Introduction

Globally, more than two billion people rely on solid biomass fuels (SBFs) to meet their daily residential energy needs, according to estimates from the World Health Organization. The dependence on SBFs is predominantly observed in lower- and middle-income countries (LMIC)1,2, and has considerable effects on health, environmental, and social welfare3. The persistent use of SBFs remains a major consequence of clean energy poverty4 and may hinder economic and human development5. In particular, in economically-challenged rural areas, households primarily use low-cost SBFs like dung cakes, firewood, chips, dried leaves, and agricultural residues for their daily cooking and heating requirements6,7,8,9. The widespread use of SBFs for cooking and heating significantly increases indoor air pollution at the household-level and aggravates outdoor pollution levels10,11. In India, the National Family Health Survey (NFHS-5, 2019–2020) reports that around 56% of rural households and 9.5% of urban residents use SBFs for daily cooking, leading to substantial indoor and outdoor air pollution. In 2017, air pollution was linked to 1.24 million deaths in India12, including 0.67 million from ambient particulate matter and 0.48 million from household air pollution.

Burning SBF can significantly contribute to the release of gaseous pollutants, including CO, NOx, and particulate matter (PM2.5). A significant and climatically relevant fraction of PM2.5 consists of light-absorbing black carbon (BC), recognized for its potent warming potential and health effects13,14,15,16. In fresh SBF-based cooking emissions, the BC to PM2.5 ratio could range between 17–41%17. Furthermore, BC has been identified as an important indicator of the health impact of combustion-related emissions18,19 and therefore can be prioritized in the design of pollution reduction strategies, particularly in areas highly dependent on SBF, due to its strong association with domestic solid fuel consumption20,21. Recent studies highlight BC’s ability to transport various toxic chemicals, establishing a robust epidemiological connection between BC-enriched aerosol exposure and increased risks of cardiovascular disease and mortality22,23. It is increasingly believed that domestic usage of SBFs poses a significant health risk, especially to women and children who are exposed to BC more frequently and over longer periods compared to other family members24,25,26. Furthermore, once released, BC-enriched aerosols affect both regional and global atmospheric radiation budgets15. Specifically in India, BC is recognized as a factor that exacerbates extreme high-temperature events and altering monsoon cycles27,28.

In rural households, several factors have been identified as influencing indoor BC concentrations during daily cooking. These can include cooking duration, the number of meals prepared, fuel type (such as dung cakes, firewood, dried leaves, or agricultural waste), type of stove and it’s efficiency, cooking practices (e.g. ignition process and timing), and kitchen ventilation efficiency13,29,30,31,32,33. Indoor BC concentrations in rural kitchens using SBFs have been reported at extremely high levels: 1-min averages of 600–1600 μg m−3 in Nepal34, peaks of 1196 μg m−3 in Punjab, India13. Filter-based indoor measurements during cooking show values ranging from 520 to 1530 μg m−3 in South India35 and up to 2770 μg m−3 in Nepal17. Studies underscore the potential use of treated wood, including waste furniture, as a fuel source36, but also highlight the risk of heavy metal-enriched particulate emissions37, traces of which have been observed in smoke residues within rural Indian kitchens38. Furthermore, previous research indicates that local and regional variations in BC are often shaped by land use patterns39 and meteorological conditions, such as the wind speed and height of the planetary boundary layer40,41,42. As a result, BC exposure concentrations for different gender or age groups can fluctuate significantly within a day or over several days, depending on the time spent in different microenvironments21. In megacities and urban regions, BC emissions are predominantly influenced by diesel traffic and industrial operations22, compared to rural areas where SBF combustion dominates. In addition to cooking, rural households in the Indo-Gangetic Plain (IGP) region use SBF extensively for water and space heating and by security personnel, especially in winter43. As a result, wintertime BC concentration variability can worsen significantly due to increased emissions from SBF-based space heating activities43,44, the local and remote transport of pollutants from agricultural fires40,45,46,47, and stable atmospheric conditions that reduce pollutant dispersion41,48,49,50. The rise in winter emissions attributable to space heating aligns with patterns observed in other parts of the world, including regions in Africa51, South America6, and North America52.

A recent model-based study22 on the IGP region emphasizes the potential benefits of mitigating BC exposure, suggesting that a reduction in current BC emissions by 87% could prevent 400,000 premature deaths annually through strategies that include the transition from SBF to cleaner alternatives, such as liquefied petroleum gas (LPG). In India, efforts to shift towards cleaner fuels, such as the distribution of LPG, have been implemented through flagship initiatives such as the Pradhan Mantri Ujjawala Yojana (PMUY) scheme. Despite significant adoption, consistent use of LPG in rural and informal urban areas is hampered by the economic instability of households, the availability of biomass, and common fuel stacking practices, where households use both SBF and LPG53,54. Fuel stacking serves as a means to manage fluctuating income levels and protect against unpredictable markets, while also maintaining traditional customs and partially benefitting from contemporary fuel sources55. Collectively, these socioeconomic and personal factors can also impact BC exposure and alter the anticipated health advantages of different intervention strategies for reducing exposure, on the individual and community-scale.

Given the lack of extensive ambient monitoring infrastructure in rural regions, personal exposure monitoring approaches have become especially important for evaluating household-level pollution. Micro-aethalometers, often used for personal or micro-environmental exposure evaluation56,57,58, offer a promising approach for monitoring BC in residences and community areas due to their portability, battery-powered operation, affordability, and ability to deliver high-resolution real-time data59,60. These devices are especially beneficial in rural and remote locations where limitations such as intermittent electricity can hinder traditional data collection61. Concurrent assessments of BC concentrations indoors and outdoors, paired with data on household layout, daily habits, and demographic and behavioral patterns, can enhance comprehension of exposure variability and its drivers. This aids in formulating targeted, feasible strategies to lower indoor BC exposure at the household level, aligning with broader policy efforts.

This research aims to measure the diurnal variations in BC exposure concentrations in rural areas both indoors and outdoors, and to explore how local meteorological conditions and types of kitchen affect the concentrations. It reports on the findings of a field campaign carried out during the winter of 2019 in two villages located in the central part of the IGP in Unnao district in Uttar Pradesh. Using portable and battery-operated micro-aethalometers, the study documents variations in BC concentration across different household kitchen designs, including both indoor and outdoor semi-enclosed kitchens that are typical of rural houses in the IGP region. Furthermore, data were collected through surveys to investigate daily cooking habits, space heating requirements, fuel consumption patterns, and socioeconomic conditions, which were then used to help interpret and examine BC concentrations observed. Based on these findings, the study proposes community-level intervention strategies to reduce the health hazards associated with BC exposure.

Methods

Study location and field campaign overview

BC measurements were carried out during a winter field campaign (9 November–12 December 2019) in two rural villages in Unnao district, Uttar Pradesh, India: (1) Bhawani Kheda (Village A) and (2) Naikani Kheda (Village B). Information about monitoring days is provided in the Supplementary Information (Fig. S1). This period coincides with the peak winter season in the Indo-Gangetic Plain (IGP), when BC emissions are typically elevated due to increased space heating, post-harvest agricultural burning, and stable atmospheric conditions that limit pollutant dispersion48,62,63. These villages are located approximately 20 km and 60 km from the major cities of Kanpur and Lucknow, respectively. The villages were selected based on three main criteria:

  1. (i)

    High reliance on SBFs for cooking and heating,

  2. (ii)

    Minimal impact from main roads or industrial sources (each village is approximately 4 km away from the nearest main road), and

  3. (iii)

    Logistical feasibility for long-term equipment installation.

The monitoring locations are indicated on the Unnao district map (see Fig. S2) with detailed village-level layout information. Both villages are small (fewer than 100 households) and predominantly surrounded by agricultural land. Some households in these villages had connections to solar microgrids that supplied essential electricity for lighting (usually one bulb) and charging mobiles.

In each village, indoor BC measurements were carried out in one household and outdoor measurements used weather-proof enclosures positioned approximately 3 meters above ground, representing the village-level BC concentration. Since only three devices were available for four monitoring sites, their deployment was rotated according to power availability and battery recharging needs. Household selection for indoor monitoring was guided by an initial site survey that focused on access to solar microgrids, the presence of backup power, and the appropriate conditions for equipment setup. To minimize disruption, neither dilution nor structural changes were implemented; devices were securely installed indoors, preserving the household’s indoor environment while effectively measuring typical SBF exposure. Despite challenges posed by intermittent power supply, auxiliary charging and regular maintenance were conducted at the local partner’s office (Boond Engineering and Development Pvt. Ltd.) to ensure continuous data collection.

Data collection

Black carbon monitoring

Monitoring BC in rural households in real-time poses challenges such as limited space, power interruptions, and sudden spikes in concentration during cooking. One suitable instrument for such conditions are compact and battery-powered micro-aethalometer, MA300. Three portable, battery-operated micro-aethalometers (Model: MA300, AethLabs, San Francisco, California) were used for BC monitoring. The MA300s have a small footprint, do not disrupt daily cooking activities, and can measure BC concentrations ranging from \(0\,\hbox {to}\, 1\,\hbox {mg}\) . In literature, BC measurements using a filter absorption photometer are commonly referred to as equivalent black carbon or eBC64. To ensure simplicity and broad interdisciplinary applicability of this study, we opted to refer to concentrations as BC instead of eBC. These devices use a filter-based dual-spot light attenuation measurement technique, and they can simultaneously evaluate real-time aerosol light absorption at five different wavelengths (375, 470, 528, 625, and 880 nm). The dual-spot filter system employs two filter spots with differential aerosol loading to calculate real-time light attenuation measurements compared to a reference (or clean) spot. This method integrates enhanced and non-linear correction mechanisms for the loading and scattering artifacts of the aethalometer59,65,66. The MA300s were operated at a flow rate of 150 mL min−1 and attenuation (ATN) limit was set to 100, with data collected at 5-min intervals. In this study, we implemented the data correction methods specific to MA300 from the three devices as described in Chakraborty et. al.59. The absorption of light at 880 nm was used for the estimation of the BC mass concentration67,68. UV measurements from the MA300 were prone to instrumental drifts, often resulting in underestimation59. Consequently, UV light absorption was not utilized for biomass smoke characterization in this study69,70.

Meteorological data products

We obtained hourly mesoscale (regional) meteorological data, specifically wind speed and planetary boundary layer height (PBLH), for the study area from the NASA meteorological reanalysis dataset MERRA271. MERRA2 offers data at a spatial resolution of \(0.5^{\circ } \times 0.625^{\circ }\) (equivalent to \(55\,\hbox {km} \times 60\,\hbox {km}\) ) and provides hourly temporal resolution, with both villages in the study found to be in the same grid cell. the meteorological data acquired from MERRA2, we calculated the hourly ventilation coefficient (VC). VC is calculated as the product of PBLH and mean wind speed41,72. This coefficient serves as a measure of the atmospheric capacity to dilute and disperse ambient air pollutants in a given region. In the past, Sujatha et. al.73 and Prakash et. al.48 have observed a negative correlation between monthly variations in VC and ambient concentration of BC in Hyderabad, a city in southern India, and in Hamirpur, a rural area in the IGP, respectively.

Household survey

In addition to collecting pollutant measurements and meteorological variables, a household survey was conducted during the BC monitoring period (Section S.1) in both villages to investigate household energy use patterns, cooking arrangements and behaviors, household layout and economic and demographic characteristics. The survey was carried out in the local language (Hindi) and consisted of individual interviews with 27 households from both villages, conducted once throughout the campaign (protocol was approved by UBC Behavioural Research Ethics Board under H19-02448). The protocol involving humans was performed in accordance with institutional guidelines and regulations. Key data obtained from the survey included the primary cooking fuel used, the size of the household, the characteristics of the cooking stoves, the kitchen facilities, and the daily cooking routines. Since this study focuses on two specific households, one from each village, our analysis will be based primarily on the findings obtained from these households. The detailed survey results are presented separately in the supplementary information (Table S1) to provide a comprehensive overview of the responses collected from the community.

Data processing

BC monitoring data from the MA300 units were adjusted according to the details provided in Section: “Black carbon monitoring”, following a procedure described in a study by Chakraborty et. al.59. In summary, the calibration of the three MA300 units (MA300A, MA300B, and MA300C) was carried out against a regulatory grade AE33 aethalometer in Vancouver during a 14-week co-location campaign that included periods with and without wildfires. Additionally, following Chakraborty et. al.59, any readings below the identified detection limit of 0.163 μg m−3were discarded. As detailed in Sections “Black carbon monitoring” and “Meteorological data products”, the MA300 system collects data every 5 min, while the MEERAv2 products provide data on an hourly basis. Therefore, to maintain consistency among the various data sources used in this research, the BC concentrations are reported as hourly averages.

During the study period, the MA300 devices were at times subjected to brief yet intense exposure to elevated levels of pollutants (> 900 μg m−3), especially during cooking activities. This led to technical problems like clogging of filter tapes (identified by a “Tape Jammed” status) and unreliable data. Any data points marked as “Tape Jammed” were omitted from the analysis. Additionally, data from the MA300 can drift59. To remove data points affected by sensor outliers, data have been identified as anomalies and have been eliminated by excluding values above \(q0.75 + 1.5\times IQR\) or below q0.75–1.5 × IQR , where q0.25 and q0.75 represent the first and third quartiles, and the interquartile range (IQR) is the difference between the third and first quartiles.

The data filtering and cleaning process led to gaps in the campaign data, with approximately 17% of the 5-min resolution data points being removed. In the supplementary information, we have included a calendar plot (Fig. S1) showing the final data products specific to each location that were used for the analysis. Moreover, for instances where indoor and outdoor measurements were taken concurrently (primarily in Village B with around 180 h of data), we conducted distinct analyses to examine the pollutant exchange dynamics between indoor and outdoor settings. All data analysis was carried out using R (version 4.2.2) and a collection of open-source packages (TidyVerse74, OpenAir75, rstatix76). The maps in Fig. S2(1,2) were created using QGIS (Version 3.40.5 “Bratislava”) and OpenStreetMap data under OBDL, and Fig. S2(3,4) were used from Google Maps with attribution.

Results

Village-level survey results

The household survey encompassed 235 questions, structured based on previous research on access and usage of household energy77,78. This work focuses solely on variables relevant to residential combustion activities, including the type of cooking fuel, kitchen design, building structure, and user perceptions. The extensive hour-long survey shed light on economic status, household energy use, various kitchen types, fuel preferences, cooking behavior, and dietary choices. Comprehensive survey outcomes are compiled and detailed in the supplementary information (Table S1).

Our survey found that 55% of the surveyed households primarily engage in agricultural practices, which involve activities such as cultivating on their own or leased land. In contrast, 44.5% of the households derive their income from non-agricultural occupations, which can range from daily labor to positions in private companies or shopkeeping. Additionally, a significant proportion (74%) reported owning livestock, which includes animals like buffalo, cows, cattle, goats, and chickens. Previous large-scale surveys have shown that households involved in land cultivation and livestock ownership tend to prefer SBF for cooking, due to convenient access to the raw materials needed for SBF preparation, such as leaves, agricultural residues, wood, and dung79.

A substantial majority of households (85%) reside in “Pucca” buildings (structures constructed with brick and mortar), with various types of kitchens. All of these households have naturally ventilated living spaces. Following the categorization outlined by Sidhu et. al.80, the household kitchens were classified into five types: (1) Indoor Kitchen (IK), (2) Indoor Kitchen with Partition (IKP), (3) Outdoor Semi-Open Air Kitchen (OSK), (4) Outdoor Enclosed Kitchen (OEK), and (5) Outdoor Open-Air Kitchen (OK). Among these, OEK was identified as the most prevalent, accounting for 52% of the surveyed households, followed by the IK type, representing 18.5%.

Our findings indicate that out of the 27 surveyed households, 22 possessed LPG connections (81%). Among the LPG connection holders, 59% of them were beneficiaries of the PMUY program. On average, households with PMUY connections paid 10.5 USD, while non-PMUY scheme beneficiaries paid 72.2 USD for the LPG connection which includes the oven and a startup cylinder. These figures are based on the 2019 exchange rate of 1 INR = 0.0142 USD. Within the group of households with LPG connections, those with PMUY connections reported significantly lower monthly expenditures compared to households without PMUY subscriptions. These households had a median family size of five members and a monthly expenditure of 85 USD (equivalent to INR 6000). Research from rural India indicates that, on average, a large LPG cylinder constitutes approximately 10% of the households’ total monthly expenses81. Despite having LPG connections, 81% of these households reported using SBF, such as dung cakes or firewood, as their primary cooking fuel. This suggests a clear practice of fuel stacking in the study area, where rural households continue to use SBF even when they have LPG connections81. It was observed that rural communities tend to prefer the flavor imparted to bread and dishes when cooked over open flames, which gives them a smoky, charred taste. The findings of the survey responses of the 27 households indicate that the main reasons for stopping the use of LPG stoves are dissatisfaction with the taste of the food and the added financial pressure.These reasons align with those documented in previous research78,81,82.

Households with traditional cooking setups that use SBF often rely on various types and combinations of ignition fuels. In Fig. 1, we present the different types of ignition fuels used by the survey respondents for their daily cooking needs with SBF. Around 60% of households utilized “Dried Leaves”, with Kerosene (47%) being the next most common ignition fuel. Over the years, the Indian government has attempted to curb kerosene usage (and the related subsidies) by progressively reducing its availability in the “Public Distribution System”83. Removing subsidies on kerosene could pose challenges, especially since it is still the cheapest option for lighting for many people in India who do not have access to electricity84. As shown in Fig. 1, these ignition fuels are easily accessible and generally require little effort and cost to acquire. Once ignited, pollutants such as PM2.5 concentration can rapidly rise because these ignition fuels often have high emission factors due to the lower initial combustion temperature and following an incomplete combustion process85,86,87. The practice of using various fuels for ignition, such as SBFs and kerosene for everyday cooking, perpetuates the dependence on SBF by encouraging fuel stacking among economically disadvantaged households.

In terms of daily cooking routines, the surveyed households generally prepare two meals each day, a trend frequently noted in prior research78,88. Participants were asked to recall and reflect on their daily cooking times. Most households (26 out of 27) reported cooking in the morning between 5:00 AM and 9:00 AM, and in the evening between 5:00 PM and 9:00 PM. The broad range in daily cooking times reported can be attributed to intra-household differences, as observed in India by Pillarisetti et. al.33, and inter-household variations within the community, as seen in Guatemala by Ruiz-Mercado et. al.89. The range and overlaps in the cooking periods can ultimately affect the local levels of outdoor pollutant concentrations and may lead to exposure crossovers.

We also investigated energy usage activities unrelated to cooking, such as heating spaces or water, and lighting. The majority of households (25 out of 27) are linked to local off-grid solar energy systems (micro-grid), which are seen as a cost-effective and environmentally friendly way to meet basic electricity demands77. However, out of these 25, only nine households reported using the micro-grid as their main source of lighting. As previously identified77, the implementation of micro-grids can lead to a decrease in the monthly usage of kerosene fuel, a substitute fuel for lighting that can release a substantial amount of BC into the indoor atmosphere43. We found that 22% of the households, or six in total, reported using kerosene as a secondary lighting source. On average, households reported that they needed to heat their home for 11.5 weeks during the winter season each year. Almost all households (25 out of 27) reported using dung cakes and firewood to meet their space heating needs. Most households preferred outdoor space heating setups, such as open burning of firewood or other SBFs in bonfires, either individually or shared among multiple homes. All these activities can contribute to indoor and outdoor BC at different times of the day.

Fig. 1
figure 1

Households employ a variety of ignition fuels in their traditional cooking arrangements when using Solid Biomass Fuels (SBF). This practice illustrates the widespread adoption of multi-fuel stacking among these households.

Full size image
Fig. 2
figure 2

Household A and Household B: Building structure, location of cooking oven (both traditional oven and LPG oven) and pollution monitor have been identified.

Full size image

Description of households with BC monitoring

We selected two specific households for investigation, referred to as Household A and Household B. The building plans for these households (Fig. 2) indicate the placement of the monitoring equipment, the location of cooking ovens, and the presence of lateral openings, such as doors, windows, or passages. Household A comprises a small family of two adult members, engaged in agricultural work. Their residence includes a single room, which also serves as an indoor kitchen. In contrast, Household B accommodates five adults and two children and uses an Outdoor Enclosed Kitchen (OEK) according to the categorization by Sidhu et. al.80. We assumed a well-mixed condition within the zone of pollutant exposure, defined as the area where emissions are present and household members spend most of their time indoors. Both of these households are LPG subscribers, however, they primarily use dung cakes as their cooking fuel.

Overall black carbon concentration

A summary of the hourly BC concentrations throughout the study at the four sites are presented in Table 1. The mean hourly concentration of BC during the study period in both indoor and outdoor sites in the villages ranged from 15 to 17 μg m−3. The minimum hourly mean concentrations in the villages (outdoor) were less than 1 μg m−3, and were four to five times higher (4.2–5.7 μg m−3) within the households, indicating differences in BC concentrations between indoor and outdoor environments. The maximum concentration in all locations ranged between 33 and 38 μg m−3. The observed household BC concentrations were consistent with the indoor measurements recorded by Arif et. al.90 in various households that used traditional cooking methods (SBF users) across the IGP during the winter season (25 ± 5 μg m−3). The ambient BC concentrations in the rural areas in Punjab were reported by Ravindra et. al.13 to vary between 0.94 and 25.21 μg m−3, while the indoor concentrations fluctuated depending on the types of kitchen and the existence of chimneys. In rural areas in Kanpur, Suresh et. al.30 reported that BC concentrations in kitchens ranged from 5.1 to 5.3 μg m−3 for forced draft cookstoves, which is lower than the concentration reported in this study using traditional cookstoves. These findings underscore that variations in BC concentrations at the household level can be influenced by multiple factors, including the type of kitchen, the type of cookstove used, and the availability of kitchen ventilation.

Table 1 Statistical summary of BC concentration (in μg m−3), coefficient of variation (CV in %), and primary cooking fuel used at each location during the study period.
Full size table

We performed a non-parametric Wilcoxon rank sum test on the collected unpaired hourly BC measurements from the two households and villages to evaluate statistical differences. Significant statistical differences (p\(\ll\)0.05) were observed in the median concentration, with Household A having a higher concentration than Household B, and Village A showing a higher concentration than Village B. Village-level concentration differences are likely due to variations in monitoring site locations and the influence of local emission sources, as further discussed in the analysis of diurnal trends (see Section “Assessment of diurnal variation of BC”). Such hyperlocal variability in BC has also been demonstrated in urban environments91.To gauge the degree of variability in the observed BC concentration at a given location, we employed the coefficient of variation (CV), which is expressed as a percentage and calculated by dividing the standard deviation by the mean. The computed CV (Table 1) for the hourly BC concentration in both villages indicated a variation 50% throughout the duration of the study. Although a similar degree of variability was observed outdoors in both villages, the indoor BC concentrations showed different variations. Household A, with its fully enclosed indoor kitchen, showed the least variability (CV = 40%), which can be attributed to a kitchen setup that limits the dispersion of pollutants from the source, resulting in a decrease in pollutant variability. However, Household B, which has an enclosed outdoor kitchen, showed a level of variability similar to the outdoor village measurements (CV = 50%), implying an enhanced exchange of pollutants between indoor and outdoor settings due to the kitchen design. Previous research has indicated that the dilution of pollutants increases in households with outdoor enclosed kitchens (OEK), resulting in a reduction in pollutant concentrations and variability13,80,86.

Activity-specific BC estimates

In rural households, significant energy is required for daily activities such as cooking, space heating, and water heating, often exceeding other energy sources like grid or microgrid electricity and LPG, leading to the use of SBFs81,92. As mentioned in Section: “Village-level survey results”, the households examined in this study frequently utilize SBFs for various daily tasks, resulting in fluctuations in BC concentrations. In India, rural communities often opt for small open outdoor fires or bonfires, especially during the winter season, to fulfill space heating needs93. However, in our study, space heating activity was not tracked due to the absence of a set schedule or location and its dependence on regional outdoor temperature, which decreased significantly from December onward (Fig. S3). We noticed a rise in the proportion of outdoor BC during the coldest temperature range in December (Fig. S4). Our results indicated that on seven out of the ten measurement days in December, 40% of the daily outdoor BC concentrations were linked to the coldest temperature range (range: \(8 \,\hbox {and}\, 14\,^{\circ }\hbox {C}\)) of the day. This result underscores the impact of increased emission activity during the colder temperatures.

To understand the primary source of BC relative to the daily routines of households, we mapped BC concentrations to activity periods reported in the survey. About 90% of the households surveyed reported cooking twice a day, allowing us to identify two separate cooking times. Based on the survey findings, we established three distinct periods related to cooking activities for our study: morning cooking (5:00–9:00 AM), evening cooking (5:00–9:00 PM), and no cooking hours, covering the rest of the day. In Fig. 3, we present box plots of BC concentration segregated by activity for each of the sampling locations.

Fig. 3
figure 3

Hourly BC concentration boxplots for various home emission activities (Morning Cooking, Evening Cooking, and No Cooking) for each measurement location: Village A, Household A, Village B, Household B, and collectively as All Locations.

Full size image

Supplementary Table S2 provides details on activity-specific BC concentrations and their variability. The Kruskal-Wallis rank sum test results showed a statistically significant difference (\(\chi ^2\) = 219.19, p \(\ll\) 0.05) in BC concentrations across the three activity periods. In households A and B, the mean BC concentrations during cooking times (both morning and evening) ranged from 21 to 23 μg m−3 (standard deviation: 6–8.5 μg m−3). The BC concentrations during cooking reported in our study are generally in line with previous studies, though there are some discrepancies. For example, Aurora et. al.20 reported BC concentrations of 60.3 ± 15.4 μg m−3 during summer cooking periods in the hilly regions of northern India. Patange et. al.94 reported BC concentrations of \(91.7 \pm 52.9\,\upmu {\textrm{g}}\,\hbox {m}^{-3}\) in households that used traditional stoves during winter, while Suresh et. al.30 noted lower BC concentrations during cooking (\(7.8 \pm 9.1\,\upmu \textrm{g}\,\hbox {m}^{-3}\)) for households that used traditional cooking ovens in summer. These studies generally located their sampling sites close to the cooking stoves, evaluating the conditions in the area where the cook breathes, thus documenting higher concentrations compared to our study, except for Suresh et. al.30 who mainly researched semi-open kitchens. We placed our sampling devices to gauge the concentration specific to the dominant exposure zone (either the kitchen or living area) within the households (the regions highlighted in Fig. 2).

Variations in BC concentrations during cooking and non-cooking periods may be indicative of differences in fuel usage patterns, activity patterns, and ventilation. Across all locations, the least variability was noted during evening cooking sessions (CV = 30%), representing a decrease of about 12% compared to morning sessions (CV = 42%). This difference between morning and evening cooking is likely influenced by local weather conditions, as discussed in Section: “Assessment of diurnal variation in BC”. Another factor that may contribute to variability is the fuel usage pattern for each cooking session. Households generally use a range of SBFs, starting with different ignition fuels (Fig. 1) and then transitioning to the main cooking fuel, such as dung cakes, firewood, agricultural waste, or a combination of fuels. These have been shown to affect overall pollutant emissions in previous studies6,20,80. Additionally, SBFs undergo a multi-stage combustion process, including ignition, flaming, and smoldering, which impacts the rate of pollutant emissions during a cooking session30,85. The comparative analysis of activity-specific variability in BC concentrations between households shows that Household B exhibited greater variability (CV = 38%, 32%, and 47% for morning, evening, and non-cooking periods) compared to Household A (CV = 30%, 26%, and 26% for morning, evening, and non-cooking periods). The observed differences in variability between the households can be attributed to variations in kitchen layout and openness of the buildings, which significantly affect the efficiency of ventilation and therefore impact household pollutant concentration13,80,86. The observed outdoor BC concentrations during the non-cooking period (Fig. 2) probably result from a combination of residual emissions from smoldering after cooking hours, outdoor biomass burning activities from heating needs, and regional air mass transport, as evidenced by similar observations at the rural IGP_CARE site (Fig. S2) in Hamirpur in Uttar Pradesh48.

Fig. 4
figure 4

Diurnal variation in the mean hourly BC concentration observed at location A (Village A and Household A) and location B (Village B and Household B) throughout the study period. Vertical error bars indicate the 95% confidence intervals calculated for each hour.

Full size image

Assessment of diurnal variation of BC

The hourly mean diurnal variation of BC for the four locations of the entire campaign is shown in Fig. 4. With the beginning of daily cooking activities in households, an increase in indoor BC concentrations was observed, but peak concentration patterns and decay profiles differed (Fig. 4). For household A, a plateau peak is observed during both morning and evening cooking times. In contrast, Household B exhibited sharp peaks in its diurnal BC concentration patterns. The plateau nature of the peak typically indicates that the emitted BC tend to remain indoors because of restricted airflow, while a sharp peak implies effective removal of pollutants from faster air movement. Furthermore, the exponential decay rate of BC after daily cooking was higher in Household B (\(0.18\,\hbox {h}^{-1}\) in the morning and \(0.15\,\hbox {h}^{-1}\) in the evening) compared to Household A (\(0.16\,\hbox {h}^{-1}\) in the morning and \(0.11\,\hbox {h}^{-1}\) in the evening). Details for the calculations can be found in the supplemental information (Section S.6). Both households exhibited pronounced decay patterns of BC in the morning compared to the evening. The distinction between daytime and nighttime exponential decay patterns has been previously observed by Rehman et. al.95 in northern Indian villages. The observed sharp exponential decay patterns and notable variations in diurnal BC concentrations in Household B can be related to its daily cooking requirements for a larger family of seven members and the presence of an outdoor enclosed kitchen, as opposed to Household A, which has only two family members and an indoor kitchen. Previous research has highlighted the influence of family size and kitchen ventilation conditions on the levels of indoor air pollutants13,86,96. In Household A, the reduced production of pollutants is counterbalanced by their accumulation due to a restricted kitchen space, resulting in higher concentrations of indoor pollutants.

The hourly diurnal variation patterns in both villages closely matched the reported cooking activities within the households, highlighting the strong impact of cooking smoke on local air quality. The range of hourly diurnal concentrations in Village A (8–\(26.1\,\upmu \textrm{g}\,\hbox {m}^{-3}\)) exceeded that of Village B (7.6–20.0 μg m−3), particularly during and after the evening time cooking period. The highest hourly BC concentrations in the diurnal cycle occurred at 8:00 AM and 7:00 PM in both villages, showing a delay of one hour compared to the peak times observed in household activities (at 7:00 AM and 6:00 PM) centered around cooking (Section: “Village-level survey results”), suggesting the influence of indoor emission on outdoor BC concentration. The period following evening cooking activities (from 8:00 PM to midnight) showed the most pronounced differences between Village A and Village B, with a statistically significant difference (p \(\ll\) 0.05 from unpaired Wilcoxon test). The observed differences in pollutant concentrations between villages can be attributed to localized emissions in the evening, and the positioning of air monitoring stations (supplementary Fig. S2). In Village A, air quality monitoring was carried out in the central part of the village, surrounded by multiple residences and a grocery store, which likely influenced the readings due to nearby emissions activities at night. In contrast, the air quality monitoring station in Village B was located on the periphery of the village, partially surrounded by agricultural fields, which reduced the chances of influence from proximate outdoor active combustion sources, with the exception of residential emissions.

Fig. 5
figure 5

Relative change (%) in hourly concentration in a day with respect to mean concentration between 1:00 AM and 4:00 AM in different measurement locations. Shaded regions refer to the cooking emission period. Blue bars indicate a positive change in concentration, and red bars indicate a negative change.

Full size image

Since most significant human combustion activities do not occur after midnight, specifically between 1:00 AM and 4:00 AM, this period can serve as a benchmark to assess the relative temporal patterns of diurnal BC compared to inactive combustion times.

Figure 5 presents the hourly variation in BC concentration relative to the mean late-night baseline concentrations (between 1:00 AM and 4:00 AM). Figure 5 shows that BC concentrations from late morning to afternoon (11:00 AM–4:00 PM) are generally lower compared to the late-night baseline. This reduction in afternoon concentrations has been consistently noted in previous research and is attributed to shifts in local meteorological conditions41,95,97. A significant relative difference exceeding 45% in village-level BC concentrations was recorded, with lowest concentration occurring between 1:00 and 3:00 PM, coinciding with the most pronounced effects of local meteorology on outdoor air (Fig. S7). Several studies41,42,72,73 used atmospheric ventilation coefficient (VC) to explain seasonal and diurnal changes in surface-level pollutant concentrations. Throughout the study period, the calculated VC showed a diurnal hourly variation (supplementary Section S.8) ranging from 483 to 2366 m2 s−1, consistent with winter diurnal profiles previously described in Gani et. al.41. The afternoon decrease in BC concentrations aligns with higher VC readings, suggesting improved pollutant dispersion due to the effects of a higher planetary boundary layer and increased wind speeds.

The Pearson correlation analysis between diurnal VC and BC concentration shows a consistent negative correlation at all study sites. The correlation coefficients were −0.37, −0.45, −0.48, and −0.72 for Household A, Household B, Village A, and Village B, respectively, indicating that a rise in VC corresponds to a fall in BC concentration. Given that the spatial resolution of MERRA is \(55\, \hbox {km}\) by \(60\, \hbox {km}\), and both Village A and B are within the same MERRA grid cell, the differences in correlation strengths between these villages could be due to variations in local outdoor emission sources. In contrast, the BC concentrations in households show a moderate linear relationship with the diurnal VC trends, with the weakest correlation seen in Household A. This indicates that the influence of VC on BC concentrations is statistically insignificant in Household A (p=0.08), where indoor cooking limits exposure to meteorological conditions. However, a statistically significant correlation between diurnal BC concentrations and atmospheric ventilation is observed in all other locations.

Implications

Human exposure to pollutants during regular daily activities, termed personal or individual exposure98, requires comprehensive data on micro-environmental pollutant concentrations and the duration spent in various micro-environments due to spatial and temporal variations in pollutant concentrations99,100. In rural India, the residential sector is a major contributor to personal exposure to PM2.5 due to the continued dependence on SBF1,21 for daily household activities. These fuels are also key sources of BC, impacting climate68,101 and health22,102 both locally and globally103,104. Therefore, BC serves as a strong marker for the design of pollutant mitigation approaches in areas heavily reliant on SBFs, given its strong association with the household’s fuel and energy consumption. Consequently, reducing BC emissions could offer immediate advantages by decelerating near-term climate warming and achieving public health objectives105.

Extensive research has explored daily exposure to BC from traditional SBF-dependent cooking methods, focusing on factors such as stove types, kitchen designs, fuel modifications, and chimney use13,20,31,90,94,106. Yet, continuous monitoring data from both indoor and outdoor environments across extended periods are limited. This study successfully used portable micro-aethalometers to measure BC concentration with precise temporal detail over extended periods covering indoor and outdoor environments, documenting the spatial and temporal fluctuations in BC concentrations and associating these variations with everyday household or community activities. In particular in India, where air quality monitoring is sparse, especially in rural areas107, devices such as the micro-aethalometer can provide comprehensive data on BC concentrations, thereby enhancing the knowledge base and supporting policymakers for decision making and the scientific community.

Our study revealed significant hourly fluctuations in BC exposure concentrations both indoors and outdoors, found to be driven by daily cooking activities, with hourly measurements ranging from \(0.5\, \hbox {to}\, 38.0\,\upmu \textrm{g}\,\hbox {m}^{-3}\) and means between \(15 \,\hbox {and}\, 17\,\upmu \textrm{g}\,\hbox {m}^{-3}\). Beyond BC monitoring, community-level surveys and information on building structure significantly improved our understanding of temporal and spatial variations in pollutant concentrations, such as strengthening the knowledge about the typical cooking hours in rural communities in a day and the presence of other emission sources, such as space heating using SBF or lighting using kerosene, in these regions.

Although mean concentrations during the campaign remained consistent in all locations, examining diurnal patterns revealed clear differences driven by kitchen design and atmospheric conditions. For instance, BC concentrations declined more rapidly after morning cooking sessions, especially in households with an outdoor enclosed kitchen. Post morning cooking activity, the consistent decline was attributed to favorable daytime ventilation conditions, such as higher planetary boundary layer heights and increased wind speeds41,73, which enhance pollutant dispersion and reduce exposure concentrations. This highlights how kitchen layout and the timing of activities influence exposure patterns, thereby guiding the development of behavior-focused interventions such as optimal cooking times or kitchen modifications.

In rural India, households use various kitchen configurations and stoves to satisfy their daily culinary requirements, influenced by elements such as quantity and taste of food, availability of cooking space and thermal comfort78,81,82. There is also a marked preference for outdoor heating systems in winter, especially in northern India43,93. The survey identified a pronounced preference in rural communities for the charred taste of bread and smoky dishes. Furthermore, SBFs (including newspapers and waste plastics) are utilized by these communities as effective heating options in colder times due to their low cost relative to other fuel types and their ready availability. The effects are particularly notable during December’s late-night hours, marked by decreasing temperatures and heightened outdoor heating use under stable atmospheric conditions. This seasonal and behavioral shift demands that cultural context be considered in advancing clean energy strategies due to the ongoing reliance on SBFs for daily energy needs, which exacerbates BC exposure.

Despite the availability of cleaner alternatives such as LPG, its continued adoption in rural households is slow and often reverts back to the use of SBF13,81,108,109, as seen in this study. The cost of LPG remains high, even with financial aid, compared to more accessible SBF sources. Observations from this study indicate faster pollutant dispersion after morning cooking sessions compared to evening ones, attributed to more stable atmospheric conditions at night that extend exposure durations. Consequently, a strategy that includes the intermittent use of cleaner fuels is recommended. In scenarios where cleaner fuels are limited, it is recommended that households transition from SBF to LPG for evening cooking, as LPG emits significantly fewer pollutants31,35, potentially reducing individual and community-level exposure. This transitional approach is especially advantageous during the winter season, when there is a higher demand for outdoor heating. The suggested modifications to cooking practices, particularly during the evening, are likely to improve health outcomes by reducing exposure to pollutants and encouraging regular use of LPG, thereby strengthening the dependence of rural communities on LPG throughout the transition period.

The scope of the measurements in this study was limited to two villages and two houses, which may not fully reflect the daily profiles of pollutant concentrations. Variations in BC concentrations can be affected by several factors, including the number of households in a community, kitchen sizes, and different cooking practices between families, all of which can lead to significant variations in observed data. Furthermore, space heating activities were not monitored in our study due to the absence of a fixed schedule or location and their dependence on outdoor temperature conditions. Although this study was limited to a small sample size and conducted during a single season (winter), the use of a mixed-methods approach to assess the temporal trends of BC concentrations in various micro-environments has offered enhanced insights for understanding and developing policies that take into account social, behavioral, and technological factors. To accurately estimate actual reductions in exposure from the proposed intermediate transition, more comprehensive studies are needed that involve a wide range of communities and numerous households over longer periods, with active implementation of the strategy. Therefore, further research is essential to devise strategies for lowering BC emissions and exposure, and to encourage intermediate transition pathways in rural areas.

Conclusion

This research provides a detailed analysis of BC concentration patterns in rural India employing concurrent continuous indoor and outdoor monitoring during an extended winter period. The results show that, despite similar average BC levels between sites, there were significant diurnal fluctuations strongly influenced by emission activities such as cooking and heating, along with weather conditions. Notable differences in BC decay between households with enclosed indoor and semi-enclosed outdoor kitchens show the impact of kitchen design and ventilation on indoor BC buildup and dissipation. Morning BC peaks fell more quickly than evening peaks, mainly due to increased daytime atmospheric mixing from a higher boundary layer and stronger winds. These findings highlight the need to consider behavioral practices, kitchen types, and meteorological factors when evaluating rural air pollution. The study points to the potential of combining time-resolved monitoring with community data to enhance the understanding of pollutant concentration dynamics. Further research across varied rural contexts, different seasons, and broader household samples are needed to generalize these results and assist targeted air quality improvements.