*Article* **Air Quality Modeling of Cooking Stove Emissions and Exposure Assessment in Rural Areas**

**Yucheng He <sup>1</sup> , Sanika Ravindra Nishandar <sup>1</sup> , Rufus David Edwards 2,\* and Marko Princevac <sup>1</sup>**


**Abstract:** Cooking stoves produce significant emissions of PM2.5 in homes, causing major health impacts in rural communities. The installation of chimneys in cooking stoves has been documented to substantially reduce indoor emissions compared to those of traditional open fires. Majority of the emissions pass through chimneys to the outdoors, while some fraction of the emissions leak directly into the indoor air, which is defined as fugitive emission. Indoor PM2.5 concentrations are then the result of such fugitive emissions and the infiltration of outdoor neighborhood pollutants. This study uses a combination of the one-contaminant box model and dispersion models to estimate the indoor PM2.5 household concentration. The results show that the contributions of outdoor infiltration to indoor PM2.5 concentrations increase with higher packing densities and ventilation rates. For a case study, under WHO recommended ventilation conditions, the 24 h average mass concentration is ~21 µg/m<sup>3</sup> , with fugitive concentration accounting for ~90% of the total exposure for highly packed communities. These results help to identify the potential benefits of intervention strategies in regions that use chimney stoves.

**Keywords:** dispersion model; health risk assessment; particulate matter; indoor air quality; cook stove; biomass burning

#### **1. Introduction**

Globally, many rural communities rely on traditional biomass burning stoves to meet household energy demands, both indoors [1] and outdoors [2], which result in a significant burden of disease [2–4]. WHO announced that an estimated 3.2 million deaths per year were attributed to household air pollution in 2020 [5]. In addition, many studies have reported high health risk associated with traditional open-fire cooking [6,7]. A large portion of the rural population utilizes traditional stoves for household needs because of socioeconomic factors including availability of fuel, the cost of the stove, and a lack of alternative energy sources such as LPG [3]. The necessity of reducing pollutants associated with stove burning has led to the development of technologies to improve combustion efficiency. The change of behavior in the kitchen can also reduce PM2.5 exposure. Although there are many different stove types, the combination of a burning chamber and a flue/chimney are commonly used in many areas of the world, in which majority of the emissions are exhausted from the chimney to the outdoor environment. Indoor PM2.5 concentrations are the result of the fraction of the emissions that leak directly into the indoor air [8], combined with the outdoor infiltrated pollutants. The outdoor neighborhood pollutants quantified in this study are from chimney emissions of an individual home and upstream neighborhood homes [9]. Often, the stacks for household chimneys installed in these regions are short, resulting in neighborhood pollution, and substantial emissions accumulate in the vicinity of homes, which enables infiltration back indoors [10,11]. The current study focuses on the

**Citation:** He, Y.; Nishandar, S.R.; Edwards, R.D.; Princevac, M. Air Quality Modeling of Cooking Stove Emissions and Exposure Assessment in Rural Areas. *Sustainability* **2023**, *15*, 5676. https://doi.org/10.3390/ su15075676

Academic Editors: José Carlos Magalhães Pires, Álvaro Gómez-Losada and Vincenzo Torretta

Received: 13 January 2023 Revised: 11 March 2023 Accepted: 15 March 2023 Published: 24 March 2023

**Copyright:** © 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

contributions of neighborhood pollution to indoor air pollution associated with different housing packing densities, which have not previously been well characterized. tion associated with different housing packing densities, which have not previously been well characterized.

accumulate in the vicinity of homes, which enables infiltration back indoors [10,11]. The current study focuses on the contributions of neighborhood pollution to indoor air pollu-

*Sustainability* **2023**, *15*, x FOR PEER REVIEW 2 of 14

Epidemiological studies have indicated that PM2.5 exposure can cause adverse health impacts through heart, respiratory, and other chronic diseases [12]. Biomass fuel burning is recognized as a major cause of chronic obstructive pulmonary disease, especially for individuals in developing countries [13]. Estimating the risk of exposure that can lead to health problems is vital to inform risk abatement strategies. The current analysis evaluates the outdoor neighborhood pollution distributions, outdoor to indoor infiltration, indoor stove fugitive contamination, and the associated PM2.5 exposure risk [14]. Subsequently, EPA health risk assessment [15] was applied to quantify the inhalation risk of PM2.5 to the female population of different age groups to contribute to the development of control strategies for air quality management in and around rural communities where solid fuels are used for cooking. Epidemiological studies have indicated that PM2.5 exposure can cause adverse health impacts through heart, respiratory, and other chronic diseases [12]. Biomass fuel burning is recognized as a major cause of chronic obstructive pulmonary disease, especially for individuals in developing countries [13]. Estimating the risk of exposure that can lead to health problems is vital to inform risk abatement strategies. The current analysis evaluates the outdoor neighborhood pollution distributions, outdoor to indoor infiltration, indoor stove fugitive contamination, and the associated PM2.5 exposure risk [14]. Subsequently, EPA health risk assessment [15] was applied to quantify the inhalation risk of PM2.5 to the female population of different age groups to contribute to the development of control strategies for air quality management in and around rural communities where solid fuels are used for cooking.

#### **2. Materials and Methods 2. Materials and Methods**

#### *2.1. Background 2.1. Background*

The disability-adjusted life years (DALYs) represent the sum of years of populations living in a status of less than good health resulting from specific causes. Figure 1 presents the household attributed DALYs per 100,000 people according to WHO released data for the year 2019 [16]. The DALYs are substantially higher in developing countries, where biomass combustion supplies the majority of household primary energy [17]. The disability-adjusted life years (DALYs) represent the sum of years of populations living in a status of less than good health resulting from specific causes. Figure 1 presents the household attributed DALYs per 100,000 people according to WHO released data for the year 2019 [16]. The DALYs are substantially higher in developing countries, where biomass combustion supplies the majority of household primary energy [17].

**Figure 1.** Household air pollution attributed DALYs per 100,000 people [16]*.* **Figure 1.** Household air pollution attributed DALYs per 100,000 people [16].

#### *2.2. Evaluation Framework*

*2.2. Evaluation Framework* The schematic of the study process is shown in Figure 2. The study is divided into the estimation of outdoor and indoor air quality. For the outdoor air quality study, meteorological parameters such as temperature, wind speed, direction, and cloud cover are required. These required micrometeorological parameters for the dispersion model inputs are derived from routine weather data [18,19]. Integrating the dispersion model with a meteorological preprocessing approximation is a good alternative when the field measurements are not available. The approximation details, together with field validation, are described in the Supplementary Material. Figure S1 shows a good agreement of the measured friction velocity with the approximation model output. A dispersion model is then deployed to quantify outdoor pollutant distribution. Such outdoor pollutants near the The schematic of the study process is shown in Figure 2. The study is divided into the estimation of outdoor and indoor air quality. For the outdoor air quality study, meteorological parameters such as temperature, wind speed, direction, and cloud cover are required. These required micrometeorological parameters for the dispersion model inputs are derived from routine weather data [18,19]. Integrating the dispersion model with a meteorological preprocessing approximation is a good alternative when the field measurements are not available. The approximation details, together with field validation, are described in the Supplementary Material. Figure S1 shows a good agreement of the measured friction velocity with the approximation model output. A dispersion model is then deployed to quantify outdoor pollutant distribution. Such outdoor pollutants near the household are the infiltration source to the indoor environment. The indoor generated PM2.5 is from indoor fugitive emissions. The infiltrated and the indoor-generated PM2.5, combine to form

*Sustainability* **2023**, *15*, x FOR PEER REVIEW 3 of 14

indoor pollution. Finally, a US EPA health risk assessment methodology [20] is utilized to quantify the potential dose and risk quotient for long-term exposure to such pollutants. [20] is utilized to quantify the potential dose and risk quotient for long-term exposure to such pollutants.

household are the infiltration source to the indoor environment. The indoor generated PM2.5 is from indoor fugitive emissions. The infiltrated and the indoor-generated PM2.5, combine to form indoor pollution. Finally, a US EPA health risk assessment methodology

#### **Figure 2.** Analysis flow*.* **Figure 2.** Analysis flow.

#### *2.3. Dispersion Model 2.3. Dispersion Model*

Dispersion models are effective, widely utilized tools to evaluate atmospheric pollution level when field measurements are not available. Such models have been utilized to quantify pollution emitted from different sources, such as cooking, traffic, and industry [17,21,22]. The ability to isolate emission sources, thus targeting the sole effect of one possible source, can inform source management and relevant policymaking. Most of the plume models do not consider complex effects of obstacles such as buildings on the dispersion of pollutants in urban or suburban areas [23]. The current study compares AER-MOD with the Quick Urban and Industrial Complex (QUIC) results to examine the influence of building morphology. QUIC accommodates building influences, rapidly enables detailed modeling of the flow field around buildings, and applies this generated wind field in a particle dispersion model. The simulation results are utilized to quantify the neighborhood infiltration because it estimates pollution dispersion in the vicinity of the buildings. AERMOD is extensively used for regulatory purposes and plays a substantial role in decision making [24]. AERMOD does not explicitly solve the flow features in the vicinity of obstacles, but accounts for obstacles through a Plume Rise Model Enhancements (PRIME) model. The current analysis compares neighborhood pollution results using both approaches. The comparison of point source dispersion among Gaussian, QUIC, and water channel evaluation is presented in Figure S2. And the comparison of contours for the outdoor emission estimation of AERMOD output with QUIC output is given in Dispersion models are effective, widely utilized tools to evaluate atmospheric pollution level when field measurements are not available. Such models have been utilized to quantify pollution emitted from different sources, such as cooking, traffic, and industry [17,21,22]. The ability to isolate emission sources, thus targeting the sole effect of one possible source, can inform source management and relevant policymaking. Most of the plume models do not consider complex effects of obstacles such as buildings on the dispersion of pollutants in urban or suburban areas [23]. The current study compares AERMOD with the Quick Urban and Industrial Complex (QUIC) results to examine the influence of building morphology. QUIC accommodates building influences, rapidly enables detailed modeling of the flow field around buildings, and applies this generated wind field in a particle dispersion model. The simulation results are utilized to quantify the neighborhood infiltration because it estimates pollution dispersion in the vicinity of the buildings. AERMOD is extensively used for regulatory purposes and plays a substantial role in decision making [24]. AERMOD does not explicitly solve the flow features in the vicinity of obstacles, but accounts for obstacles through a Plume Rise Model Enhancements (PRIME) model. The current analysis compares neighborhood pollution results using both approaches. The comparison of point source dispersion among Gaussian, QUIC, and water channel evaluation is presented in Figure S2. And the comparison of contours for the outdoor emission estimation of AERMOD output with QUIC output is given in Figure S3.

Figure S3. QUIC has broad applications, primarily in modeling wind flow and dispersion patterns in urban or suburban areas, providing building-scale results that can show pollutant concentrations and their interaction with eddies in the built environment [23,25]. QUIC is a fast-response dispersion model that is comprised of a wind field model QUIC-URB and a dispersion model QUIC-PLUME. The flow patterns modeled by QUIC-URB have been validated with USEPA wind tunnel measurements [26]. QUIC-PLUME has also been validated through many experiments and modeling comparisons. For example, Zajic et al. QUIC has broad applications, primarily in modeling wind flow and dispersion patterns in urban or suburban areas, providing building-scale results that can show pollutant concentrations and their interaction with eddies in the built environment [23,25]. QUIC is a fast-response dispersion model that is comprised of a wind field model QUIC-URB and a dispersion model QUIC-PLUME. The flow patterns modeled by QUIC-URB have been validated with USEPA wind tunnel measurements [26]. QUIC-PLUME has also been validated through many experiments and modeling comparisons. For example, Zajic et al. compared QUIC-PLUME output to a Gaussian plume model output at different atmospheric stability, with the results being in good agreement for unstable and neutral atmospheric stabilities [27].

AERMOD was developed by the U.S. Environmental Protection Agency (EPA) in conjunction with the American Meteorological Society (AMS) to incorporate scientific advances into a dispersion model for regulatory applications [28]. AERMOD is a regulatory model with superior performance to other models in a 17 field study databases [29]. Many studies have integrated meteorological preprocessing to estimate AERMOD required inputs. Kumar et al. integrated a weather forecast model, using data-driven predicted meteorological data as inputs [30]. The current meteorological approximation has been validated for different built environments including relatively uniform terrain, as well as urban canopies [18,31].
