Year long monitoring of indoor air quality and ventilation in school classrooms in Victoria, Australia

Abstract

Children under 15 years spend over six hours per day in school buildings and are particularly vulnerable to the effects of poor indoor air quality (IAQ). However long-term monitoring of indoor environmental conditions is limited in Australian schools. This study evaluates IAQ in ten Australian school classrooms in the temperate settings of Victoria. Temperature, relative humidity, CO₂ concentration, PM_2.5 and PM₁₀ were monitored over a period of one year. The temperature and relative humidity values in the monitored classrooms ranged between 18.7°C and 23.8°C and 44–63 RH%, respectively, indicating that the conditions have been mechanically controlled. The average air change (ACH) rates in the 10 school classrooms varied considerably (range: 0.7–4.83 ACH). The ventilation rates ranged from 1.6 Ls⁻¹ to 11.5 Ls⁻¹ per person with 7 out of the 10 classrooms showing ventilation rates below the recommended minimum. This study points to the need for increasing ventilation rates in classrooms.

KEYWORDS:

• Indoor air quality
• school classrooms
• CO₂ concentration
• ventilation
• air change rates
• particulate matter

Introduction

Given the growing awareness of the adverse impacts of poor air quality there is an increasing interest in improving indoor air quality (IAQ) in school buildings. Children under 15 years constitute approximately 19% of the Australian population (ABS 2019). Students in Years 1–12 receive a minimum of 25 h of instruction per week and spend up to 1075 h indoors in school classrooms each year (DET 2017). This equates to approximately 25% of an average lifetime (Cheryan et al. 2014). Poor IAQ can cause acute and chronic health effects, and vulnerable members of the community such as children are particularly susceptible to adverse respiratory effects (Annesi-Maesano et al. 2012; Clausen, Toftum, and Bek€o 2009; Fsadni et al. 2018; Madureira et al. 2015; Paciência et al. 2019; Simoni et al. 2010). In a review that studied the effects of particulate matter (PM) on young children, short-term exposure to PM was found to adversely impact respiratory health; and for long-term exposure, fine particles (e.g. PM_2.5) represented a higher risk factor than coarse particles (e.g. PM₁₀) (Liu et al. 2018). Chatzidiakou et al. (2014), found that adverse health impacts such as respiratory disease and illness-related absenteeism were reported at concentrations of NO₂, O₃ and CO below the current guidelines. Palumbo et al. (2018) investigated the effect of classroom conditions in Europe and found increased risks for allergy and flulike symptoms in hot classrooms in the heating season, and reduced frequency of allergy symptoms with perceived good outdoor air quality. Given the number of hours that children spend in school classrooms, there is a growing concern about the conditions of indoor air quality in school buildings and how it impacts student health, well-being and learning ability (Csobod et al. 2014; Wargocki and Wyon 2013). Furthermore, the on-going Covid-19 pandemic has brought about an increased focus on indoor air quality in schools due to the significance of viral transmission via small airborne microdroplets.

Poor air quality also impacts the cognitive performance of students. Studies conducted on the schools in United States found a significant relationship between student achievement and behaviour and the thermal environment of a classroom, where air-conditioned schools consistently reported higher achievement (Weinstein 1979; Young et al. 2003). In a recent meta-analysis of 18 studies that investigated classroom thermal conditions and children performance, Wargocki, Porras-Salazar, and Contreras-Espinoza (2019) noted that children’s performance can increase by 20% if classroom temperatures are lowered from 30°C to 20°C. In a more recent review of relationship between air quality and learning outcome of children, Wargocki et al. (2020) found 12% improvement in the speed and 2% improvement with respect to error in the performance of psychological tests when CO₂ concentration was reduced from 2100 to 900 ppm. More recent thermal comfort studies of New South Wales’ schools in Australia found neutral and preferred temperature for children lower than the same for adults (de Dear et al. 2015). Previous studies on child health and pollution in Australia (Barnett et al. 2005; Knibbs et al. 2018) typically rely on exposures measured through outdoor stations which do not adequately characterize the personal exposure to indoor air pollutants. Laiman et al. (2014) investigated the particle concentration of 25 naturally ventilated schools in the humid subtropical Brisbane and found that overall air change rates during school hours were higher than the acceptable guideline of 0.7/h by ANSI/ASHRAE Standard 62.1. The air change rate was higher during warmer months compared to colder months. However, in the cool temperate Victoria, high indoor CO₂ concentrations (> 2700 ppm) were found during winter due to inadequate ventilation (Luther and Atkinson 2012). This is in agreement with the New Zealand studies where ventilation rates were found to be 16 times lower than the recommended value and CO₂ concentrations exceeded 3500 ppm (Wang et al. 2016). Luther, Horan, and Tokede (2018) assessed 24 classrooms from six different schools during the winter period and found that air change rates, carbon dioxide exhalation rate and the number of students in the classroom are the three most important parameters for the CO₂ concentration of a school classroom during its occupancy. None of the Australian studies performed longitudinal measurements of air quality parameters. Moreover, there has been an increasing tendency to retrofit naturally ventilated classrooms with air conditioning units due to community complaints about thermal discomfort during the summer. This study addresses these gaps in the literature. It investigates IAQ factors (i.e. CO₂, PM) in school classrooms in the temperate setting of Victoria.

Methods

Ten classrooms in five schools (classrooms A and B from schools S1, S2, S3, S4 and S5) were selected for the study. Out of the five schools, four (S1, S2, S3, S4) are primary schools and one secondary school (S5). The study received ethics approval from the University and Department of Education and Training. This study did not include student perceptions of air quality using a questionnaire although subjective assessment through monitoring of student attention and concentration was part of the study. However, the focus of this paper is IAQ assessment through objective measurements. The IAQ parameters of air temperature, relative humidity, carbon dioxide (CO₂) concentration levels, PM_2.5 and PM₁₀ concentration levels were continuously monitored for one school year from February to December 2019. Both long-term (i.e. one year) and weekly measurements were conducted. For the long-term monitoring, battery-operated data loggers (HOBO MX1102) were used and set to 15-minute sampling intervals (ASHRAE 2009). The sensors on these loggers were air temperature, relative humidity, carbon dioxide (CO₂) concentration levels and the unit was placed on a wall at least 2m from the closest student table (DIN 2007) away from doors and windows, and installed at 1.2–1.5 m high close to breathing zone height of students and not directly underneath supply air diffusers. All the sensors were co-located in a room for cross comparison of their measurements and assess their consistency and reliability. Particulate matter was also monitored for one week during each of the four school terms (representing summer, autumn, winter and spring seasons) using an Aerocet handheld particle counter Model 531S. The datalogger, installed close to the HOBO datalogger at 1.0–1.2 m high, was programmed to record data at 15-minute sampling intervals. A TSI DustTrak 8533 was also placed adjacent to the Aerocet at table-top height of 0.85 mm, to provide comparative data and set the k factor required for Aerocet particle counter. Figure 1 shows the typical setup in the classroom. Instrumentation specification is shown in Table 1 and typical layout of equipment set up in the classroom is shown in Figure S1 in the supplementary files.

Figure 1. Typical equipment set up in the classroom.

Table 1. Instrument specifications.

Type of instrument	Parameter	Measuring range and accuracy
HOBO MX1102	Air temperature (°C)	±0.21°C from 0° to 50°C
	Relative humidity (%)	±2% from 20% to 80%
	CO2 concentration (ppm)	±50 ppm ±5% of reading at 25°C
Aerocet Handheld Particle Counter Model 531S	PM2.5 PM10	± 10% ± 10%
TSI Dust Trak 8533 Aerosol Monitor	PM2.5 PM10	Particle size range 0.1–15 µm Aerosol concentration range 0.001–150 mg/m^3 Resolution ±0.1% of reading or 0.001 mg/m^3 (whichever is greater)
Testo 480	Air temperature (°C) Relative humidity (%) CO2 concentration (ppm) Globe temperature (°C) Air velocity (ms^−1)	0 to +50°C, ±0.5°C 0 to 100RH%, ±(1.8RH% + 0.7% or measured value) 0 to +10,000, ±(75 ppm CO2 + 3% of measured value) 0 to +120°C, Class 1 0 to +5 ms^−1, ±(0.03 ms^−1–4% of measured value)

The air change rate (ACH) for each term was calculated. Three different methods namely build-up method, steady-state method and decay method are available for utilizing CO₂ as a tracer gas for determining the air exchange rate in a given space (Hänninen 2013). At very low air change rates (which may be the case for some of the classrooms in this study), it can take long time to reach the steady state. However, for consistency across the school classrooms, the steady-state method was used for calculating the ACH in each classroom. Steady-state method uses the peak-analysis approach based on a mass balance model (Seppänen 2006; ASTM 2018; Roulet and Flavio 2002) as shown in Equation (1). The peak-analysis approach assumes that CO₂ concentration reaches steady state in classrooms. (1) $a=(Ng/V)/(C_s\text{--}{C}_o)$(1) where a = air change rate, h⁻¹ N = number of people (occupants) in the space g = indoor CO₂ generation rate per person, mLs⁻¹ V = space volume, m³ C_s = Final (steady-state) CO₂ concentration in the space, ppm C_o = CO₂ concentration in outdoor air.

In this study, the average CO₂ concentration recorded in the classrooms during the school hours was used as the steady-state value of CO₂ (C_s). CO₂ concentration at Cape Grim (BoM and CSIRO 2018), was used as base outdoor concentration (C_o). The values recorded at Cape Grim is 407 ppm (BoM and CSIRO 2018), and this is found to be equivalent to urban concentrations in Melbourne based on measurements conducted in 2004 by Coutts, Beringer, and Tapper (2007). The CO₂ generation rates of 0.00285 Ls⁻¹ for children and 0.0052 L/s⁻¹ for the teachers in the occupied classrooms were taken according to ASTM D6245 (2018). These CO₂ generation rates correspond to the average ages of 6–11 for students and 21–50 for the teachers, with an activity level 1.2 met (ASTM 2018, Table 4, p5). The air flow rates in ls⁻¹ were calculated from the ACH.

Participating schools

Four of the schools are located in Melbourne area at distances of 3, 7, 13 and 17 kms from the central business district (see Figure S2 in supplementary files). School S1 is in regional Victoria (85 km from Melbourne). The locations belong to Cfb climate as per Koppen–Geiger classification system. Table S1 (supplementary file) shows the external weather conditions corresponding to the four school terms. Monthly average, minimum and maximum air temperatures and relative humidity recorded at 9 am and 3 pm are reported in the table. School S1 is in a residential suburban area south of Geelong CBD. School S2 is located in a residential suburban area southeast of the CBD next to a busy thoroughfare and a bus stop. School S3 is in an inner-city residential area east of the CBD. Similar to S2, School S4 is located in the residential suburban area northwest of the CBD but next to a minor residential road. School S5 is in a residential suburban area north of Melbourne CBD. The five school buildings were built in the last 10–20 years and the building structure and construction are typical of the school building stock in terms of age, typology and construction specification.All schools are provided with a mechanical air conditioning system. Classrooms in S2 and S3 were also equipped with supplementary electric convection heaters. Table 2 shows the physical properties including area, volume, construction details and type of ventilation systems in each of the classrooms.

Table 2. School classroom characteristics.

School	Classroom	A (m^2)	CH (m)	V (m^3)	Construction	Ventilation
S1	S1A	75.5	2.3–3.3	202.4	Ground floor location; standalone building; metal cladding walls with insulation; perforated metal ceiling; east and west windows; loop pile carpet tile floors	Relocatable classroom, Openable windows – west rear windows are always kept close (windowpanes are covered with student work); east front windows and doors were at times kept open; split air conditioning system; ceiling fans
	S1B	75.5	2.3–3.3	202.4	Ground floor location; standalone building; metal cladding walls with insulation; perforated metal ceiling; east and west windows; loop pile carpet tile floors
S2	S2A	81.0	2.6	210.6	First level location; fibre cement cladding walls; clear single glazed north-east windows with micro-venetian blinds; loop pile carpet tile floors	Openable windows – windows closed in S2A and mostly open in S2B, doors open in S2A and S2B; cassette type (ceiling mounted) split air conditioning – reverse cycle; supplementary floor mounted electric convection heater.
	S2B	81.0	2.6	210.6	First level location; fibre cement cladding walls; clear single glazed north-east windows with micro-venetian blinds; loop pile carpet tile floors
S3	S3A	58.7	3.0	176.2	First level location, brick veneer walls; clear single glazed north windows; loop pile carpet tile floors	Openable windows – windows and doors closed in S3A, windows and door open in S3B; cassette type (ceiling mounted) split air conditioning – reverse cycle; supplementary floor mounted electric convection heater; ceiling fans
	S3B	54.9	3.0	164.7	First level location, brick veneer; clear single glazed south windows; loop pile carpet tile floors
S4	S4A	80.0	2.7	216.0	First level location, brick veneer walls; clear single glazed south windows; loop pile carpet floor finish	Openable windows – windows closed, doors sometimes open to the corridor in S4A and S4B; cassette type (ceiling mounted) split air conditioning – reverse cycle
	S4B	81.3	2.7	219.5	First level location, fibre cement cladding walls; clear single glazed north windows; loop pile carpet floor finish
S5	S5A	61.7	2.9	178.0	Ground floor location; brick veneer walls; no windows; loop pile carpet floor finish	Classrooms not provided with windows, south facing doors are often open to external corridor; cassette type (ceiling mounted) split air conditioning – reverse cycle; ceiling fans
	S5B	63.0	3.10	195.4	Ground floor location; brick veneer walls; no windows; loop pile carpet floor finish

Results

Indoor thermal conditions

The school classrooms were continuously monitored for an academic year. The indoor air temperature and relative humidity of the 10 classrooms during school hours (from 9:00am to 3:30pm) across the four school terms are shown in Table S2. In Term 1, the average indoor air temperatures, were generally consistent across the 10 classrooms ranging from 21.8°C to 23.8°C with average relative humidity levels of 50–55 RH%. Lower average temperatures were measured in Terms 2 and 3, 18.9–21°C and 18.7–21.1°C, respectively, with corresponding humidity levels of 51–63 RH% and 44–61 RH%. Term 4 indoor conditions were similar to Term 1 with an average indoor temperature range of 20.1°C to 22.3°C and 49–56 RH%. The indoor air temperature and relative humidity profiles are shown in Figures 2 and 3, respectively. The figure shows that the middle 50% of measured conditions for both indoor air temperatures and relative humidity levels are consistent across the four school terms in the 10 classrooms. Both mean and median measures are closely aligned. These indicate that the conditions in the monitored classrooms have been mechanically controlled.

Figure 2. Variations of air temperature in the school classrooms for four school terms during school hours (9:00am–3:30pm).

Figure 3. Variations of relative humidity in the school classrooms for four school terms during school hours (9:00am–3:30pm).

Air quality parameters

CO₂ concentration levels

The CO₂ concentration levels in the 10 classrooms across the four school terms are shown in Figure 4 with mean values ranging from 657 to 2235 ppm (Table S2). The CO₂ concentration levels varied significantly between the five schools with both classrooms in S3 and S5 showing the lowest levels, 912–1179 ppm in S3A and 925–977 ppm in S3B, 657–1019 ppm in S5B and 671–980 ppm in S5A, and similarly in S2B with 882–1247 ppm. Both classrooms in S4 showed the highest average CO₂ concentration levels ranging from 1519 to 2235 ppm. However, all 10 classrooms showed maximum levels of CO₂ concentrations over 1700ppm during school hours. The highest maximum CO₂ concentration levels were recorded in both S4 classrooms ranging from 3415 to 5000 ppm across the four school terms. Both School S1 classrooms showed high maximum levels ranging from 3033 to 4216 ppm and both S2 classrooms also showed high maximum levels ranging from 1417 to 4195 ppm. Figures 5 and 6 show the hourly CO₂ concentration levels averaged for each term for classrooms S3B (with lowest CO₂ concentration levels) and S4A (with highest CO₂ concentration levels), respectively. The concentration levels start building up from 9 am as classes begin and the highest levels were found during mid-day, just before the lunch break. Another peak was also observed around 3.30 pm just before school ends. These peak levels of CO₂ concentration were similar across the five schools and typical in occupied schools. Based on the reported classroom occupancy and teacher records, operation of air conditioning systems, opening of windows as well leaving doors wide open may have contributed to the fluctuations and peak levels of the indoor CO₂ concentration levels.

Figure 4. CO₂ concentration levels in the school classrooms for four school terms during school hours (9:00am–3:30pm).

Figure 5. Hourly average CO₂ concentration levels for S3B.

Figure 6. Hourly average CO₂ concentration levels for S4A.

Calculation of ventilation rates

The six classrooms have floor area sizes ranging 55–81 m² and volume of 165–220 m³ (Table 1) occupied by 15–27 school children (Table 3). The summary of the calculated air change rates (h⁻¹) and air flow rates (l s⁻¹) during school hours across the four school terms are shown in Tables 3 and 4 respectively. The calculated ventilation rates ranged from 0.7 ACH to 4.83 ACH. For S4 classrooms with the highest CO₂ concentrations, the air change rates during the four terms were 0.7 ACH to 1.14 ACH. The S3B classroom in school S3 has air change rates of 2.85 ACH to 3.12 ACH. Highest ACH of 4.83 was observed in S5A classroom. The occupant behaviour in terms of opening and closing of doors and windows has a major influence on the ventilation rates. Windows and doors were left open in classroom S2B most of the time. Classroom S3B and S5A have the highest level of ACH rate throughout the year and as seen in Table 3. The teacher leaves the windows slightly open and also leaves the door to the corridors open most of the time for S2B and S3B enabling cross ventilation. Windows are closed in both S4A and S4B whereas doors are left open to the corridor in the first floor. This did not help in reducing the CO₂ concentration levels. Comparing the ACH across the four terms, it can be seen that summer months (T1) have higher ACH than winter months (T3) in all the school classrooms except S5 which is a secondary school. In S5, both classrooms are situated in the ground floor. The doors which pen to outside are kept open frequently for students to move from one classroom to other during change of subjects.

Table 3. Calculated air change rates.

				Air change rates (h^−1)
School	Classroom	V (m^3)	N	T1	T2	T3	T4
S1	S1A	202.4	27	1.48	1.23	1.11	1.32
	S1B	202.4	26	1.44	1.31	1.19	1.26
S2	S2A	210.6	22	1.17	0.86	1.01	1.03
	S2B	210.6	22	2.25	1.71	1.28	1.74
S3	S3A	176.2	15	1.73	1.44	1.32	1.13
	S3B	164.7	26	3.12	2.94	2.85	3.02
S4	S4A	216.0	27	0.83	0.72	0.70	1.03
	S4B	219.5	27	0.88	0.87	0.99	1.14
S5	S5A	178.0	22	2.21	3.02	4.83	2.26
	S5B	195.0	17	1.46	2.37	3.59	1.61

Table 4. Calculated ventilation rate values per student.

			Ventilation rate (Ls^−1 ·p)
School	Classroom	N	T1	T2	T3	T4
S1	S1A	27	3.1	2.6	2.3	2.8
	S1B	26	3.1	2.8	2.6	2.7
	Average		3.1	2.7	2.5	2.8
S2	S2A	22	3.1	2.3	2.7	2.7
	S2B	22	6.0	4.5	3.4	4.6
	Average		4.5	3.4	3.1	3.7
S3	S3A	15	5.6	4.7	4.3	3.7
	S3B	26	5.5	5.2	5.0	5.3
	Average		5.6	4.9	4.7	4.5
S4	S4A	27	1.9	1.6	1.6	2.3
	S4B	27	2.0	1.9	2.2	2.6
	Average		2.0	1.8	1.9	2.5
S5	S5A	22	5.0	6.8	10.9	5.1
	S5B	17	4.6	7.6	11.5	5.1
	Average		4.8	7.2	11.2	5.1

Table 4 shows the ventilation rates per person. The average calculated ventilation rate per person ranged from 1.6 Ls⁻¹ to 11.5 Ls⁻¹. In comparison to the values found in a similar study in naturally ventilated classrooms in Brisbane (Laiman et al. 2014), the ventilation rates in this study were higher. However, ANSI/ASHRAE Standard 62.1 recommends 5 Ls⁻¹·person as minimum ventilation rate for classrooms (ASHRAE 2016). The resulting ventilation rate per person in this study as shown in Table 4 were <5 Ls⁻¹ per student in majority of the classrooms except S5, indicating insufficient ventilation. School S4 shows average ventilation rates of 1.8–2.5 Ls⁻¹ per student. Both classrooms in school S3 have ventilation rates ranging from 4.5 to 5.6 Ls⁻¹ which marginally meet the minimum ASHRAE Standard 62.1 requirement. The classrooms in S3 also exceed the specified ventilation rate of 4 Ls⁻¹ per child established in the SINPHONIE project (Csobod et al. 2014), whereas this specified rate marginally aligns with those found in S2, with average values ranging from 2.9 Ls⁻¹ to 3.8 Ls⁻¹.

Particulate matter

Figures 7 and 8 show the range of PM₁₀ and PM_2.5, respectively, during each one-week monitoring period. The PM levels varied considerably between the five schools. Mean PM_2.5 levels ranged from 6.3 to 39.9 µg/m³ and mean PM₁₀ levels ranged from 7.23 to 47.17 µg/m³ across the four terms. Term1 which is predominantly summer-autumn period showed higher levels of particles, particularly finer PM_2.5 particles. External PM levels were also monitored for a day during Terms 2, 3 and 4 and the results are shown in Table S3 in supplementary files. The external weather conditions had significant influence on outdoor particle concentration levels and the lowest levels were observed during Term 2 (autumn/winter period) as a result of rainy days. The indoor–outdoor ratios (I/O) were calculated for room B in each school and this was estimated to be 0.5–4.5 and 0.2–16.6 for PM_2.5 and PM₁₀, respectively (Figures 9 and 10). The ratio was generally lower for S2 and S5 compared to S1, S3 and S4 suggesting lower indoor pollutant levels and possibly a tighter building envelope in S2. This also corresponds to the higher ventilation rates in S2 and S5 as shown in Tables 3 and 4. The indoor PM levels depend on the HVAC system removal rate by ventilation and filtration, envelope tightness, the amount of pollutants already inside the room, number of occupants, type of activities, etc. (Cheong and Lau 2003).

Figure 7. PM₁₀ levels in the schools during four school terms.

Figure 8. PM_2.5 levels in the schools during four school terms.

Figure 9. PM₁₀ indoor–outdoor ratio.

Figure 10. PM_2.5 indoor–outdoor ratio.

Discussions

Indoor carbon dioxide (CO₂) concentration measurements are commonly used as indicators of indoor ventilation in relation to the occupancy of the room. When concentration levels exceed 1000 ppm, insufficient ventilation and unacceptable conditions in relation to odours removal are indicated (ASTM 2018). Outdoor carbon dioxide (CO₂) concentration levels typically range between 400 and 500 ppm, and typical indoor CO₂ concentration levels range between 500 and 1500 ppm (Seppänen 2006). The commonly referenced guideline value for CO₂ of 1000 ppm is based on the 650 ppm concentration difference with the outdoor CO₂ concentration of 350 ppm (Seppänen 2006). However, with outdoor CO₂ concentrations increasing (NOAA 2020), the indoor-outdoor concentration difference of 650 ppm would translate to a higher concentration than 1000 ppm. The difference of 480 ppm to 1833ppm above outdoor levels found in this study indicate that IAQ in these 10 school classrooms can be categorized as ‘Medium’ (IDA 2, 400–600 ppm) for S3 and S5; ‘Low’ (IDA 4, >1000 ppm) to ‘Acceptable’ (IDA 3, 600–1000 ppm) for S1 and S2; and ‘Low’ (IDA 4) with over 1000 ppm for S4, following the classifications of IAQ according to EN 13779 (CEN 2007). ASHRAE Standard 62.1 (2006) recommends a steady-state CO₂ concentration in a space no greater than about 700 ppm above outdoor air levels with ventilation rate to be held to 5 Ls⁻¹ per person for classrooms (age 9 plus). CIBSE BB101 (2018) Guidelines stipulate that mechanical system must adhere to daily average of 1000 ppm and natural ventilation system cannot exceed a daily average of 1500 ppm. Australian Standard AS 1668.2 (2012) specifies a minimum floor area requirement per occupant and recommends minimum outdoor airflow rate between 10 and 12 Ls⁻¹ per person. For example, minimum floor area of 2 m² per person in classrooms and 12 Ls⁻¹ per person serving persons up to 16 years of age. However, there is no information on minimum CO₂ concentrations or other indoor air pollutants exposure levels. As cited in MOE (2007), New Zealand standards (1990) specifies 8 Ls⁻¹ per person as fresh air requirement in a class with an estimated maximum occupancy of 50 students per 100 m², and adopts the benchmark of 1000 ppm CO₂ concentration levels.

Most of the previous studies on IAQ in school classrooms were conducted for a short period of time. This study investigated the indoor conditions in classrooms for a whole academic year representing four terms covering four seasons. The results in this study agree with other studies conducted in schools with markedly lower ventilation rates compared to the minimum recommended value as defined by ASHRAE Standard 62.1. In naturally ventilated schools in Portugal, ventilation rates of 2.4 Ls⁻¹ per student were observed in winter (Canha et al. 2013). Similar results were also found in a study of primary schools in Greece, where average ventilation rates ranged from 0.8 Ls⁻¹ to 2.4 Ls⁻¹ per student in non-heating period and even lower ventilation rates of 0.1 Ls⁻¹–1.1 Ls⁻¹ per student during the heating period (Kalimeri et al. 2016). A study of school classrooms in northern Italy during the cold period found a median ventilation rate of 3.0 Ls⁻¹ per student with a range of 0.7–7.7 Ls⁻¹ per student (Rovelli et al. 2014). In a study of 140 classrooms in Southwestern United States, ventilation rates ranging 3.1–4.5 Ls⁻¹ per person were found (Haverinen-Shaughnessy and Shaughnessy 2015). Hänninen et al. (2017) in a study of Albanian classrooms observed average ventilation rates ranging from 0.8 to 3.6 (median 1.8) Ls⁻¹ per student.

The ACH in this study varied significantly between schools with higher ACH observed in the summer months than winter months except in S5 which is a secondary school. Furthermore, two classrooms in the same schools showed different trends in different seasons. This shows that many other occupant-related factors such as the opening and closing of classroom doors and windows as well as operation of air conditioning systems may have contributed to the differences in air change rates. In Albanian schools, very low ventilation rates during the cold season were associated with the lack of heating and uncomfortable low temperature in classrooms (Hänninen et al. 2017). However, all the classrooms in this study were installed with heating system and some of them were retrofitted with split air conditioning systems which are switched on by the class teachers during the summer months. Yet, these artificial heating and cooling systems did not provide sufficient ventilation to the classrooms. Also, it was noted that in S2B and S3B, the class teachers left the windows slightly open throughout the year and this may be the reason for consistently higher ACH in those classrooms. For an average classroom size of 200 m³ volume and 24 students, 5 Ls⁻¹ per person as recommended by ANSI/ASHRAE Standard 62 is equivalent to an air change rate of 2.16 ACH. This was met only for three out of the 10 classrooms investigated in this study.

PM_2.5 levels were consistently high during the summer-autumn period in Term1 whereas PM₁₀ was higher in Term1 only in S2 and S5. PM levels were generally higher in S4 that had the lowest ACH during the four seasons. WHO (2006) recommends maximum levels of 20 µg/m³ per year and 50 µg/m³ per 24-hour for PM₁₀ levels; for PM_2.5, the recommended levels are not more than 10 µg/m⁻³ per year and 25 µg/m³ for 24 h. As there are no guidelines for the levels of indoor pollutants in Australia, outdoor (ambient) guidelines are normally used as a reference. It was found that 24 h average were within the WHO guidelines for all the schools. It was noted that I/O levels were lower for S2 and high for S4. This is mainly because outdoor PM levels were lower in S4. Outdoor PM_2.5 levels increased in the morning drop-off period, afternoon pick-up time and also when students were outside playing in the playground. However, no consistent patterns were observed between various seasons. In a study involving two primary schools in Greece, Kalimeri et al. (2016) found that I/O ratio for PM_2.5 ranged from 0.3 to 7.9. The authors state that the high I/O ratio values can be attributed to either higher indoor levels due to resuspension from indoor activities or to the relatively low outdoor concentration during heating period. Rovelli et al. (2014) in a study of seven schools in Milan found that the ratios between the five-day indoor and outdoor PM_2.5 levels ranged from 0.48 to 1.50. The authors found that particle fractions affected by indoor resuspension phenomena are a concern in school classrooms. Trompetter et al. (2018) investigated the effect of ventilation on PM levels in two classrooms in a New Zealand school, one ventilated and the other non-ventilated, for three weeks period during winter. PM₁₀ concentration during school hours was found to be significantly higher, but PM₁₀ concentrations were 66% lower in the ventilated classroom as compared to the unventilated control classroom. The authors noted that the increased PM₁₀ was from the soil tracked in from outside on children’s footwear and re-suspended during activities within the classrooms and recommended that carpet cleaning and dust reducing carpet should be used to mitigate dust. High activity in densely occupied classroom spaces causes indoor generation of particles (Polidori et al. 2013). MacNeill et al. (2016) found that traffic-related pollutant concentrations can significantly reduce by changing the timing of ventilation in schools located near major roads. Polidori et al. (2013) state that for minimizing exposure of children to indoor pollutants originated from outside, highly effective air filtration devices may be installed in schools. Identification of specific sources of particles would require further control experiments targeting particular activities, extended to outside school hours.

Limitations

This study has limitations particularly associated with PM monitoring and subjective assessments of perceptions. Systematic monitoring of PM concentration levels with respect to indoor activities is required for a comprehensive understanding of indoor and outdoor pollutant sources. It was noted that PM concentration levels were influenced by the local activities within the room rather than outdoor sources. Close monitoring of PM levels with the existing HVAC system turned off can help to understand the main pollutant levels. Additionally, a questionnaire survey will help to understand how the students perceive the indoor environment. Using steady-state method for ventilation rate estimation is another limitation of this study. As steady state may not be reached and maintained in all the classrooms, this should be confirmed by plotting the time series CO₂ concentration in each term in each of the classroom. Subsequently, ventilation rates should be estimated using build-up and decay methods to check the variability.

Conclusions

This paper investigated IAQ performance in schools using one-year long field monitoring of selected indoor environmental parameters, particularly CO₂ and PM concentration levels in 10 classrooms in the temperate climate of Victoria, Australia. The monitoring of the classrooms provided a comprehensive understanding of the actual conditions of the classrooms in different school settings and an overview of the various physical and operational factors affecting ventilation rates and air quality in various seasons. The average CO₂ concentration levels in the different classrooms ranged from 912 to 2235 ppm, while the maximum concentration levels reached up to 5000 ppm during certain times of the occupied hours. The ventilation rates were below the recommended standards in seven out of the 10 classrooms investigated. This was consistent across the season in the four school terms with slight improvement in ventilation during the summer months. The PM levels were within the guidelines stipulated by the World Health Organization and found to be very much influenced by the activities and movement of students inside the room, and the opening and closing of windows and doors. Future studies to investigate the sources of PM and practices to minimize particulate production in classrooms are proposed.

This study noted a performance gap between design expectations and building operation where classrooms were retrofitted with heating and cooling systems without any provision of fresh outdoor air. Currently, most schools rely on teachers opening the windows to provide fresh air to students. But that is not always practical during extreme weather events. Furthermore, extreme weather events such as heat waves, thunderstorms and bushfires, as well as planned burning prior to the fire season, are becoming more common due to climate change, resulting in phenomena such as increased smoke haze that affect IAQ in classrooms. An integrated approach is required for designing and operating the HVAC systems. Provision of mechanical or hybrid ventilation should be part of the standards for designing new schools. Such provision of mechanical ventilation caters for adverse weather conditions, when doors and windows are closed. The perceived cost associated with implementing mechanical ventilation strategy can be considered as a barrier, which will be outweighed by the long-term health and socio-economic benefits. It is important to provide clear instructions and operational guidelines for principals and teachers. A summary of the operational guidelines includes:

• Checking if the artificial heating/cooling system has provision for ventilation, otherwise supplying fresh air through artificial ventilation or operable windows;

• Installing CO₂ monitor with traffic light system which can give some indication of when windows should be open;

• Using CO₂ controlled ventilation system where air exchange rate can be automatically reduced during low occupancy;

• Leaving windows left open for 15–20 min during the break when outdoor temperature is not too high or too low.

The findings from this study will inform the operation and maintenance of classrooms and HVAC systems including provision of fresh air. The results of this study indicate the need to further investigate the indoor conditions and air quality in school classrooms for providing guidelines for building construction, envelope air tightness, cooling policy and ventilation standards. The observed poor ventilation in classrooms is likely to be associated with detrimental effects on health and cognitive performance of students. The next stage of the study involves subjective assessment by analysing student attention and concentration data and investigate how IAQ impacts the attention and concentration performance of the students.

Acknowledgements

This project was funded by the Victorian Department of Environment, Land, Water and Planning. We would like to thank the principals and teachers of the participating schools.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability

All data generated or analysed during this study are included in this article (and its supplementary information files).

Correction Statement

This article has been corrected with minor changes. These changes do not impact the academic content of the article.

Additional information

Funding

This work was supported by the Department of Environment, Land, Water and Planning, State Government of Victoria.