Seasonal variation in indoor concentrations of air pollutants in residential buildings

ABSTRACT

Indoor concentrations of PM₁₀, PM_2.5, CO, and CO₂ were measured in 25 naturally ventilated urban residences during the winter and summer seasons in Alexandria, Egypt. Ambient air samples were also collected simultaneously for comparison to indoor measurements. Furthermore, data for air exchange rates, home characteristics, and indoor activities during sampling were collected. It was found that the average indoor PM₁₀, PM_2.5, CO, and CO₂ concentrations for all homes in winter were 119.4 ± 30.9 μg/m³, 85.2 ± 25.8 μg/m³, 1.6 ± 0.8 ppm, and 692.4 ± 144.6 ppm, respectively. During summer, the average indoor levels were 98.8 ± 21.8 μg/m³, 67.8 ± 14.9 μg/m³, 0.5 ± 0.5 ppm, and 558.2 ± 66.2 ppm, respectively. The results indicate that the indoor daily averages of PM₁₀ and PM_2.5 concentrations were higher than the World Health Organization (WHO) guidelines for all selected homes in the two sampling periods. For CO and CO₂ levels, the indoor daily averages for all monitored homes were less than the WHO guideline and the American National Standards Institute/American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc. (ANSI/ASHRAE) Standard 62.1, respectively. A strong seasonal variability was observed, with air quality being particularly poor in winter. Due to increased ventilation rates in summer, indoor levels of air pollutants were strongly dependent on ambient levels, while in winter the indoor concentrations were more strongly affected by indoor sources due to increased human activities and poor ventilation. In addition, stronger indoor/outdoor correlation of air pollutants’ levels was found in summer than in winter probably due to higher ventilation and infiltration in the summer. The study also attempted to understand the potential sources and the various determinants that influence indoor PM, CO, and CO₂ concentrations in the two seasons. The findings can assist policymakers to better understand the indoor air pollution problem and to provide a sound basis for the development of proper national IAQ standards in Egypt.

Implications: Personal exposure is considerably influenced by indoor air pollution which increases health risks. Assessment of indoor air quality has become a more significant issue in Egypt as people tend to spend most of their time inside buildings, especially in their homes. Currently, there is a lack of research on residential indoor air quality in Egyptian cities in terms of the spatial and temporal variation which prevents an accurate assessment of the current situation to develop effective mitigation measures and to establish national indoor air quality standards. This article is considered the first research studying the effect of seasonality on indoor concentrations of PM₁₀, PM_2.5, CO, and CO₂ in urban residences in Alexandria. It also studies the indoor/outdoor relationship of air pollutants’ levels and identifies their major sources as well as the various determinants that influence their indoor concentrations.

Introduction

Ambient air pollution and its potential health effects have attracted great attention worldwide for long periods of time. Most people spend much of their time (about 90%) within enclosed buildings, while vulnerable populations such as children, the elderly, and the infirm often spend entire days indoors (Odeh and Hussein 2016; Schweizer et al. 2007). Thus, indoor air pollution has increasingly become a public health concern as it may pose more harmful effects and constitute the major risk factor to personal exposure (Faria et al. 2020; WHO 2010). Indoor air quality (IAQ) is largely influenced by many factors such as indoor emission sources, building characteristics including design and ventilation, and outdoor air quality. Additionally, physical parameters (relative humidity and temperature) and occupants’ behavior and habits cannot be ignored (Leung 2015). Many studies have found associations between poor IAQ and several negative health effects, including sick building syndrome (SBS), asthma exacerbation, and increased blood pressure (Rumchev et al. 2018). Most indoor air pollutants cause respiratory and cardiovascular problems, and the severity of effect varies according to their toxicity, concentration, and duration of exposure, and the health status of the individuals exposed (Maynard 2019; Vardoulakis et al. 2020).

Particulate matter (PM) is one of the main air pollutants related to adverse effects on human health both indoors and outdoors (Franck et al. 2011; Pope, Ezzati, and Dockery 2009). Indoor airborne PM consists of particles that were penetrated from outdoor sources and those generated indoors. Particles generated indoors can be released either directly from indoor sources and are known as primary PM or generated by chemical reactions of gaseous precursors (i.e., gas-to-particle conversion) emitted indoors and/or outdoors, and these particles are called secondary PM. PM with an aerodynamic diameter less than 10 μm is referred to as PM₁₀ which is divided into a coarse fraction with an aerodynamic diameter 2.5–10 μm (PM_2.5–10), and a fine fraction of less than 2.5 μm in aerodynamic diameter (PM_2.5). Fine PM fraction often contains hazardous substances and can penetrate deeper into the lungs (Ghio, Carraway, and Madden 2012; Yang et al. 2018). Therefore, fine particles have been linked to increased respiratory and cardiovascular morbidity and mortality (Cincinelli and Martellini 2017; Logue et al. 2012; WHO 2006). Sources of ambient PM in urban environments include vehicle emissions, industrial processes, construction activities, road dust from brake abrasion and tire friction, re-suspension of deposited particles, and long-range transport (Charron and Harrison 2005). In residential buildings, indoor PM is released by various human activities such as cooking, smoking, vacuuming, candles and incense burning, cleaning, and walking (Barraza et al. 2014; He et al. 2004). Indoor PM levels are affected by many factors including ambient PM concentrations, generation rate of PM from various indoor sources, ventilation rates, particle penetration rate from outdoor environments, and particle deposition and re-suspension mechanisms in indoor environments (Fromme et al. 2007; Rivas et al. 2019).

Carbon monoxide (CO) is a colorless and odorless gaseous air pollutant that is emitted by incomplete combustion of fossil fuels. Transportation and industry are the main ambient sources of CO in urban areas. Major indoor sources of CO include tobacco smoke, gas stoves, and any other burners of carbon-containing fuels (WHO 2010). Indoor exposure to high levels of CO exceeding 30 ppm is rare and largely related to urban homes with poorly adjusted stoves or close to significant CO emitting sources (e.g., attached garages or main roads) (ATSDR, 2009). For indoor environments close to major traffic roads, indoor CO concentrations will be largely influenced by outdoor CO levels particularly for those with little or no significant indoor sources (Zhong, Yang, and Kang 2013). Exposure to CO causes various health effects as it impairs the ability of blood to deliver oxygen to vital organs through reaction with the blood’s oxygen carrier known as hemoglobin (Hb) to form carboxyhemoglobin (COHb), thus affecting the cardiovascular, pulmonary, and nervous systems (Maynard, Myers, and Ross 2016). This depends largely on CO concentration, exposure period, and health status of exposed individuals (Kurti et al. 2016; Reboul et al. 2012). Carbon dioxide (CO₂) is mainly produced by both combustion processes and human metabolism. Therefore, CO₂ concentrations in occupied indoor environments are higher than those found outdoors (Fisk 2017). The indoor-outdoor difference in CO₂ concentration increases as the ventilation rate per occupant decreases (Satish et al. 2012). The American National Standards Institute/American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc. ANSI/ASHRAE Standard 62.1 (2013) recommends that building ventilation rates are sufficient to keep indoor CO₂ levels less than 700 ppm above outdoor CO₂ concentrations which typically range from 300–500 ppm. The standard considers CO₂ to be a surrogate for other occupant-generated pollutants, such as human bioeffluents (body odors), and not as an agent of adverse health effects (Persily 2015). As a natural product of human respiration, indoor occupant density is the most important source of indoor CO₂ elevations (Abdel-Salam 2015). Average indoor CO₂ levels in normal indoor settings typically range from 500 to 1,000 ppm, but can exceed 2,000 ppm with increased room occupant densities and reduced ventilation rates (Persily and de Jonge 2017). Recent trends in building designs, including increasing air tightness accompanied with low building ventilation rates to save energy and slow climate change, can reduce IAQ and increase indoor CO₂ concentrations (Papachristos 2015; Shrubsole et al. 2019). Indoor CO₂ concentrations indicate if the rate of outdoor air supply to the indoors is adequate to dilute indoor air pollutants (Gall et al. 2016).

IAQ data are generally scarce in Egyptian cities and the problem is less understood. There is currently a lack of research studies in Egypt to investigate IAQ in urban homes. Better understanding of major influential factors and variables controlling IAQ in domestic households is necessary to develop effective control measures and identify priority issues for researchers and policymakers. Therefore, the current work aims to determine the quality of indoor air in a number of residential homes of an urban area during different seasons in Alexandria, Egypt. This is considered the first research studying the effect of seasonal variation on indoor levels of air pollutants in residential buildings in Alexandria. This research also aims to study the relationship between indoor and outdoor concentrations of air pollutants and identification of their main sources in the selected residences. Furthermore, the study investigates potential determinant factors controlling indoor air quality for urban homes.

Experimental methods

Study description and sampling sites

The current study was performed in the city of Alexandria, Egypt. Alexandria is the second largest Mediterranean port city located in the north central part of Egypt (31º12ʹ N 29º55ʹE) about 200 km to the north west of Cairo, with a total area of about 2,679 km² and a total population of about 5.3 million (CAPMAS 2019). It is characterized by a Mediterranean temperate climate with warm humid summer which lacks wet precipitation and mild wet winter. The annual average temperature is 21°C. August is the hottest month with an average temperature of 28°C, whereas January is the coldest month with an average temperature of 14°C. Indoor air samples of PM₁₀, PM_2.5, CO, and CO₂ were collected in 25 homes located in a densely populated residential and commercial urban area known as Sidi Bishr (31º15ʹ28”N 29º58ʹ57”E), which is about 10 km to the east of the city center and is located in the Montaza district. The study area contains a mixture of residential buildings and commercial shops. Air sampling was conducted during two measurement periods in 2019: the winter season (January to March) and the summer season (July to September). The sampling sites are all private flats situated in the first floor (about 5 m high from the ground) of multistory residential buildings and with varying distances from heavy traffic roads. No attached garages were found in those residential buildings. Indoor air samples were collected in the living room of each home and at a height of approximately 1.5 m above the floor, to simulate sampling in the breathing zone of the occupants, and away from the doors in order to avoid disturbances from air currents. During indoor residence times, occupants are assumed to spend waking hours in the living room which is their main activity place and it usually has high occupant density (Tang and Wang 2018; Wang et al. 2008). It is considered a major indoor microenvironment inside dwellings where occupants are exposed to various air pollutants from both outdoor and indoor sources (e.g., smoking, cleaning, cooking) (Mölter et al. 2012). A common architectural design of residential buildings in the city is to provide living rooms with balconies which act as a significant source of natural ventilation. Outdoor air samples were collected at the same height in the balconies, just outside the living rooms. Indoor and outdoor air measurements of PM₁₀, PM_2.5, CO, and CO₂ were conducted simultaneously for 24 h at each home on 3 occasions in each season followed by calculation of seasonal average values. All selected homes were inhabited during the two measurement periods and no pets were found. Prior to sampling, paired air samplers were tested side-by-side in the lab for the verification of the calibration factors, which were then used to correct the readings, if needed, to give true results.

Information on the general conditions and different indoor activities in the selected residences was collected using a questionnaire filled by the occupants. This included the condition and age of the buildings, type of ventilation, size of living rooms, number of occupants, number of smokers, frequency of cleaning and cooking, and any other potentially significant factors. All homes were supplied with natural gas for cooking and other activities of the occupants. Cooking activities in the sampling sites were conducted by gas stoves 2–3 times per day. Floors of all sampled living rooms, either wood or tile, were covered with carpets. All homes relied on natural ventilation through opening windows and doors but with different ventilation times during the two sampling periods. Electric heaters were the only type of space heaters used infrequently by occupants during the winter season that have insignificant effect on IAQ (Ni et al. 2016; Ruiz et al. 2010). The main characteristics of the sampling sites are given in Table 1.

Table 1. Main characteristics of all selected sites

Site No.	No. of occupants	No. of smokers	Home age (yr)	Volume of living room (m^3)	Frequency of home cleaning	Exhaust fan/hood in the kitchen	Distance from major roads (m)
Home 1	6	0	7	32	Infrequent	No	40
Home 2	2	0	6	56	Infrequent	Yes	180
Home 3	3	0	9	45	1 x day	Yes	240
Home 4	3	0	11	52	1 x day	Yes	130
Home 5	4	0	8	40	Infrequent	Yes	70
Home 6	3	1	13	36	1 x day	No	140
Home 7	5	2	10	27	Infrequent	No	90
Home 8	4	0	5	40	1 x day	No	260
Home 9	3	0	14	55	1 x day	Yes	200
Home 10	5	0	12	42	Infrequent	No	50
Home 11	4	0	9	52	Infrequent	Yes	160
Home 12	3	1	10	47	1 x day	No	200
Home 13	2	0	3	63	1 x day	Yes	220
Home 14	5	0	4	72	Infrequent	Yes	170
Home 15	4	0	7	35	1 x day	Yes	150
Home 16	1	0	9	80	1 x day	Yes	180
Home 17	6	1	11	55	Infrequent	No	130
Home 18	3	0	2	67	Infrequent	Yes	250
Home 19	2	0	6	70	1 x day	Yes	60
Home 20	5	1	19	45	Infrequent	No	110
Home 21	4	1	15	48	1 x day	Yes	80
Home 22	5	1	10	37	Infrequent	No	70
Home 23	3	0	5	60	1 x day	Yes	250
Home 24	3	0	15	65	1 x day	Yes	20
Home 25	4	2	18	41	Infrequent	No	100

Sampling and analytical methods

Carbon monoxide and carbon dioxide

Indoor and outdoor CO and CO₂ concentrations were monitored simultaneously by using two portable Q-Trak monitors (model 8551; TSI Inc., Shoreview, MN, USA). These monitors detect CO₂ concentrations based on the mechanism of non-dispersive infrared (NDIR). The monitor is also equipped with an electro-chemical sensor for CO measurement. The selected sampling interval of the two monitors was programmed for 1 min. Zero and span checks of the Q-Trak monitors were performed prior to each field campaign. In addition, a multi-point calibration of the two monitors was carried out by the manufacturer in order to obtain accurate measurements. The Q-Trak monitor measured CO₂ with a resolution of 1 ppm within the range of 0–5,000 ppm and with an accuracy of ± 3% of reading at 25°C (uncertainty of ± 0.36%/°C change in temperature). CO was measured with a resolution of 1 ppm within the range of 0–500 ppm, with an accuracy of ± 3% of reading (uncertainty of ± 0.5%/°C away from calibration temperature), and with a repeatability of ± 2% of reading.

Particulate matter (PM₁₀ and PM_2.5)

Indoor and outdoor air sampling of two PM fractions (PM₁₀ and PM_2.5) were also conducted simultaneously using the Marple PM₁₀ and PM_2.5 environmental monitor samplers (model 200 Personal Environmental Monitor (PEM), MSP Co., Minneapolis, MN), respectively (Marple et al. 1987). The two sampled aerosol fractions were collected on 37-mm diameter Teflon membrane filters with 2 µm pore size, at a flow rate of 10 L/min for 24 h in each household. The flow rate was regularly monitored at the beginning and end of each sampling period using a calibrated airflow meter (Model 320–100, SKC Inc., PA, USA). Before the weighing process, all filters were conditioned for 24 h in a controlled temperature and RH room of about 20°C and 45%, respectively. Before and after sampling, gravimetric analysis was carried out three times to both unexposed (blank) and air-exposed (sampled) filters using a high-precision microbalance (Mettler Toledo AT261; Mettler-Toledo, Inc., Columbus, OH) (a readability of 0.01 mg, an accuracy of ± 0.03 mg, and a repeatability of 0.015 mg), followed by calculation of average values. Unexposed control filters were used to eliminate weighing errors produced by differences in temperature and RH between weighings. All used filters were handled and processed with extreme care to avoid any inadvertent sample contamination or particle loss before weighing.

Air exchange rate

Ventilation is the process of supplying fresh air to any confined space and removing contaminated air from it. Therefore, this process is very important for good IAQ and a healthy indoor environment. Natural ventilation is the mechanism by which indoor air is exchanged with outdoor air as a result of pressure differences caused by wind and/or buoyancy forces (Almeida, Barreira, and Moreira 2017; Saha et al. 2013). Air exchange rate (AER) or ventilation rate was measured in the living rooms of all selected residences by the tracer gas decay method. In this method, the tracer gas should be detected easily at low concentrations and should not also be available in the background air. Therefore, sulfur hexafluoride (SF₆) was released as an odorless and nontoxic tracer gas in the center of the living rooms and at a height of about 1.5 m. The tracer gas was mixed and uniformly distributed in each living room to a concentration of about 10 ppm. Then, the gas was allowed to decay and the rate of decay of the SF₆ was measured by the photo-acoustic multi-gas monitor (Innova 1412i; LumaSense Technologies, Inc., Santa Clara, CA, USA), and the measured decay rate was used to calculate the AER. The method was applied at the beginning and end of the measurement periods followed by calculation of average values. This method is well established and fully described in the standards ISO 12,569 (ISO 2000) and ASTM E 741 (ASTM 2011).

Statistical analysis

All descriptive statistics and statistical analyses of data were performed by Statistical Package for Social Sciences (SPSS) for Windows (version 20.0, SPSS, Inc., NY, USA). Before applying the statistical tests, the distribution of investigated parameters was evaluated using the Shapiro–Wilk test. Statistical analyses were applied using two-tailed tests at a 0.05 level of significance. Correlations among measured air pollutants as well as between indoor levels of air pollutants and other influencing indoor and outdoor factors were analyzed by the Pearson bivariate correlation coefficient (r). Seasonal differences between indoor concentrations of each air pollutant were calculated and t-test was used to determine statistical significance. Similarly, differences in average indoor concentrations of air pollutants regarding provision of exhaust fans or hoods in the domestic kitchens as well as frequency of cleaning were also calculated and statistically tested.

Results and discussion

Indoor and outdoor concentrations of air pollutants

The average indoor PM₁₀ and PM_2.5 concentrations for all sampling sites over the two study periods were 109.1 ± 28.4 μg/m³ (range: 70–190 μg/m³; median: 100 μg/m³) and 76.5 ± 22.6 μg/m³ (range: 47–145 μg/m³; median: 75 μg/m³), respectively. The corresponding average outdoor levels were 138.5 ± 26.3 μg/m³ (range: 91–197 μg/m³; median: 135 μg/m³) and 79.7 ± 14.4 μg/m³ (range: 55–108 μg/m³; median: 79.5 μg/m³), respectively. The average PM_2.5 contribution to PM₁₀ concentrations was 58.1 ± 7.4% for outdoors and 70.0 ± 8.7% for indoors. Therefore, the fine PM fraction (PM_2.5) originated mainly from combustion sources were the dominating fraction in the outdoor and indoor PM₁₀. Also, indoor PM₁₀ and PM_2.5 concentrations in all selected households were found to be strongly correlated (r = 0.90, p < .01). Significant correlation was also found between outdoor levels of PM₁₀ and PM_2.5 (r = 0.80, p < .01). This is probably due to their emissions from similar sources.

For CO and CO₂, the average indoor concentrations were 1.1 ± 0.9 ppm (range: 0.1–3.1 ppm; median: 1.0 ppm) and 625.3 ± 130.3 ppm (range: 460–966 ppm; median: 596 ppm), respectively. The corresponding average outdoor levels were 1.3 ± 0.6 ppm (range: 0.2–2.7 ppm; median: 1.3 ppm) and 452.1 ± 31.8 ppm (range: 398–519 ppm; median: 444.5 ppm), respectively. Currently, there are no IAQ standards established in Egypt. The indoor daily averages of PM₁₀ and PM_2.5 concentrations for all selected homes in the two sampling periods were higher than the WHO guidelines for an averaging time of 24 h (50 μg/m³ for PM₁₀ and 25 μg/m³ for PM_2.5) (WHO 2006). For CO levels, the indoor daily averages for all homes were less than the 24 h WHO guideline of 6 ppm (WHO 2010). The indoor CO₂ concentrations were also compared to the ANSI/ASHRAE Standard 62.1 and no exceedances were found for all sampling sites.

Seasonal effect on indoor levels of air pollutants

The daily average concentrations of PM₁₀, PM_2.5, CO, and CO₂ that were measured inside and outside all the monitoring sites during the winter and summer seasons are shown in Table 2. Outdoor concentrations of PM and gaseous pollutants were found to be higher during winter because of lower ambient temperatures than summer. Lower outdoor temperatures will favor atmospheric stability and lower mixing layer height in the winter season, which prevent the dispersion of air pollutants (Chithra and Nagendra 2014). There was a marked seasonal variation in outdoor CO levels (p < .0001), whereas no significant seasonal difference was found between outdoor levels for PM and CO₂. It can also be seen from Table 2 that the indoor concentrations of the pollutants measured in this study varied widely between the two measurement periods. The highest indoor concentrations of all target air pollutants were observed during the winter season, whereas their lowest concentrations were observed during summer. Seasonal variation of the indoor concentrations of PM₁₀, PM_2.5, CO, and CO₂ in all sampling sites are shown in Figures 1–4, respectively. Error bars were drawn on the figures to show the variation in indoor levels of air pollutants for each observed site during the two seasons. The daily average concentrations of all measured air pollutants were higher in winter than in summer. On average, indoor concentrations were decreased by 20.6 μg/m³ for PM₁₀, 17.4 μg/m³ for PM_2.5, 1.1 ppm for CO, and 134.2 ppm for CO₂ in summer compared to winter. These differences were found to be highly statistically significant (p < .0001). These large seasonal variations in the indoor levels could be due to different ventilation practice in winter and summer as well as contribution from different indoor pollution sources (Keeler et al. 2002; Tong et al. 2018). Based on highly variable weather conditions, previous studies reported a significant increase in time spent indoors at home during the winter compared to the summer with consequent seasonal differences in indoor time-activity patterns (Leech et al. 2002; Matz et al. 2014; Schweizer et al. 2007). Due to increased ventilation rates in summer, indoor levels of air pollutants may strongly depend on outdoor concentrations, while in winter the indoor levels may be more strongly affected by indoor sources due to increased human activities and low ventilation times (Fromme et al. 2007; Massey et al. 2012). Therefore, air exchange rates were measured for all sampling sites during the two measurement periods.

Table 2. Summary statistics for indoor and outdoor concentrations of target air pollutants measured in 25 homes during winter and summer seasons

	Winter				Summer
Parameters	Mean ± S.D.	Min.	Max.	Median	Mean ± S.D.	Min.	Max.	Median
PM10 indoor (µg/m^3)	119.4 ± 30.9	81	190	110	98.8 ± 21.8	70	165	95
PM10 outdoor (µg/m^3)	140.4 ± 24.6	102	187	140	136.5 ± 28.2	91	197	129
PM2.5 indoor (µg/m^3)	85.2 ± 25.8	51	145	81	67.8 ± 14.9	47	98	63
PM2.5 outdoor (µg/m^3)	80.9 ± 15.0	55	108	81	78.5 ± 13.9	55	106	77
CO indoor (ppm)	1.6 ± 0.8	0.4	3.1	1.4	0.5 ± 0.5	0.1	1.8	0.3
CO outdoor (ppm)	1.5 ± 0.6	0.5	2.7	1.5	1.1 ± 0.6	0.2	2.3	1.0
CO2 indoor (ppm)	692.4 ± 144.6	501	966	618	558.2 ± 66.2	460	676	561
CO2 outdoor (ppm)	453.5 ± 33.5	405	519	448	450.7 ± 30.7	398	515	441

Figure 1. Seasonal variation of the indoor concentrations of PM₁₀ in all 25 selected homes

Figure 2. Seasonal variation of the indoor concentrations of PM_2.5 in all 25 selected homes

Figure 3. Seasonal variation of the indoor concentrations of CO in all 25 selected homes

Figure 4. Seasonal variation of the indoor concentrations of CO₂ in all 25 selected homes

AER measurements are important in determining the exchange of PM and gaseous pollutants between outdoor and indoor air (You et al. 2012). The average AERs for all studied households in the summer and winter seasons were 4.5 ± 0.9 h⁻¹ (range: 2.8–5.6 h⁻¹; median: 4.4 h⁻¹) and 1.1 ± 0.4 h⁻¹ (range: 0.4–1.9 h⁻¹; median: 1.1 h⁻¹), respectively. The average AER was significantly higher (p < .0001) during the summer than in the winter. The increased ambient temperatures during the summer season led to higher ventilation times and AERs in all selected residences than in the cold season, and this is in agreement with previous studies (Isaacs et al. 2013; Langer et al. 2016; Yamamoto et al. 2010). For comparison with previous studies from the scientific literature, Table 3 summarizes findings of previous IAQ studies conducted in living rooms of residences in other countries. The results obtained in the current study were compared with indoor concentrations measured in households from other countries (See Table 3). Generally, the indoor PM₁₀ and PM_2.5 concentrations measured in homes of this study were higher than those found in other countries except in Harbin, China (Tang and Wang 2018) and Agra, India (Massey et al. 2012). For CO and CO₂, the indoor levels measured in households of Alexandria were lower than those found in other cities except for CO measured in United Arab Emirates (Yeatts et al. 2012).

Table 3. Summary of findings in previous published IAQ studies

Reference	No. of households	Location	Mean indoor concentrations	Main pollution sources
Byun et al. (2010)	50	Korea	PM10 (μg/m^3): summer: 45.3	Cooking, cleaning, smoking
Cattaneo et al. (2011)	60	Lodi (River Po Valley) Italy	PM2.5 (μg/m^3): summer: 21.6; winter: 46.3 PM10 (μg/m^3): summer: 29.7; winter: 53.2 CO (ppm): summer: 1.4; winter: 1.7 CO2 (ppm): summer: 569.6; winter: 877.9	Road traffic, heating, tobacco smoking, cooking, cleaning
Héroux et al. (2010)	145	Regina, Canada	PM2.5 (μg/m^3): winter: 5.5; summer: 6.4	Cooking/cooking appliances, smoking, candles
MacNeill et al. (2014)	50	Halifax, Canada	PM2.5 (μg/m^3): winter: 6.8; summer: 10.1	Cooking, heating, cleaning
Massey et al. (2012)	10 (5 roadside and 5 urban)	Agra, India	Roadside houses, urban houses: PM2.5 (μg/m^3): summer: 119, 80; winter: 207, 146 PM10 (μg/m^3): summer: 218, 137; winter: 304, 254	Cooking, smoking, outdoor vehicular traffic, garbage burning
Pickett and Bell (2011)	10	Connecticut, New York, Vermont; USA	CO (ppm): summer: 0.8 CO2 (ppm): summer: 663.2	Gas stoves, attached garage, smoking, outdoor sources
Tong et al. (2018)	55	Hong Kong	PM2.5 (μg/m^3): summer: 25.3; winter: 34.2	Outdoor concentrations, incense burning, cooking
Tunno et al. (2015)	21	Pittsburgh, USA	PM2.5 (μg/m^3): summer: 25.8; winter: 18.9	Cooking, smoking, cleaning, outdoor sources
Tang and Wang (2018)	10	Harbin, China	PM2.5 (μg/m^3): winter: 313 PM10 (μg/m^3): winter: 347 CO (ppm): winter: 3.8 CO2 (ppm): winter: 1,029	Cooking, smoking, heating, outdoor sources
Yeatts et al. (2012)	628	United Arab Emirates (all 7 emirates)	Median concentrations: PM2.5 (μg/m^3): winter: 6.2 PM10 (μg/m^3): winter: 43.9 CO (ppm): winter: 0.8	Incense, gas stove, smoking

Relationships between indoor and outdoor concentrations

Indoor concentration levels of air pollutants are generally affected by corresponding outdoor levels and indoor sources. In the summer season, daily average outdoor PM₁₀, PM_2.5, and CO concentrations exceeded the corresponding indoor levels by 100%, 92%, and 88%, respectively. In comparison, outdoor levels of PM₁₀, PM_2.5, and CO during the winter season exceeded indoor levels by only 84%, 56%, and 56%, respectively. The presence of indoor sources may dramatically increase the indoor levels if adequate ventilation is not available (Yang et al. 2009). For CO₂, indoor concentrations exceeded outdoor concentrations in both seasons for all monitoring sites. The indoor CO₂ levels are not harmful, but are indicators of ventilation and occupancy in the sampling sites (Mendes et al. 2013). High indoor CO₂ concentrations were mainly attributed to indoor sources and also as a result of crowded conditions combined with insufficient ventilation.

During the summer season, a statistically significant correlation was found between indoor and outdoor concentrations of PM₁₀ (r = 0.76, p < .01), PM_2.5 (r = 0.85, p < .01), CO (r = 0.72, p < .01), and CO₂ (r = 0.45, p < .05). Significant, but relatively weaker, correlations were also found during the winter season between indoor and outdoor concentrations of PM₁₀ (r = 0.60, p < .05), PM_2.5 (r = 0.71, p < .01), CO (r = 0.58, p < .05), and CO₂ (r = 0.40, p < .05). The correlation coefficient values between indoor and outdoor concentrations can be used as an indicator of the degree to which levels of air pollutants measured indoors are attributed to outdoor air. Stronger indoor/outdoor correlations were found in the summer period rather than in winter due to higher AERs through frequent opening of doors and windows, which allows better exchange of air between outdoors and indoors. In addition, higher ambient temperatures during the summer season led to higher air infiltration through openings and fissures in the building structure caused by differences in air pressure between outdoor and indoor environments (Cattaneo et al. 2011; Urso et al. 2015). There were consistently strong correlations between indoor and outdoor PM and CO levels during summer. This suggests that during the summer campaign there was a stronger relationship between outdoor PM and CO sources, mainly from nearby traffic emissions, and their indoor concentrations compared to winter. This supports the conclusion that vehicle emissions, migrating indoors from the outdoor air, are the most important factor influencing indoor concentrations of PM and CO during the summer. For CO₂, it was not very obvious that contribution from traffic emissions to indoor CO₂ concentrations was as important as for PM and CO. However, several potential indoor CO₂ sources, including combustion processes and human metabolism, are present during both winter and summer seasons, and the weak indoor/outdoor correlations obtained for CO₂ indicate the significance of these sources.

The indoor/outdoor (I/O) mass concentration ratios were also calculated to evaluate the difference between indoor and outdoor concentrations. The I/O ratios for PM₁₀, PM_2.5, CO, and CO₂ were calculated during the summer and winter seasons for all sampling sites and are shown in Figure 5(a,b). Box plots were used to show the minimum, lower quartile (the 25th percentile), median, upper quartile (the 75th percentile), and maximum values. The I/O ratio varies according to many factors such as intensity of indoor sources, outdoor levels, ventilation rate, penetration factor from outdoor air, and deposition rate (Chen and Zhao 2011). Other factors such as location, building design, and seasonal effect are also important (Massey et al. 2012). The average I/O ratios for all sampling sites during the summer season were 0.73 ± 0.09 (range 0.60–0.90; median: 0.72) for PM₁₀, 0.86 ± 0.10 (range 0.63–1.10; median: 0.87) for PM_2.5, 0.54 ± 0.37 (range 0.16–1.17; median: 0.50) for CO, and 1.24 ± 0.13 (range 1.01–1.47; median: 1.19) for CO₂. During the winter season, the average I/O ratios for all sampling sites were 0.85 ± 0.13 (range 0.57–1.09; median: 0.84) for PM₁₀, 1.05 ± 0.23 (range 0.72–1.46; median: 0.98) for PM_2.5, 1.03 ± 0.29 (range 0.52–1.53; median: 0.90) for CO, and 1.53 ± 0.29 (range 1.15–2.09; median: 1.42) for CO₂. Higher I/O ratios found in the winter season could be attributed to the lower ventilation rates compared to the summer season as occupants tend to keep windows and doors closed during the winter to protect them from extreme cold outside and thus achieve the required level of thermal comfort (Elbayoumi, Ramli, and Yusof 2015). During the warm summer, doors and windows were kept mostly open to provide cooler and more ventilated rooms. Furthermore, higher I/O ratios suggest additional sources in the indoor environments. Homes with I/O ratios > 1 indicate significant contribution from indoor activities, such as cooking, smoking and cleaning, whereas I/O ratios < 1 imply that the outdoor concentrations are more than the indoor ones.

Figure 5. Box plots for I/O PM₁₀, PM_2.5, CO, and CO₂ ratios for all sampling sites during (a) summer and (b) winter

Fine PM fraction (PM_2.5) showed higher average I/O ratios than that of the PM₁₀ for all homes in the two measurement periods. This is due to the higher emissions of PM_2.5 from indoor combustion sources; in addition to the much higher infiltration of ambient PM_2.5 from nearby heavy traffic roads into the indoor environments as they have smaller size than PM₁₀ (Hoek et al. 2008). Also, the coarse PM₁₀ particles can more easily fall out by sedimentation or deposit on windows and doors during their penetration from the outdoor air as they have higher deposition velocities than PM_2.5 (Thatcher and Layton 1995). Moreover, indoor PM₁₀ levels may be reduced as a result of their higher deposition onto indoor furnished surfaces and floors (Lai 2002). For CO, the average I/O concentration ratio in the summer was about half its value in the winter for all homes under investigation, suggesting that the main source of CO in the summer season was external from nearby traffic emissions, whereas a more contribution from indoor sources combined with low ventilation occurred in the winter. The I/O ratios of CO₂ were above 1 in all observed homes during the two measurement campaigns, and the correlations between the indoor and outdoor concentrations, though statistically significant, were weak. This implies that indoor sources could contribute to the indoor CO₂ levels more than those from outdoors.

Relationships between indoor air levels and other influencing factors

Correlations among indoor levels of all measured air pollutants as well as between indoor air levels and other indoor and outdoor influential factors are shown in Table 4. Strong positive correlations were observed between indoor PM₁₀ and PM_2.5 in both seasons, suggesting their releases from similar emission sources. Furthermore, Table 4 illustrates positive associations between indoor PM and gaseous parameters (CO and CO₂). In winter, stronger correlations were observed between indoor concentrations of measured parameters than in summer. Low ventilation or insufficient outdoor air supply in winter may cause the more significant correlations between measured pollutants released from the more intense indoor activities (Kabir et al. 2012). Considering the significant, but relatively weaker, correlations between the indoor pollutants’ concentrations in summer, these levels might be caused by the inflow of outdoor air which probably contains considerable amounts from nearby heavy traffic emissions. Higher ambient temperature in summer will cause ambient outdoor air to be forced into the buildings through windows, doors, and slits due to the push resulted from thermal gradient (Chan 2002; Gupta and Cheong 2007).

Table 4. Correlations between the measured indoor air pollutants and influential factors during two seasons

	Winter				Summer
Air pollutants	PM10	PM2.5	CO	CO2	PM10	PM2.5	CO	CO2
PM10 (μg/m^3)	1				1
PM2.5 (μg/m^3)	0.91**	1			0.84**	1
CO (ppm)	0.82**	0.93**	1		0.60*	0.66**	1
CO2 (ppm)	0.84**	0.87**	0.85**	1	0.58*	0.54*	0.70**	1
Home age (yr)	0.44*	0.58*	0.50*	0.30	0.48*	0.64**	0.65**	0.36
Distance from roads (m)	-0.41*	-0.55*	-0.43*	-0.38	-0.54**	-0.67**	-0.63**	-0.40*
No. of occupants	0.76**	0.60**	0.52*	0.78**	0.68**	0.45*	0.41*	0.69**
No. of smokers	0.80**	0.86**	0.83**	0.81**	0.71**	0.73**	0.74**	0.71**
Room volume (m^3)	-0.57**	-0.52**	-0.62**	-0.68**	-0.52**	-0.42*	-0.57**	-0.65**
AER (h^−1)	-0.64**	-0.71**	-0.82**	-0.85**	0.56*	0.65**	0.71**	0.29

*Correlation is significant at the 0.05 level (2-tailed).

**Correlation is significant at the 0.01 level (2-tailed).

Some home characteristics significantly affected concentrations of air pollutants indoors. Volume of living rooms was found to be inversely associated with indoor air pollutants’ levels both in summer and winter (See Table 4). This could be related to more dilution to indoor concentrations in larger rooms (Kozielska et al. 2020; Nasir and Colbeck 2013). Home age was also investigated as an important building-related factor that could affect IAQ due to the infiltration through cracks and fissures in the building envelope (Lai et al. 2012). Age of selected homes was found to range from 2–19 yr (See Table 1). A statistically significant correlation was realized between home age and indoor levels of PM and CO during the two measurement periods with higher correlation for fine PM fraction in both seasons than PM₁₀. No significant correlation was found between age of residential buildings and indoor CO₂ concentrations in both seasons. This indicates that indoor CO₂ levels are influenced by other important factors such as ventilation and indoor sources (Abdel-Salam 2015). Ventilation rate is another important building-related characteristic that was found to considerably affect indoor PM, CO, and CO₂ levels during the two measurement periods. In summer, significant correlation was found between AER and indoor PM and CO concentrations, while no significant correlation was found with indoor CO₂ levels. In winter, statistically significant, but negative, correlations were found between AER and indoor concentrations of all measured pollutants. Another important finding of the current study was the significant negative correlation of indoor levels of PM₁₀, PM_2.5, and CO with distance from main traffic roads in both seasons with stronger correlations during summer period. This is probably due to both higher ventilation and infiltration of outdoor air during warm season. Road traffic is also related to coarse particles through tire friction, brake abrasion, and resuspension of deposited road dust which is enhanced by the increased dryness as well as higher thermal circulation and wind speed during the summer months (Artíñano et al. 2004; Titos et al. 2014). One notable exception was found for CO₂, where a relatively weak significant correlation was observed during the summer period, while no significant correlation was found in the winter. Many previous studies already showed that indoor air quality in naturally ventilated residences could be largely influenced by proximity to nearby heavy traffic roads of high vehicular emissions (Chang 2002; Garcia et al. 2013).

The trend of increased indoor levels of air pollutants measured in this study during the winter season when compared with the summer season implies that several other factors influence IAQ during the winter, such as indoor activities and duration of human occupancy. The density of human occupancy, with people tending to spend more time indoors in the winter months than in the summer months, together with poor ventilation, can significantly affect IAQ in the winter. Indoor activities that generate air pollutants include smoking, the use of gas stoves for cooking, and cleaning. The strong positive correlations observed between indoor levels of air pollutants and number of smokers with higher association in winter than in summer (Table 4) confirmed previous findings demonstrating that smoking is a major contributor to indoor-generated PM and other combustion-related gaseous pollutants (Héroux et al. 2010; Langer et al. 2016). In the present study, smoking was recorded in only eight homes and I/O ratios of target air pollutants were calculated in smoking and nonsmoking households during winter and summer seasons and presented in Table 5. The effect of smoking as a significant source of PM, particularly finer particles, CO, and CO₂ was generally linked to higher I/O ratios in smoking homes when compared to nonsmoking homes in the two seasons, as illustrated in Table 5. In smoking homes, I/O ratios for PM_2.5, CO, and CO₂ in winter were significantly greater than 1, indicating additional indoor sources. Also, in homes where occupants smoke, the average indoor concentrations of PM₁₀, PM_2.5, CO, and CO₂ were higher than those measured in homes with no smokers by 45%, 60%, 125% and 41%, respectively in winter and by 36%, 44%, 72%, and 20%, respectively in summer.

Table 5. Summary statistics for indoor/outdoor (I/O) concentration ratios of target air pollutants in smoking and nonsmoking homes during winter and summer seasons

		I/O ratio (Winter)				I/O ratio (Summer)
Homes	Parameters	Mean ± S.D.	Min.	Max.	Median	Mean ± S.D.	Min.	Max.	Median
Smoking	PM10	0.97 ± 0.09	0.84	1.09	0.99	0.81 ± 0.07	0.68	0.90	0.83
	PM2.5	1.32 ± 0.13	1.09	1.46	1.36	0.97 ± 0.06	0.90	1.10	0.95
	CO	1.37 ± 0.12	1.22	1.53	1.38	0.83 ± 0.29	0.46	1.17	0.86
	CO2	1.85 ± 0.16	1.61	2.09	1.85	1.36 ± 0.08	1.22	1.46	1.38
Nonsmoking	PM10	0.79 ± 0.11	0.57	0.96	0.83	0.69 ± 0.07	0.60	0.81	0.70
	PM2.5	0.92 ± 0.12	0.72	1.23	0.92	0.81 ± 0.06	0.63	0.93	0.82
	CO	0.87 ± 0.18	0.52	1.33	0.85	0.37 ± 0.19	0.16	0.75	0.33
	CO2	1.37 ± 0.20	1.15	1.8	1.32	1.19 ± 0.12	1.01	1.47	1.16

Another important indoor combustion activity is cooking as one of the most significant indoor sources influencing IAQ. The contribution of residential cooking to air pollutants’ levels measured in the living rooms depends largely on many factors such as closing or opening the kitchen’s door, home layout, type of cooking, and use of exhaust fans or smoke hoods in the kitchens (Buonanno, Morawska, and Stabile 2009; Ferro et al. 2009). According to the questionnaire filled by the occupants, kitchen doors were kept open during cooking activities in all homes under investigation. Also, no information was collected on the method of cooking in the current study. Kitchen exhaust ventilation systems were found in 60% of the selected households (see Table 1). The use of such air exhaust ventilation systems in domestic kitchens during conducting cooking could result in significant reductions in cooking emissions (Sun and Wallace 2020; Willers et al. 2006). Indeed, homes supplied with exhaust fans or range hoods in kitchens had significantly lower indoor concentrations of PM₁₀ (average = 87.3 vs. 115.9 μg/m³; p < .0004 for summer, and 102.1 vs. 145.5 μg/m³; p < .0001 for winter), PM_2.5 (average = 60.4 vs. 78.9 μg/m³; p < .001 for summer, and 70.3 vs. 107.6 μg/m³; p < .0001 for winter), CO (average = 0.2 vs. 1.0 ppm; p < .0001 for summer, and 1.1 vs. 2.5 ppm; p < .0001 for winter), and CO₂ (average = 516.8 vs. 620.2 ppm; p < .0001 for summer, and 599.1 vs. 832.4 ppm; p < .0001 for winter) when compared with homes not provided with ventilation fans or hoods installed in the kitchens. Apparently, the use of these exhaust ventilation provisions in kitchens during stove operation can have a significant impact on indoor air pollution levels, favoring the removal of combustion-generated products, particularly in winter under conditions of reduced ventilation. In homes 1, 8, and 10, where there were no fans or smoke hoods installed in the kitchens as well as with no smokers, the indoor levels of combustion-generated pollutants including PM_2.5, CO, and CO₂ were probably attributable to cooking activity particularly in winter. In such homes, average I/O ratios in winter were equal to 1.09 ± 0.12 (range: 1.02–1.23; median: 1.03) for PM_2.5, 1.16 ± 0.15 (range: 1.05–1.33; median: 1.11) for CO, and 1.74 ± 0.06 (range: 1.69–1.80; median: 1.73) for CO₂. In summer, average I/O ratios were equal to 0.82 ± 0.05 (range: 0.78–0.87; median: 0.81) for PM_2.5, 0.58 ± 0.25 (range: 0.29–0.75; median: 0.69) for CO, and 1.36 ± 0.10 (range: 1.28–1.47; median: 1.34) for CO₂. The effect of cooking using gas stoves was more obvious for these homes in winter with average I/O ratios above 1 and higher than those calculated in summer because indoor pollutants’ concentrations in winter were more strongly affected by indoor sources due to poor ventilation and increased human activities.

Cleaning is a routine indoor activity in which vacuuming and sweeping are associated with higher indoor levels of PM through resuspension of previously deposited dust particles from domestic floors, carpets, and furniture (Corsi, Siegel, and Chiang 2008; Knibbs et al. 2012). In the present study, frequency of cleaning was on a daily basis in 52% of the monitored residences; whereas it was irregular in the remaining homes (See Table 1). Infrequent cleaning, defined as longer periods between cleaning activities, can be responsible for increased concentrations of indoor PM (particularly coarse fraction) as it may allow much more dust particles to deposit and accumulate on floors and other indoor surfaces followed by continuous resuspension by human activities including walking and movement (Lewis et al. 2018; Long, Suh, and Koutrakis 2000). As expected, daily cleaned homes had significantly lower indoor concentrations of PM₁₀ than less frequently cleaned homes (average = 89.1 vs. 109.3 μg/m³; p < .017 in summer, and 103.2 vs. 137.1 μg/m³; p < .004 in winter). Lower, but not statistically significant, indoor concentrations of PM_2.5 were found in daily cleaned homes than those for infrequently cleaned homes (average = 64.1 vs. 71.8 μg/m³ in summer, and 76.5 vs. 94.6 μg/m³ in winter). This is probably due to the various significant indoor and outdoor sources influencing indoor PM_2.5 concentrations.

Human occupancy and crowding is another important factor that could elevate indoor levels of measured air pollutants through increased human activities. Significant positive correlations were found between number of occupants and indoor pollutants’ levels in both seasons with higher associations occurred during the winter (see Table 4). Occupants’ activities including walking and movement could lead to resuspension of previously settled PM from floors, furniture and other different horizontal indoor surfaces (Qian, Peccia, and Ferro 2014; Rosati, Thornburg, and Rodes 2008). Coarse particles were found to be more affected by the increased number of occupants than fine particles in both seasons. Previous studies in the literature also showed similar results that human occupancy is considered an important determinant of coarse airborne particles particularly in the winter as a result of reduced ventilation in households, with resuspension rates increasing with particle size (Thatcher and Layton 1995; Urso et al. 2015). Furthermore, higher number of occupants also led to increased respiration which in turn increased indoor CO₂ concentrations. In homes where there is higher human occupancy, high CO₂ levels are typically taken to be indicative of poor ventilation.

The current study, however, has some limitations that could be addressed in future research. One limitation of this study is that a few air pollutants were only measured in the living rooms of homes under investigation. Another issue is that air measurements were undertaken in only one room, and hence may not be representative of the whole household. Although several major determinants that influence indoor pollutants’ concentrations were identified and assessed in the present study, other possible determinants may need to be considered. It should be noted that lack of resources and budget constraints were the main reasons for these limitations. However, this study can be a good foundation for more elaborate future research with a wider scope to enhance our knowledge about indoor air pollution and allow accurate assessment of the problem in order to provide a sound basis for setting proper national IAQ standards.

Conclusion

Indoor and outdoor concentrations of PM₁₀, PM_2.5, CO, and CO₂ were measured simultaneously in 25 naturally ventilated urban homes in Alexandria during the winter and summer seasons. This study revealed that the indoor PM₁₀ and PM_2.5 concentrations were noticeably higher than the WHO guidelines in all selected homes during the two sampling periods. For indoor CO and CO₂ levels, none of the values exceeded the WHO guideline and the ANSI/ASHRAE Standard 62.1, respectively. The effect of seasonality was apparent in this study, where there was a general pattern of increasing average indoor and outdoor concentrations from summer to winter for all target pollutants. The observed seasonal trend in indoor levels of air pollutants is most likely attributed to the different ventilation practice across the two seasons and is obviously reflected by changes in indoor CO₂ levels. Due to frequent ventilation in the summer season, the indoor concentrations of air pollutants were strongly influenced by outdoor sources. In contrast, the levels of indoor air pollution in the winter season were affected more tightly by indoor activities. This is due to the high occupancy, with people tending to spend more time indoors in the winter months than in the summer months, coupled with poor ventilation. The current study also showed that the indoor/outdoor correlations of air pollutants’ levels were higher for all selected homes during the summer than winter. It is likely that the lower indoor/outdoor correlations observed for air pollutants during the winter season were largely due to limited ventilation. Furthermore, indoor combustion sources including smoking and cooking were found to considerably increase indoor concentrations of all selected pollutants in both seasons. It was also observed that the use of exhaust fans or range hoods in the domestic kitchens during cooking led to significant reduction in indoor levels of all target pollutants in both seasons. Another routine indoor activity is cleaning that was found to significantly reduce indoor PM₁₀ levels when undertaken more frequently. Several important determinants were also identified and correlated with indoor levels of PM, CO, and CO₂ in both seasons such as number of occupants and room volume. Other determinants including distance from major roads, home age, and AER were also significantly correlated with indoor PM and CO concentrations in both seasons, whereas a strong association was only realized between AER and indoor CO₂ levels in winter. As this is the first research attempting to study the effect of seasonal variation on IAQ in residential buildings in Alexandria, more elaborate research is required to confirm these results, identify more determinants that influence indoor pollution levels, and assess how these findings can be translated into effective mitigation measures.

Disclosure statement

No potential conflict of interest was reported by the author.

Additional information

Notes on contributors

Mahmoud M.M. Abdel-Salam

Mahmoud M.M. Abdel-Salam is an Associate Professor in the Department of Environmental Sciences, Faculty of Science, Alexandria University, Alexandria, Egypt.