Publications·November 30, 2007
The World Health Organization's 2004 Global and Regional Burden of Disease Report estimates that acute respiratory infections from indoor air pollution (pollution from burning wood, animal dung, and other bio-fuels) kill a million children annually in developing countries, inflicting a particularly heavy toll on poor families in South Asia and Africa. This paper reports on an experiment that studied the use of construction materials, space configurations, cooking locations, and household ventilation practices (use of doors and windows) as potentially-important determinants of indoor air pollution. Results from controlled experiments in Bangladesh are analyzed to test whether changes in these determinants can have significant effects on indoor air pollution. Analysis of the data shows, for example, that pollution from the cooking area diffuses into living spaces rapidly and completely. Furthermore, it is important to factor in the interaction between outdoor and indoor air pollution. Among fuels, seasonal conditions seem to affect the relative severity of pollution from wood, dung, and other biomass fuels. However, there is no ambiguity about their collective impact. All are far dirtier than clean fuels. The analysis concludes that if cooking with clean fuels is not possible, then building the kitchen with porous construction material and providing proper ventilation in cooking areas will yield a better indoor health environment.
1 Why the study matters
The WHO attributes over one million child deaths a year to indoor air pollution (IAP) from solid-fuel cooking, with South Asia among the hardest-hit regions. Prior national monitoring in Bangladesh showed household PM₁₀ levels frequently above 300 µg m⁻³, far exceeding safety guidelines
. Yet evidence on how house design, kitchen layout and fuel choices jointly shape exposure was scarce. This World Bank research team therefore built and instrumented full-scale replica houses in a peri-urban Bangladeshi village and ran 485 controlled cooking experiments to isolate the impact of structural and behavioural factors on particulate pollution.
Key words: indoor air pollution, PM₁₀, biomass fuels, cooking location, building materials, ventilation, seasonality, poverty, Bangladesh, controlled experiment
2 Experimental design
Site & dwellings. Four archetypal one-room houses—thatch, mud, tin and brick—were erected on a common axis to capture prevailing winds (north–south in winter, south–north in summer). Roofing could be swapped among thatch, tin and concrete to examine permeability effects. Kitchens were created in four configurations: within-dwelling, attached lean-to, detached shed and fully open-air
.
Fuels & seasons. Operators cooked with clean fuels (kerosene/LPG) and three biomass categories—wood, cow-dung cakes and “other” residues (rice husk, jute). Tests ran April 2005 – June 2006, skipping the monsoon, so data cover both the high-dust dry season (Nov–Mar) and the low-dust wet season (Apr–Jun / Oct)
.
Instrumentation. Each kitchen and living space housed MiniVol 24-h gravimetric samplers and pDR-1000 nephelometers logging PM₁₀ every two minutes; 76 ambient-air readings captured background variability
. The resulting dataset included 221–700 µg m⁻³ winter kitchen means versus 30–310 µg m⁻³ summer means, underscoring seasonal contrasts
.
3 Core findings
3.1 Kitchen exposure
Multivariate regressions revealed that seasonality dominates absolute concentrations: kitchen PM₁₀ constants averaged ≈156 µg m⁻³ in the dry season and 55 µg m⁻³ in the wet season
. After controlling for weather:
Layout matters—but not as expected. In the dusty season, indoor kitchens were cleaner (–187 µg m⁻³) than open or detached ones because walls filtered polluted outdoor air; attached lean-tos were intermediate (–49 µg m⁻³). The hierarchy reversed in the wet season when outdoor air was cleaner
.
Building materials moderate pollution. Brick walls added ≈60 µg m⁻³ and mud walls ≈34 µg m⁻³ relative to thatch/tin in the dry season; differences vanished or flipped in the wet season. Tin roofs yielded a small but significant –23 µg m⁻³ benefit over thatch in the dry months
.
Fuel choice is crucial. Compared with kerosene/LPG, wood and dung each raised PM₁₀ by ~70 µg m⁻³, and mixed residues by ~90 µg m⁻³ in the dry season. In the wet season dung became the dirtiest (+109 µg m⁻³)
. Real-time data confirmed that wood and dung also produced sharper pollution spikes than other biomasses
.
3.2 Living-space exposure
Because smoke diffuses rapidly, living-room PM₁₀ rose by 0.5–0.6 µg m⁻³ for every 1 µg m⁻³ increase in the adjoining kitchen, but only in the dry season; in the wet months the link weakened dramatically. Mud and tin walls actually lowered living-room pollution versus thatch, while ceiling-fans trimmed PM₁₀ by ~20 µg m⁻³ in dusty months but made no difference in the wet season.
3.3 Exposure patterns
pDR-1000 stats showed that attached kitchens produce the highest peak intensities, while detached kitchens create longer but flatter exposure curves. Building materials per se did not alter peak-versus-sustained loads
.
4 Interpretation & policy insights
Outdoor air is part of the indoor problem. When ambient PM₁₀ is high (dry season), walls shield occupants; when ambient air is cleaner (wet season) the same walls trap smoke. Any intervention must therefore couple household measures with community-level pollution control
.
Clean fuels remain the gold standard. All biomass fuels were “far dirtier than clean fuels,” whatever the season
. Yet affordability barriers mean incremental structural fixes offer near-term relief.
Design tweaks pay off. For households unable to switch fuels:
Use porous wall materials (thatch/tin) and tin roofs to enhance passive ventilation in dry months.
Site kitchens inside or attached during dusty months, but detached/open once rains arrive.
Install ceiling fans in living areas for dilution ventilation during the dry season.
Collective solutions multiply gains. Villages could:
Bulk-purchase cleaner fuels or efficient stoves;
Centralise cooking in well-vented facilities with tall stacks;
Rotate women in cooking roles to spread exposure risks;
Pair health education with micro-credit so male household heads—also heavily exposed—see the immediate benefits
.
5 Contribution to knowledge
This is the first rigorously monitored study to disentangle fuel, structure, layout and season in a low-income country setting. Its 18-month dataset provides actionable coefficients for exposure modelling and underscores why single-factor interventions (e.g., promoting chimneys without tackling ambient dust) may disappoint. The findings have already informed World Bank indoor-air strategies and underscore the need for integrated “health-energy-housing” programmes across South Asia.