Can commercial buildings cope with Australian bushfires? An IAQ analysis

The Australian 2019/20 summer witnessed an extraordinary bushfire season characterised by unprecedented duration, geographical reach and impact. The aftermath of the bushfires includes increased health-related implications on people due to short and long exposure to poor air quality. The current advice from the Australian authorities in such events is to remain indoors, as it was assumed indoor air quality (IAQ) is healthy. This paper examined that assumption and presents the case study of an office building in Canberra subjected to the 2019/20 bushfires, responding to the need of understanding the ability of air-conditioning buildings to cope with such unprecedented and extreme weather events. Measured data for indoor concentration of CO 2 , PM 10 and PM 2.5 recorded a prolonged period of concerning levels, as well as extreme concentration peaks. This poses a significant risk to the occupants’ health. The values showed peaks up to 12 times higher for PM 10 and 24 times higher for PM 2.5 than the recommended critical thresholds. The infiltration factor and protection performance analysis suggest that old filtering systems and low airtightness levels are not optimal in protecting the indoor environment from outdoor air pollutants. Results show that the concentration of the outdoor pollutant significantly exceeded thresholds for a prolonged time, posing a health risk to the population. The case study presented has been partially able to protect its occupants thanks to the heating, ventilation and air-conditioning (HVAC) response, but this brings up the concern about all those spaces where mechanical ventilation may not be installed or have the same filtering system. Evidence calls for an urgent


BACKGROUND
The Australian 2019-20 bushfire season is now known as the 'black summer' (Vardoulakis et al. 2020b). This refers to a series of extreme bushfire events that started in June 2019 and declared contained only in March 2020. A combination of an unprecedented extended period of high temperatures and drought contributed to the increase of magnitude and duration of bushfires and their subsequent impacts (Yu et al. 2020). It is estimated that the fire burnt around 18.6 million ha, with catastrophic consequences to fauna and flora along with significant environmental and economic losses to affected areas. Indeed, the fires destroyed 5900 buildings, threatened the wildlife population due to habitat loss and caused at least 34 fatalities, exceeding any prior bushfire event (Jolly et al. 2015;Ladds et al. 2017).
However, the hidden impact of bushfire lies in the indirect costs associated with the adverse health of bushfire smoke haze, such as increased morbidity and increased hospitalisations, which have been estimated to be around A$1.95 billion for the 2019-20 Australian bushfire (Johnston et al. 2021). Even short-term exposures to smoke haze can exacerbate respiratory conditions, such as asthma and eye irritation, while long-term exposures can impact the cardiovascular system leading to myocardial infarction and cardiac arrest (Reid et al. 2016;Brown 2018;Cheong et al. 2019;Hu et al. 2018), increase in psychological disorders and adverse birth outcomes (Cascio 2018;Finlay et al. 2012;Karthikeyan et al. 2006).
A significant factor that must be considered when assessing the risks associated with bushfires smoke relates to the broad geographical extension where the health impacts may be observed (Adam et al. 2021). The smoke can cover areas hundreds of kilometres away from the actual bushfire zone, reaching major urban centres and, consequently, representing a potential health hazard for millions of people and becoming a significant issue for the public health system (Dennekamp & Abramson 2011).
Unfortunately, it is unlikely, though, those events from 2019-20 will be one-off considering the current trends (Vardoulakis et al. 2020b), as the 2019-20 events only confirmed the warning of previous research predicting an amplification of the global risk of bushfire because of climate change (Yu et al. 2020;Jones et al. 2013). Indeed, the observation of the global fires from 1979 to 2013 highlighted that the length of the bushfire season has been prolonged by up to 20% during this period (Jolly et al. 2015), indicating an increasing trend of more frequent and more severe bushfire events (Keywood et al. 2013;Flannigan et al. 2009). Following this tendency, the number of days characterised by an extreme risk of bushfire will increase by 15-70% by 2050 and by more than 100% by 2100 (Jones et al. 2013). Thus, understanding the implications and the impacts of bushfires on the economy, environment and health is of utmost importance to assess the suitability of the current strategies adopted to reduce the health risk during such events. Australian bushfires highlighted the urgent need for an inquiry into the effectiveness of buildings in protecting occupants from exposure to poor air quality. This paper presents the case study of an office building in Canberra subjected to the 2019-20 bushfires to evaluate the ability of mechanical ventilation systems in providing healthy indoor air quality (IAQ).
585 Brambilla et al. Buildings and Cities DOI: 10.5334/bc.87 when windows and doors are closed, the rate of infiltration is reduced, depending on the level of airtightness of the construction itself (Adam et al. 2021). Recently, several studies focused on residential buildings performance (Zhang & Stewart 2020;Rajagopalan & Goodman 2021;Shrestha et al. 2019;Fisk & Chan 2017). However, considering that people spend one-third of their time at the workplace, the common assumption that air-conditioned commercial buildings can protect their occupants from particulates and other pollutants, especially during extended smoke events, should be investigated. Offices must also be able to offer a safe and healthy environment, protecting their occupants from hazardous outdoor pollution.
Commercial buildings are more likely to rely on mechanical ventilation, offering the potential to prevent pollutants' infiltration. A strategy that can be adopted is to apply positive pressurisation of the indoors by increasing air intake above the exhaust and, hence, reducing infiltration for the outdoors (Davison et al. 2021). However, this may not be possible in all cases, increasing the reliance on filters.
Filtering systems can potentially aid in reducing the circulation of particulate matter (PM) and pollutants from bushfires in the indoor environment (Barn et al. 2016). In standard heating, ventilation and air-conditioning (HVAC) systems, the filters are placed in the return-air ducts, and they only function when the system is in operation. Thus, during bushfire events, the HVAC needs to be constantly switched on during the occupation to allow for adequate PM filtrations (Joseph et al. 2020). The efficiency of the filters in removing the PM from the air is expressed through their minimum efficiency reporting values (MERV) from 1 to 16 (ASHRAE 2017). The higher the MERV ratings, the more efficient is the filter, with high-efficiency particulate air (HEPA) filters being categorised as MERV 16. They have been proved to be more effective at reducing indoor PM concentrations (ASHRAE 2017). Further, ultrafine particles can be efficiently filtered by special air filters with high gas permeance and low-pressure drop (Wang et al. 2017(Wang et al. , 2018. HEPA filters, however, are usually employed in healthcare or industrial settings (Joseph et al. 2020), with MERV filters more likely to be installed in standard commercial buildings.
However, the literature presents contradicting findings on the efficiency of MERV filters during extreme bushfire events. Studies on the IAQ during the 2013 Singapore haze episode demonstrate that mechanical ventilation systems with MERV 7 filters have been unable to offer adequate protection from the outdoor PM, with an estimated removal efficiency of 30% for particles with diameters < 0.1 μm (Chen et al. 2016). Contrarily, some evidence shows that either fan-coil units equipped with high-grade filters (Cao et al. 2016) or portable air cleaners (Stauffer et al. 2020) may be effective in removing indoor PM.
These findings indicate that filter efficiency is key for reducing the indoor concentrations of pollutants that originated outdoors. However, conclusive evidence and research on the commercial sector are still lacking, which represents a significant gap that prevents the generation of informed management strategies.
The ability of the built environment to protect from outdoor pollutions must be considered and evaluated through a holistic approach. Furthermore, remaining indoors is a strategy that becomes less effective if the smoke haze lingers for consecutive days (Brown 2018), as the quality of the indoor air is also severely affected during non-hazy days (Sharma & Balasubramanian 2019). When the high levels of air pollutants persist over weeks, it is impractical and also introduces several side issues. Keeping the windows closed may result in lower air exchange, but can lead to building overheating, higher levels of indoor humidity, higher CO 2 concentrations and accumulation of PM 2.5 trapped indoors (Vardoulakis et al. 2020a;Reisen et al. 2019). When a mechanical ventilation system is available, the high reliance on HVAC for air filtering leads to increased energy consumption (Joseph et al. 2020). It is widely recognised that these extreme weather events are exacerbated by climate change, in both magnitude and frequency (Vardoulakis et al. 2020b). Thus, increased use of HVAC systems would lead to increased carbon emissions with implications for climate change mitigation, resulting in a vicious cycle that may be difficult to interrupt while preserving people's health.

SCOPE
The majority of the research aimed at analysing and evaluating the suitability of the current health advice in cases of extreme bushfire events has focused on residential buildings. Few projects consider commercial building case studies, mainly testing ad hoc filtration systems. A lack of evidence exists on the ability of office buildings to protect their occupants from outdoor pollution using the mechanical ventilation systems already in place.
During the 2019-20 events, the public health recommendation to stay indoors was a very clear response to poor air quality conditions found outdoors. However, a lack of data supports the assumption that IAQ conditions delivered by buildings were actually safe during the extreme events of 2019-20 because much of the attention was placed on thresholds found outdoors and not indoors. This study provides measured data about the performance of the Australian built environment under extreme weather events based on the case study of a governmental building in Canberra during the 'black summer'. This analysis relies upon the monitored indoor concentration of CO 2 , PM 10 and PM 2.5 and it investigates the ability of the indoor environment to cope with the prolonged and extreme outdoor smoke conditions. Considering the likelihood of this event being repeated as a result of climate change, it is important to understand how buildings coped.
The ultimate goal of the analysis is to understand the resilience of the built environment to future events, informing the next generation of health advice and management strategies to be adopted during extreme weather events.

METHODS
This research showcases the performance of a governmental office building before and during the 2019-20 bushfires season, from December 2019 to February 2020. The building is located in Canberra, one of the locations hardly hit by the smoke haze during the 'black summer'. The case study is a fully occupied two-storey mixed-use building, with spaces used for laboratories and offices, and it hosts approximately 450 workers, when fully occupied. The construction technique and architectural design represent the average Australian office building. The building is equipped with a central mechanical HVAC system, estimated to be 15 years old, without a dedicated air filtration stage before the mixing plenum. The HVAC is equipped with F5 filters (AS 1324(AS .1-2001Australian Standard 2001), corresponding to a grade MERV 8-9. During the bushfire events, the ventilation system was promptly set on the heavy filtration of outdoor air, but it was not possible to set small pressurisation of the indoor environment, achieved through lower return air volume and higher outside air volume. Hence, the only active protection from outdoor air pollutants infiltration was offered by the filters. This makes the case study highly relevant because it represents the case of an average office building equipped with a relatively old HVAC system, and not optimally operated, offering the possibility to evaluate the protection offered by the system envelope/HVAC.

INDOOR POLLUTANTS THRESHOLDS
The air pollutants concentrations thresholds are indicated by the National Environment Protection Measure for Ambient Air Quality (Air NEPM) (ADEE 2016) and reported in the Indoor Air Quality Handbook, a publication released by the Australian Building Code Board (ABCB) to provide construction industry participants with non-mandatory advice and guidance on specific topics (ABCB 2018).

CO 2
CO 2 is a natural chemical component of the atmosphere and non-toxic in low concentrations; in standard conditions, the average CO 2 outdoor level in the air is around 400 ppm (ABCB 2018). CO 2 is formed and released in the atmosphere by biological and natural processes, such as 587 Brambilla et al. Buildings and Cities DOI: 10.5334/bc.87 respiration, decay processes and volcanos, but also by anthropogenic activities, such as fossil fuel combustion. The latter is the primary source of CO 2 in outdoor environments, while indoor is usually dominated by human metabolic functions. A high concentration of CO 2 can cause headaches, dizziness, confusion, dyspnoea, disorientation, hypertension and ultimately loss of consciousness (ABCB 2018).
CO 2 is often used as a driving parameter for mechanical ventilation and to monitor indoor occupancy. ASHRAE Standard 62.1-2019 recommends a maximum CO 2 level of 1000 ppm to maintain IAQ (ASHRAE 2019), but the National Construction Code of Australia further reduces it to 850 ppm over an eight-hour period (ABCB 2013). This level is based on an increment of 450 ppm above the background CO 2 concentration, representing an adequately ventilated building (ABCB 2018).
In this study, CO 2 concentrations were used to understand the occupancy of the building and to estimate the mechanical ventilation adopted based on daily values.

PM 10 and PM 2.5
Airborne particles are referred to as particulate matters (PMs), and they include dust, dirt, sand, smoke and liquid droplets, with varying sizes and visibility. They are usually grouped based on their size, dividing between (ABCB 2018): • coarse particle PM 10 : diameter < 10 μm • fine particles PM 2.5 : diameter < 2.5 μm • ultrafine particles: diameter < 0.1 μm.
These types of pollutants can be produced by natural processes or anthropogenic activities. Bushfires and dust storms are among the first, and they are the most relevant for the Australian context due to the natural wildfires and the dust transported from the arid central regions.
The WHO (2009) set the yearly thresholds based on the lowest level associated with a relative increment of lung cancer mortality, which is PM 2.5 25 μg/m 3 and PM 10 50 μg/m 3 . A recent study concluded that PM 2.5 levels four times higher than this value may induce up to 6% of daily allcause mortality, 5% of cardiovascular mortality and 6% of respiratory mortality (Liu et al. 2019). Another recent investigation found a linear correlation between PM 2.5 and mortality from lung cancer; indeed, it reported that for every 10 μg/m 3 increment in fine particle concentration, there is an increased risk of death of 8% (Pope et al. 2002).
In Australia, air quality thresholds defined by the Air NEPM (ADEE 2016) had been updated in 2016 to include both the annual and the daily limits as reported in Table 1.
The 2016 revision also started a process for the reduction of the PM 2.5 limits to 20 μg/m 3 (24 h) and 7 μg/m 3 (one year) to be achieved by 2025, and to develop a population exposure metric that could be nationally consistent and agreed between the different territorial jurisdiction.

MEASUREMENTS
The monitored parameters include indoor and outdoor, CO 2 , PM 2.5 and PM 10 , measured at 1.5 m above the floor. The accuracy of the device is reported in Each room of the building was equipped with a monitoring device, placed in the centre of the area with careful consideration of the displacement regarding supply air diffusers. This study presents the monitored data of an open-plan office area located on the ground floor, because it is representative of a more consistently and homogenously occupied floor across the monitoring period.
The data were collected at one-minute intervals throughout the bushfire season. In January, the indoor device was switched off for the first 15 days, while the rest of the series does not present gaps of more than 15 consecutive minutes, which allowed us to build a complete time series of hourly averages. The gap does not influence the analysis, because the findings are based on the difference between indoor and outdoor concentrations rather than comparing the monthly trends.

RESULTS AND DISCUSSION
3.1 CO 2 Figure 1 shows that the outdoor CO 2 levels were consistently above the standard value of 400 ppm for the whole bushfire season, with a significant peak during January, when the CO 2 concentration increases by almost a factor of 2. The Australian Capital Territory (ACT) Emergency Service Agency (ESA) officially declared the end of the bushfire season on 31 March 2020, but the emergency was under control by February. Figure 1 reveals that the levels of outdoor CO 2 dropped to the normal value as soon as the bushfire was extinguished, indicating that the possible harm for humans in these events is directly related to the presence of the pollutant's source. Although the recorded CO 2 concentration indoors is high, the value never exceeded the critical threshold of 850 ppm as an eight-hour averaged value (ABCB 2013). CO 2 levels never achieved a concerning value to be considered a primary source of harm for human health. Figure 1 also reports that the indoor CO 2 concentrations were constantly above the outdoor levels, which can be expected as CO 2 is also    produced by human respiration and is the main driver for ventilation in buildings. In this case, however, it is possible to notice a discrepancy with the outdoor data. Indeed, the highest outdoor CO 2 median values are reported toward the end of the bushfire season, while it seems the opposite indoors, even if the building has been occupied for most of the bushfire season. These trends are determined by use of the mechanical ventilation in the system, which has been used at its maximum capacity to compensate for the smoke infiltration through the envelope.

PM 10 AND PM 2.5
The concentration of PM in the air is usually expressed through two indicators: the coarse and the fine PM, respectively, PM 10 and PM 2.5 . In the case study presented here, trends observed for the two PM indicators are similar; however, the magnitude and frequency of the thresholds' exceedances must be analysed separately to identify the possible source of health risks and HVAC resilience. Figure 2 shows the outdoor PM concentration during the monitoring period, alongside the maximum and minimum monitored values and the acceptable thresholds. The air pollutants concentrations exceeded the yearly thresholds previously indicated. Although the Air NEPM quality standard (ADEE 2016) admits an exceptional exceedance of the limit, it is still unclear what are its long-term consequences. The 2019-20 bushfire season was unusually long, and PM concentrations outdoors have been consistently above the thresholds for more than two consecutive months. Concerning levels of both PM 10 and PM 2.5 have been detected during the whole summer, with December being the worst month. The daily thresholds are set on the 24 h averages and, during this month, the detected values ranged cyclically from just above the recommended limit to 12 times higher for PM 10 and 24 times higher for PM 2.5 (approximately 600 μg/m 3 in both cases). However, looking at the peak values, the situation appears even more severe, with PM 10 peaks up to 24 times and PM 2.5 peaks up to 48 times higher than the limits (approximately 1200 μg/m 3 ). This trend is observed to continue in January and mid-February when most of the fires were contained, but it asks the question about the health effects of longterm and cyclical exposure to a high concentration of air pollutants combined with short-term exposure to extremely high concentrations. Indeed, the lack of an hourly indication on critical PM concentrations limits the possibility to identify the possible issues at a finer scale and, consequently, to define a more suitable strategy to reduce harmful exposures. This issue is also acknowledged by the Environmental Protection Agency of the State of Victoria (EPA), which states that: there is currently no national standard for the one-hour PM 10 average, and it uses the value 80 μg/m 3 to trigger a 'poor' air quality category.
( EPA 2019) This can also indicate the magnitude of the issue, as PM 10 has exceeded this level for more than one consecutive month. Figure 3 shows a summary of the indoor and outdoor monthly PM 2.5 values recorded. As shown in Figure 2, PM 2.5 and PM 10 are very similar, hence the first can be used as a proxy for the second, and the same conclusions apply. 591 Brambilla et al. Buildings and Cities DOI: 10.5334/bc.87 The first observation regards the scale of the values: outdoor concentrations reached peaks up to 10 times higher than measured values indoors, indicating that the building offered a certain degree of protection from the extreme and most severe moments. Considering the outdoor concentrations, it is possible to notice a difference in the month with the highest peaks and the highest average. The highest peaks were observed during January, but the month with the highest average concentration is December. This may be due to the partial failure of the monitoring system, even if the outdoor data indicate that December has been the worst month for long periods of outdoor exposure, and January posed an increased pressure on the buildings to resist air pollutants' infiltration thereby keeping indoor conditions acceptable. Figure 4 shows the PM 10 and PM 2.5 indoor concentrations in regard to the outdoor average during the extreme peak observed in December.
One of the major issues related to the presence of PM 10 and PM 2.5 is the persistence of these pollutants in the air because they generally take a long time to settle. Rain may help the process, but Australian summers are generally hot and dry, preventing water from settling or dissolving the particles. The fine and light nature of PM 10 and PM 2.5 makes it very easy for these pollutants entrained into the air by wind or disturbances consequently entering buildings through infiltrations, openings and ventilation systems. This is clearly seen in the delay of the peaks.
The daily limits for PM 10 and PM 2.5 , respectively 50 and 28 μg/m 3 (ADEE 2016), are exceeded. This indicates that the building and the ventilation strategy have not been able to provide an adequate level of protection during teak events. Outdoor levels reached everyday peaks up to 1200 μg/m 3 , except for the 20th, where it ranged consistently between 150 and 200 μg/m 3 . From Figure 4 it is possible to notice that indoor trends do not follow the outdoors. Indeed, indoor peaks are observed during the night-time, which poses pressure on the HVAC systems in the morning, just before working hours. These night-time peaks are due to the ventilation schedule employed in commercial buildings, where the HVAC system is usually switched off during non-working hours. However, the data monitored during the bushfire question the suitability of these energy-efficient measures during extreme events, where constant ventilation of the indoor spaces is required.
The high values detected during 20 December are consistent with the delay identified previously, which not only indicates that the building has a slower reaction time (peak of the precedent day), but also that the building and the mechanical system have difficulties in coping with constant pressure, given by the lower but steadier outdoor value detected on the same day.

STATISTICAL ANALYSIS AND INFILTRATION MODELLING
To better assess the ability of the building to provide a healthy indoor environment, the particle infiltration factor (F inf ) and protection effectiveness (PE) are calculated. The particle infiltration factor is averaged daily, dividing occupied and unoccupied hours. The hourly averaged concentration of pollutants has been used to determine the hours of occupancy, which are assumed to correspond to the hours where the HVAC system was in operation. A multiple linear regression is used to determine the coefficient of penetration of outdoor particles (a 1 ) and the coefficient decay of indoor particles (a 2 ) ( Barn et al. 2008;Switzer & Ott 1992), following the equation: where C in is the concentration of PM indoors; C out the concentration of PM outdoors; and S i the indoor PM generation, assumed to be null. Based on (1), the infiltration factor F inf has been determined as: F inf for occupied and unoccupied hours is equal to 0.36 and 0.50. respectively, which means that the infiltration rate is higher during the unoccupied hours when the HVAC system is off. This result is easily understandable because the calculation accounts for the filtering system by assessing the rate of the air pollutants' concentration. These values appear to be aligned with the literature: in a comparative study the seasonal infiltration factor was calculated for 21 homes with and without HEPA filters (Barn et al. 2008), resulting in F inf ranging from 0.61 to 0.28 without filters, and from 0.19 to 0.10 when HEPA filters were installed. The results of the presented case study support the literature, with F inf higher when there is a filtering system in place. 593 Brambilla et al. Buildings and Cities DOI: 10.5334/bc.87 Further, these values can be used to calculate the factor PE, expressed as a percentage: unoccupied -occupied / unoccupied 100 * 0.5 -036 / 0.5 28% (3) This value represents the efficacy of the system building and HVAC to protect indoor occupants from outdoor pollutants' infiltration, and it accounts for the filters' efficiency and the envelope's airtightness. Indeed, PE depends on both the ability of the HVAC system to filter the pollutant penetrated indoors and the ability of the envelope to reduce this penetration. Hence, PE can be a good indicator of the overall building performance, which, in this case, shows a capacity to reduce by 28% the air pollutants' infiltration. Compared with a more sophisticated filtering system, this value is lower than the average (Barn et al. 2008). However, it reflects a typical Australian case, where the HVAC system installed is estimated to be 15 years old and the airtightness of the envelope is not optimal (Ambrose & Syme 2017). This prompts a new question about the necessity to expand the research on bushfire responses and intervention strategies to a broader cohort of buildings, accounting for different ages, quality, design, operational and construction practices.
A multiple regression analysis was also conducted to predict the relationship between outdoor PM 2.5 and CO 2 (independent variables) and indoor PM 2.5 data (dependent variables) for unoccupied and occupied hours ( Table 3).
The multiple regression analyses predict the relationship between outdoor PM 2.5 and CO 2 values, and this correlation is found to be statistically significant (R 2 ranked 0.23, p < 0.0005 for unoccupied hours; 0.14, p < 0.0005 for occupied hours). The R 2 values show that there is a stronger relationship between outdoor and indoor variables during the unoccupied hours compared with occupied ones. This might be since, when the HVAC system was switched off, the windows were kept closed, leaving uncontrolled infiltration as the only exchange between the indoor and outdoor environments. However, during the occupied hours when the HVAC system is active, the analysis shows a weak relationship between the dependent and independent variables. In this case, it is the performance of the filters to determine whether the relationship between the variable is weak or strong. Accordingly, the multiple regression analysis confirms the infiltration models.

SUMMARY OF THE RESULTS
In this study a commercial building in Canberra (AU) was monitored during the extreme bushfire events of 2019-20. The results show the following: • Indoor CO 2 levels do not constitute a source of harm for the occupants' health during extreme bushfire events, mainly due to the high reliance on the mechanical ventilation system, preferred over natural ventilation to protect the indoors and left active for all the occupied hours as a response to the high filtration needs.
• Indoor PM 2.5 and PM 10 have shown similar trends, hence one can be used as a proxy for the other. The main reason is the filtering system: the case study is equipped with F5 filters, which can capture particles with diameters < 1 μm, thus showing a similar efficacy for both air pollutants.
• Outdoor PM consistently exceeded the acceptable thresholds (ADEE 2016) for the whole monitoring period, with averages up to 12 (PM 10 ) and 24 (PM 2.5 ) times higher than the limits, and peak values up to 24 (PM 10 ) and 48 (PM 2.5 ) times higher, indicating a serious health hazard for both long exposures and peak values.  Table 3 Multiple regression analyses of the relationship between indoor PM 2.5 and outdoor PM 2.5 and CO 2 data for occupied and unoccupied hours.

CONCLUSIONS
The 2019-20 Australian bushfire season has prompted unprecedented recorded pollutant concentrations outdoors and indoors, posing a significant threat to human health. The severity and duration of the bushfire season have also placed extreme pressure on buildings and their heating, ventilation and air-conditioning (HVAC) systems to cope with such environmental stress and protect against the smoke for a prolonged period of time. This study contributes to the discussion about the ability of our indoor environments to cope with extreme weather events. It reports on indoor air quality (IAQ): measured levels of indoor air pollutants (such as PM 2.5 and PM 10 ) detected in an office building during the Australian bushfire season of 2019-20.
Monitored levels of indoor air pollutants, such as CO 2 , PM 10 and PM 2.5 , have been compared with the outdoor levels and recommended thresholds to understand the health risks posed on the population during this period. The findings show that concentrations of outdoor air pollutants consistently exceeded the indicated thresholds for more than two consecutive months, with peaks up to 24 and 48 times higher than considered safe limits, indicating a significant high health risk. Besides, it was observed that indoor PM 10 and PM 2.5 concentrations also exceeded the thresholds. The particle infiltration factor (F inf ) and protection effectiveness (PE) calculations indicated that the infiltration system was working effectively when unoccupied and occupied periods are compared, with an overall efficiency of the system envelope and filter up to 39%. The multiple regression analyses showed that outdoor PM 2.5 and CO 2 concentrations had a higher impact on indoor PM 2.5 concentration during the unoccupied hours.
Furthermore, the indoor environment shows a delay in the peaks, suggesting that air pollutants had difficulty settling and, most likely, that they accumulate indoors, further aggravated by the small air exchange with the outdoors. These statistically significant findings highlighted the importance of HVAC and infiltration system, especially under these extreme weather conditions. The high concentration of indoor air pollutants during unoccupied hours overthrows the common assumptions that trivialise the envelope air leakage in office buildings.
The current health policy recommends the public to stay indoors during extreme hazy days, which strongly relies on the assumption that the indoors is healthier and more efficient in protecting from air pollutants. However, Australian buildings are generally leaky and without a mechanical ventilation system. If HVAC is installed, occupants are less prepared and lack instruction on how to manage it during such extreme weather events. According to the results, the monitored office building could reduce high levels of air pollutants, but the concerning prolonged concentration of air pollutants questions the general approach currently adopted.
Considering that the likelihood of these extreme events is increasing due to climate change, understanding how buildings behaved during this period is of utmost importance to identify a suitable strategy to reduce the health risks associated with fire-related air pollutants.
595 Brambilla et al. Buildings and Cities DOI: 10.5334/bc.87 Future research should aim to investigate the performance of a broader set of buildings, drawing on the knowledge and insights gained from this analysis.
The results from this paper suggest that: • the built environment is not equipped to cope with the duration and magnitude such as those from the Australian bushfire season of 2019-2020 • the evidence exposes the unreliability of the current health advice • urgent actions and a thorough rethinking are needed in relation to the protection that buildings should provide (especially IAQ) during extreme events associated with climate change.