consume 7-12 cylinders per year. Thus, the average HH in our study probably cooks 25-50% of their meals with LPG.
In Koppal, LPG refills among the beneficiaries of Pradhan Mantri Ujjwala Yojana (PMUY), India’s massive LPG subsidy program, is roughly 2.3 cylinders per year. This use rate is similar to self-reported refills in our intervention participants and less than half that of the general rural population, who used ~4.7 cylinders per household per year.
In addition, though some theories of technology adoption state that adopters increase the use of a new technology with experience, we found no observable increase in LPG consumption among rural consumers from one year to the next. Thus, at least in the near-term, the evidence indicates that stacking with LPG and polluting fuels persists for several years in this population. These results suggest that policy revisions are needed to encourage increased LPG consumption use are needed for both PMUY and general rural consumers in order to achieve health and environmental benefits.
Objective 3) Measure in situ impacts of stove adoption on air pollution, and climate-forcing
- Control versus Intervention HHs
We conducted a HH-level paired analysis to evaluate the effectiveness of the intervention in KA by calculating the difference in pollutant concentration in each HH between follow-ups (F1 and F2) and baseline (BL) measurements while controlling for inter-household variability in indoor air quality. We find PM2.5 decreases occurred in all groups between BL and both F1 and F2. However, decreases were larger in intervention HHs than control HHs in F1 only (p = 0.1). This indicates that the intervention was inconsistent, with a discernable improvement after roughly 1 year, but no improvement after the second year.
In HP, we were unable to collect baseline kitchen PM2.5 concentrations. However, we can compare the distributions of PM2.5 between control and intervention HHs during F1 and F2. We can also compare between HP and KA. We found the distribution of PM2.5 in study HHs in HP was significantly lower in intervention HHs than control HHs during F2, but not in F1, which was opposite of the pattern observed in KA.
- Impact of LPG use
We found PM2.5 concentrations were ~50% lower in HHs using LPG stoves relative to those without LPG. Based on survey data, we were also able to classify LPG use by exclusive, primary and secondary. During F1 in HP, PM2.5 was 78% and 59% lower in HHs exclusively using LPG compared to HHs with secondary and primary use respectively.
The trend is similar for F2 in HP and follow-ups in KA. We also see significant reduction in black carbon absorption (Bap) and black carbon to PM ratio (BC/PM2.5) for the exclusive LPG group compared to primary and secondary group in all follow-ups in both locations.
- Effect of chimneys
Chimneys also affect HAP concentrations and we found that chimneys featured in the study in several ways. First, several intervention stoves featured built-in chimneys. Of these, the most popular option was the Himanshu Tandoor (HT), a hybrid heating/cooking stove offered only in HP. We found mean indoor PM2.5 and BC absorption were 66% and 60% lower (p < 0.05) in HHs where HT was used exclusively relative to TT in F1. We do not have a sufficient sample (N= 2) to make a conclusion in F2. A comparison based on primary stove use also shows 46% lower mean indoor PM2.5 (p < 0.05) in HHs having HT as primary stoves relative to TT.
In addition, though chimney installation was not a part of the intervention in KA, roughly 60% of participating HHs had an inset hearth with a chimney in their kitchen, which resulted in lower PM2.5 concentrations. Mean PM2.5 was ~50% lower in HHs with a chimney than those without. Interestingly, HH with secondary stoves with a chimney showed more reduction in indoor PM2.5 and Bap relative to HH with a primary stove with a chimney.
- Inter-period and Inter-site variability
We also observed inter-period variability in indoor air quality. For example, in HP, Bap and BC/PM2.5 in HHs using traditional tandoor (TT) stoves exclusively were 50% and 69% higher respectively in F1 than F2. We see the same trend for traditional open woodfires stoves. Cooking characteristics and ventilation had the biggest influence on inter- site variability.
- Emission factors of biomass stoves
There was no consistent trend in PM2.5 emission factors (EF) among biomass stoves between the measurement periods. PM2.5 OC EF tracked PM2.5 closely, which is consistent with OC-dominated PM. CO EF showed a bit less stove-to-stove variability across measurement periods.
We tested TT and HT stoves in HP. TT EFs did not vary significantly between measurement periods; however, TTs had the highest intra-season variability in PM2.5 EFs driven by extremely high emitting events. The lack of significant difference between emissions from the TT and HT indicates that the improved HT design will not necessarily reduce atmospheric pollution, although the HT did lead to significantly lower kitchen concentrations of PM2.5.
The tests included three tests of the forced-draft TERI stove. STEMS measurements had minimal variability, but results need to be interpreted with caution because of the small sample size (N=3). The PM2.5 EF of TERI and Prakti (5.8 ± 0.2 and 6.3 ± 2.5 g kg-1) were comparable to PM2.5 EFs of rocket and gasifier stoves in a previous field study and were not significantly lower than the three stone fires (TSFs). A number of studies found similar results for alternative biomass stoves. HT stoves also did not show consistent reductions relative to TSF stoves. While, PM2.5 EF of HT stoves was 46% and 20% lower than that of TSF at BL (p = 0.02) and F2 (p = 0.05) respectively, it was 66% higher in F1 (p = 0.07). The trend of OC EF was similar to PM2.5 EF for these stoves consistent with the fact that PM2.5 was dominated by organic carbon in the emission (OC/PM: 40±10% and 37±10% for HT and TSF respectively). Other pollutant EFs (e.g., CO EF, EC EF) and properties (e.g., EC/TC, SSA) did not show any significant difference between these two stoves. The basic chimney stoves (TT) also had lower mean PM2.5 and OC EF than TSF at BL and F2, though the distributions were not significantly different. Chimney stoves were also different than TERI gasifier with regard to EC EF. For example, mean EC EF of TT was 3 times higher (p = 0.03) than TERI at F1. The same is true for the HT stove (p=0.08).
- In-use emissions from LPG stoves
LPG stoves had the lowest EFs for all pollutants across the study. Mean PM2.5 EF (1.8 ± 2.4 g kg-1) and CO EF (34.3 ± 23.1 g kg-1) from LPG tests were 76% and 60% lower respectively than biomass stove tests. In addition, LPG has a higher calorific value and the LPG stoves have better thermal efficiency than most biomass stoves. If we adjust EFs to account for these differences using published calorific values and stove efficiencies, the differences between LPG and biomass stoves increase. Mean PM2.5 and CO from LPG are 0.07 and 1.38 g MJd-1 respectively, which are 94-98% and 90-97% lower than biomass stoves.
In-home mean PM2.5 EFs and CO EFs in this study were ~36 and ~2 times higher respectively than lab EFs measured by Shen et al. Notably, the variability of in-home LPG EFs is similar to variability in biomass stoves, with higher variability in F2 than F1. Mean CO and PM2.5 EFs were also higher in F2 than F1 and the EF distributions were significantly different in some cases (p =0.001 for PM2.5 EF in HP and p = 0.02 for CO EF in KA). This difference might be attributed to degradation in stove performance with time considering 8- to 10 months passed between F1 and F2.
Objective 4) Model the impacts of widespread stove adoption on regional and global climate through a range of scenarios directly informed by data from the field.
The simulations show that when BC is not a site for ice nucleation (IN), global and regional solid fuel aerosol emissions have a net cooling impact of -141 ± 4 mWm-2 and -12 ± 4 mWm-2 respectively. Impacts are dominated by aerosol-indirect and semi-direct effects (AIE and SDE) from enhanced cloud condensation nuclei concentrations for the formation of liquid and mixed-phase clouds, and suppression of convective transport of water vapor from the lower to upper troposphere and lower stratosphere. However, when BC does behave as a source of IN, the net climate impacts of global and regional solid fuels are -51 ± 210 and 0.3 ± 29 mWm-2 respectively. Uncertainty is calculated from sensitivity simulations that alter the maximum freezing efficiency of BC across a plausible range (0.01-0.1). Thus, the overall sign and magnitude of forcing caused by solid fuel emissions globally, and from India, remains ambiguous pending improved understanding of the role BC plays in ice nucleation.
Journal Articles on this Report : 10 Displayed | Download in RIS Format
Other project views: | All 30 publications | 13 publications in selected types | All 13 journal articles |
---|
Type | Citation | ||
---|---|---|---|
|
Islam M, Wathore R, Zerriffi H, Marshall J, Bailis R, Grieshop A. In-use emissions from biomass and LPG stoves measured during a large, multi-year cookstove intervention study in Rural India. SCIENCE OF THE TOTAL ENVIRONMENT 2021;758(143698). |
R835421 (Final) |
Exit Exit |
|
Islam M, Wathore R, Zerriffi H, Marchall J, Bailis R, Grieshop A. Assessing the Effects of Stove use Patterns and Kitchen Chimneys on Indoor Air Quality during a Multiyear Cookstove Randomized Control Trial in Rural India. ENVIRONMENTAL SCIENCE & TECHNOLOGY 2022;45:150-158. |
R835421 (Final) |
Exit Exit |
|
Islam M, Meyestanio S, Saleh R, Grieshop A. Quantifying brown carbon light absorption in real-world biofuel combustion emissions. AEROSOL SCIENCE AND TECHNOLOGY 2022;56(6):502-516. |
R835421 (Final) |
Exit Exit |
|
Kar A, Pachauri S, Bailis R, Zerriffi H. Capital cost subsidies through India's Ujjwala cooking gas programme promote rapid adoption of liquefied petroleum gas but not regular use. NATURE ENERGY 2020;5(2):125-126. |
R835421 (Final) |
Exit Exit |
|
Singh D, Zerriffi H, Bailis R, LaMay V. Forest, farms and fuelwood:Measuring changes in fuelwood collection and consumption behavior from a clean cooking intervention. ENERGY FOR SUSTAINABLE DEVELOPMENT 2021;61:196-205. |
R835421 (Final) |
Exit Exit |
|
Singh D, Aung T and Zerriffi H. Resource Collection Polygons:A spatial analysis of woodfuel collection patterns. Energy for Sustainable Development 2018; 45:150-158. |
R835421 (Final) |
Exit Exit |
|
Jagadish A, Dwivedi P, McEntire KD and Chandar M. Agent-based modeling of “cleaner” cookstove adoption and woodfuel use:An integrative empirical approach. Forest Policy and Economics 2019; 106:101972. |
R835421 (Final) |
Exit Exit |
|
Menghwani V, Zerriffi H, Dwivedi P, Marshall JD, Grieshop A and Bailis R. Determinants of Cookstoves and Fuel Choice Among Rural Households in India. Ecohealth 2019; 16(1):21-60. |
R835421 (Final) |
Exit Exit |
|
Kar A, Pachauri S, Bailis R and Zerriffi H. Using sales data to assess cooking gas adoption and the impact of India’s Ujjwala programme in rural Karnataka. Nature Energy 2019; 4(9):806-814. |
R835421 (Final) |
Exit Exit |
|
Jagadish A and Dwivedi P. Deconstructing networks, unearthing consensus:Diffusion of “cleaner” cookstoves in rural Himalayas of India. Energy, Sustainability and Society 2019; 9(1). |
R835421 (Final) |
Exit Exit |
Supplemental Keywords:
AirProgress and Final Reports:
Original AbstractThe perspectives, information and conclusions conveyed in research project abstracts, progress reports, final reports, journal abstracts and journal publications convey the viewpoints of the principal investigator and may not represent the views and policies of ORD and EPA. Conclusions drawn by the principal investigators have not been reviewed by the Agency.
Project Research Results
- 2019 Progress Report
- 2018 Progress Report
- 2017 Progress Report
- 2016 Progress Report
- 2015 Progress Report
- 2014 Progress Report
- Original Abstract
13 journal articles for this project