v13.1.01 Stewart Evans

International Journal of Ventilation ISSN 1473-3315 Volume 13 No 1 June 2014 ________________________________________________________________________________________________________________________

Building Ventilation Strategies to Protect Public Health during Chemical Emergencies James Stewart-Evans Centre for Radiation, Chemicals and Environmental Hazards (CRCE), Environmental Hazards and Emergencies Department (EHED), Nottingham City Hospital, Hucknall Road, Nottingham NG5 1PB, United Kingdom Abstract Releases of airborne chemicals can rapidly affect wide areas, leading to exposures that may adversely affect public health. A strategy of sheltering indoors has often successfully protected public health, but in some cases it has been ineffective. This paper explores the role of ventilation as one of a number of factors that affect shelter effectiveness. Ventilation directly influences indoor exposure during chemical emergencies, and theoretical incident and ventilation scenarios are used to show how air exchange rates and chemical doseresponse characteristics determine indoor dose and effects on health. For chemicals for which the likelihood of adverse health effects is driven by exposure concentration (peak exposure), sheltering in place for short periods can be an effective protective measure, because the dilution of outdoor air mixing with indoor air reduces the maximum concentration indoors. It is important to minimise ventilation before an outdoor hazard arrives. Delays in maximising ventilation after it has passed are less likely to cause extra harm, since most protection has already resulted from the lower peak exposure concentration indoors during the hazard’s passage outdoors. For chemicals for which the likelihood of adverse health effects is driven by both time and concentration (cumulative exposure), which include those that exhibit a linear dose-response relationship, it is particularly important for people to both promptly minimise ventilation before an outdoor hazard arrives and maximise ventilation after it has passed, in order to minimise their cumulative exposure. Most UK residential properties are naturally ventilated, but there is an increasing trend towards the use of mechanical ventilation in new-build properties that are more airtight than their predecessors. The installation, maintenance and use of such systems pose a number of considerations. The potential for ventilation filters to offer additional protection is worthy of further attention. Key words: ventilation strategies, emergency preparedness, emergency response, chemical emergencies, public health, shelter-in-place. 1. Introduction

offers decreases over time. This is a particular consideration during long-running incidents.

Releases of airborne chemicals have the potential to rapidly affect wide areas, leading to exposures that can adversely affect public health. The emergency preparedness for, and response to, acute chemical incidents has the primary aim of protecting the health of the population by preventing, or minimising, adverse health outcomes such as injuries and deaths.

A building’s air-exchange rate is a critical determinant of the protection that it provides to people sheltering from an outdoor airborne hazard. It is important for people to optimise ventilation during chemical emergencies in order to maximise the protection offered by sheltering. This paper explores the role of ventilation as one of a number of factors that affect shelter effectiveness. It presents examples of the effect of different ventilation strategies on indoor chemical concentrations and chemical dose in two hypothetical release scenarios, a 30 minute release and a 120 minute release, for three chemical dose-response relationships, in order to show the value of adopting a ventilation strategy that maximises the protection offered by sheltering.

Sheltering-in-place is an accepted and commonly implemented protective strategy. It has often successfully protected public health, but in some cases it has been ineffective (see, for example, Dunning and Oswalt, 2007). Sheltering is a particularly effective strategy for public protection in short-lived incidents, but the protection that it

________________________________________________________________________________________________________________________

1

J Stewart-Evans ________________________________________________________________________________________________________________________

2. Discussion

some windows open when outdoor temperatures are comfortable (Johnson and Long, 2005). Air exchange is lower when there are lower temperatures because people attempt to minimise the ingress of cold air from outdoors. In Johnson and Long’s study, at times when the outdoor temperature was less comfortable (38 °C), only 20% or so of the residences that they examined had at least one window or door open. A large-sample survey by Price and Sherman (2006) on ventilation behaviour in new California houses (all built in 2003) also found a similar seasonal dependence in the percentage of residences with open windows. In summer and spring, their survey showed that roughly 40% of homes had at least one window open for >2 h between 6 am to 6 pm weekdays. In winter, only 15% of homes had at least one window open for >2 h during the same time period. Chan (2006) found that the estimated median air-exchange rate of houses in the US varies by a factor of two from its summer high (~1 ac/h) to its winter low (~0.5 ac/h).

2.1 The Effect of Ventilation on Air Exchange

Building ventilation is typically provided using trickle vents, extract vents, mechanical ventilation and other ‘designed’ openings (NHBC, 2009a). They are controllable so that ventilation rates can be increased or decreased to suit the occupier. In less permeable buildings there is a higher requirement for ventilation; airtight buildings are more likely to use mechanical ventilation. In order to examine the effect of ventilation strategies during chemical emergencies, the air-exchange rates associated with modelled normal, minimised and maximised ventilation require definition. Studies in the UK and other countries have shown large variability in residential airtightness and building air exchange rates (Brown, 1997; Stephen, 2000; Chan et al., 2005). This variability has a significant effect when attempting to estimate the level of protection provided by buildings: some studies have used probability distributions when predicting effects (Chan et al., 2007b; Brelih, 2012). The US Environmental Protection Agency noted that air exchange rates varied by US region, recommending a median (rather than arithmetic mean, due to skewing caused by high upper values) air exchange rate of 0.45 air changes per hour (ac/h) for residential properties and a mean value of 1.5 ac/h for nonresidential buildings (EPA, 2011). In the UK, the Environment Agency have used 0.5 ac/h for residential properties, 1 ac/h for offices and 2 ac/h for schools when modelling exposure related to ground gas in non-emergency scenarios (Environment Agency, 2005). US sources have collated estimates of residential air exchange as part of building parameters to inform estimates of indoor exposure to radiation (Biwer et al., 2002), and a number of other studies have examined or estimated the reduced level of air exchange likely to be found in properties when people have implemented sheltering advice. An overview of related studies is provided in Public Health England’s recent report Protective actions in acute chemical and radiological incidents: evacuate or shelter-in-place? (Stewart-Evans et al., 2013 (in press)).

For the purposes of this paper, 1 ac/h is used to represent a normal rate of air exchange. An air change rate of between 0.5 and 1.5 ac/h for a residential dwelling will usually be sufficient to control condensation (The British Standards Institution, 2005); a controlled rate of 0.5 and 1 is recommended as energy efficient (LMU, 2009). When considering the effects of air exchange on indoor exposure, it is important to note that a value of 1 ac/h does not mean that the indoor concentration will equal a constant outdoor concentration after one hour. This is because as air enters a building, interior mixing and exfiltration take place at the same time. Therefore, the time required to replace air inside a building is not a linear function of air exchange. After one air exchange the indoor concentration will still be lower; Sorensen et al. (2004) calculate that 63% of the air will be exchanged after one hour at 1 ac/h and that 95% of the air will be exchanged after three hours at 1 ac/h. 2.1.2 Minimised Ventilation

A rate of 0.5 ac/h is used to represent minimised ventilation. Based on studies examining infiltration and ventilation, this appears a reasonable value to use for typical air exchange due to infiltration alone in UK residential properties built within the last decade and as a lower bound for older buildings in which the occupants have taken steps to minimise

2.1.1 Normal Ventilation

Natural ventilation is highest in the summer months when higher indoor temperatures mean that people are more likely to leave windows and vents open. One study found that up to 35% of households have

________________________________________________________________________________________________________________________

2


be adopted by sheltering populations. These comprise:

air exchange, as the majority of English housing stock is indeed older and shows high variability in air exchange rates (Stephen, 2000; Department for Communities and Local Government, 2012).

1. Optimum ventilation. Ventilation is minimised before the outdoor hazard arrives. Ventilation is maximised as soon as the outdoor hazard has passed. This is an ideal scenario. In practice, knowing when the outdoor concentration becomes less than that indoors and when to ventilate is particularly difficult owing to imperfect real-time information about the release, and the difficulty of predicting transport and dispersion rates.

2.1.3 Maximised Ventilation

Chan (2008) found that mechanical ventilation increased air exchange by 2–10 times compared to modelled air infiltration rates. However, when open, airflows through windows, doors and other designed openings in the building envelope can dominate the air-exchange rates in residences (Chan, 2006). Also open windows are likely to increase air exchange rates to a greater extent than exhaust and heating and cooling fans (Wallace et al., 2002). A number of studies have examined the effects of opening windows on air exchange: Howard-Reed et al. (2002) measured the air-exchange rates in two residences and found that opening one window increased the air-exchange rates by 0.1 to 2.8 h-1 in one residence, and by 0.5 to 1.7 h-1 in another residence.

2. Constant ventilation. Ventilation is unchanged before, during and after the presence of the outdoor hazard. This represents situations in which building occupants are unaware of the presence of an outdoor hazard or are unable to take steps to change building ventilation rates. 3. No purge. Ventilation is minimised before the outdoor hazard arrives. Ventilation remains minimised after the outdoor hazard has passed. This represents situations in which building occupants shelter promptly but do not take steps to terminate sheltering; this may occur if emergency responders do not issue instructions for the public to stop sheltering when a hazard has passed.

A rate of 4 ac/h is used to represent maximised ventilation (i.e., rapid ventilation using openable windows and doors). Current Building Regulations mandate this as the minimum purge ventilation rate per room directly to outside (HM Government, 2010). Rates may, of course, differ.

4. 15 minute ventilation delay. Ventilation is minimised 15 minutes after the outdoor hazard arrives. Ventilation is maximised as soon as the outdoor hazard has left. This represents situations in which there is some delay before building occupants shelter; this may occur if initial instructions for the public to start sheltering are delayed after an incident occurs or due to the time required for people to move indoors and take steps to reduce ventilation. The 15 minute time period is arbitrary – previous studies have found that individual households may take only 5-15 minutes to implement sheltering once they decide to do so (Sorensen, 1988; Sorensen et al.; 2004, Argonne National Laboratory, 2006) – in reality this delay could be shorter or longer.

2.2 The Effect of Ventilation on the Effectiveness of Sheltering Indoors from an Outdoor Airborne Chemical Hazard

A successful sheltering strategy requires that two distinct actions are taken without delay to maximise the passive protection a building provides: •

Reducing the indoor-outdoor air exchange rate before an outdoor hazard arrives. This is achieved by closing windows and doors and turning off fans, air conditioners and combustion heaters. Once this has been done, the residual air exchange rate will depend on a building’s airtightness and infiltration driven by wind pressure and the indoor-outdoor temperature differential.

•

Increasing the indoor-outdoor air exchange rate as soon as the outdoor hazard has passed. This is achieved by opening windows and doors and turning on fans to ventilate the building.

5. 15 minute purge delay. Ventilation is minimised before the outdoor hazard arrives. Ventilation is maximised 15 minutes after the outdoor hazard has passed. This represents situations in which building occupants shelter promptly but do not terminate sheltering until after an outdoor hazard has passed; this is a credible scenario

This paper examines the effects of six ventilation strategies that reflect different behaviours that may

________________________________________________________________________________________________________________________

3

J Stewart-Evans ________________________________________________________________________________________________________________________

Clearly, this is a simplification. Buildings can consist of a number of compartments (rooms) each subject to their own air exchange rates; mixing is not instantaneous. This model does not account for attenuation processes (such as filtration, deposition and sorbtion) that can reduce indoor concentrations and increase the time over which sheltering is effective. As such, it overestimates indoor concentrations – in reality, indoor exposures are likely to be lower than those presented in these examples, which are intended to illustrate the role of air exchange and ventilation. Given the assumptions described previously, Figure 1 shows the effect of the six ventilation strategies on indoor concentration over a 120 minute period, the first 30 minutes of which a building is exposed to a constant elevated outdoor concentration of a hazard before the outdoor concentration returns to zero, representative of a short-lived release.

because the optimum time to end sheltering will vary between properties and no one time is best for all. Because it is rarely feasible to stagger “stop sheltering” messages, emergency responders are likely to issue blanket advice to all properties. Again, in reality this delay could be shorter or longer. 6. 15 minute ventilation and purge delay. Ventilation is minimised 15 minutes after the outdoor hazard arrives. Ventilation is maximised 15 minutes after the outdoor hazard has passed. This represents situations in which there is a delay both before building occupants shelter and before they stop sheltering. Indoor concentrations are calculated using a simple well-mixed-box model (as described by Chan et al., 2007a). This is based on a mass-balance. The outdoor concentration is taken to be one value representative of a homogeneous outdoor concentration around the building shell. The building is represented by a box (one room) subject to a set air exchange rate. The outdoor and indoor concentration and air exchange rate are used to determine indoor concentration over time by calculating ingress (the chemical mass entering the building as outdoor air moves indoors) and egress (the chemical mass leaving the building as indoor air simultaneously moves outdoors), assuming that mixing of air indoors is instantaneous.

After its arrival, the airborne hazardous substance enters the building as air from outdoors replaces air indoors. Air exchange is a limiting factor that means that the peak (maximum) concentration that people are exposed to is lower indoors than outdoors, with lower air exchange rates associated with lower peak indoor concentrations. The highest level of protection from peak concentration is offered by reducing ventilation before an outdoor hazard arrives (scenarios 1, 3 and 5). Delays in reducing ventilation are undesirable as they lead to

Figure 1. The effect of ventilation strategies on indoor concentrations (30 minute release). ________________________________________________________________________________________________________________________

4


higher peak indoor concentrations, but it is still beneficial to reduce ventilation whenever outdoor concentrations are higher than those indoors, as this will lower the peak indoor concentration. Once an outdoor hazard has passed, concentrations indoors are higher than those outdoors. Indoor concentration can be reduced quickest by immediately maximising ventilation at this point. Delays in maximising ventilation are undesirable as they preserve higher indoor concentrations for longer (scenarios 4 and 6), but it is still beneficial to increase ventilation whenever indoor concentrations are higher than those outdoors, as this will lower total exposure. If ventilation remains minimised (there is no purge, scenario 3), the reduced air-exchange rate means that indoor concentrations are maintained for significantly longer periods than other scenarios: in this example the indoor concentration returns to zero within 3½ hours in scenarios 1 and 4-6; if there is no purge (scenario 3) the indoor concentration takes over 17 hours to return to zero.

factor’ (DRF), calculated by dividing the indoor dose by the outdoor dose. An optimum ventilation strategy (scenario 1) minimises both indoor concentration and dose, offering the highest level of protection. Higher air exchange rates during the period when an outdoor hazard exists (scenarios 2,4 and 6) result in higher doses, and in this scenario the delayed reduction of ventilation in scenario 4 and 6 has little effect in lowering dose, partly because the outdoor hazard passes relatively soon after ventilation is reduced (15 minutes later). Delays in maximising ventilation after the outdoor hazard has passed (scenarios 3, 5 and 6) result in higher doses. If ventilation is minimised before a hazard arrives but there is no purge ventilation after it has passed (scenario 3), this can have a significant adverse effect: the dose shows a comparatively high rate of increase after the outdoor hazard has passed due to the slow decline in indoor concentration. This is particularly significant when longer exposure periods are considered, as the indoor dose tends towards the outdoor dose.

When considering the likelihood of adverse health effects, the exposure concentration (c) is not the only consideration. It is also important to consider dose, which Haber’s Law states is a function of concentration and time (c x t). Figure 2 uses the same incident scenario and assumes a linear doseresponse relationship to present the ‘dose reduction

Figures 3 and 4 examine identical ventilation scenarios during a longer-lived release: a 240 minute period, the first 120 minutes of which a building is exposed to a constant elevated outdoor concentration of a hazard before the outdoor concentration returns to zero.

Figure 2. The effect of ventilation strategies on indoor dose response factor (DRF) (30 minute release).

________________________________________________________________________________________________________________________

5

J Stewart-Evans ________________________________________________________________________________________________________________________

resulting from optimum ventilation behaviour (scenario 1) than unchanged, constant ventilation (scenario 2)). If ventilation remains minimised (scenario 3), there is a pronounced effect on indoor concentration, which remains significantly higher than those scenarios in which ventilation is maximised (scenarios 4, 5 and 6) or, to a lesser extent, constant (scenario 2).

Figure 3 more vividly illustrates the importance of timing. When outdoor concentrations are elevated for longer periods, indoor concentrations are higher, all other things being equal. A 15 minute delay in minimising ventilation (scenario 4) results in a higher peak indoor concentration, but this is less significant than for the shorter 30 minute release (i.e., the peak concentration is closer to that

Figure 3. The effect of ventilation strategies on indoor concentrations (120 minute release).

Figure 4. The effect of ventilation strategies on indoor dose reduction factor (120 minute release). ________________________________________________________________________________________________________________________

6


because the TL is more dependent on exposure duration than the behaviour of indoor concentrations over time. Increasing values of n indicate that peak concentrations play an increasing role in toxic effect; adverse effects can be greatly increased by brief exposures to high concentrations. For example, for an n value of 1, if the concentration is doubled, the substance would have the same toxic effect in one half of the time, but for an n value of 2, it would have the same toxic effect in one quarter of the time.

In comparison with Figure 2, Figure 4 illustrates how, even if it is delayed, a reduction of ventilation (scenarios 4 and 6) has a greater effect on dose reduction when the release is prolonged. This is because the delay itself is a smaller proportion of the period when an outdoor hazard is present; hence the increased dose associated with it is less significant. Similarly to Figure 2, constant ventilation (scenario 2) offers the least dose reduction during or after the period when the outdoor hazard is present. A significantly higher dose is experienced if ventilation is not maximised once an outdoor hazard has passed (scenario 3); over longer exposure periods the dose associated with this scenario may even exceed that associated with constant ventilation.

The effect of different hypothetical dose-response characteristics (cumulative, n = 0.5; linear, n = 1; and peak, n = 2 and n = 3) on the level of dose reduction in each of the six ventilation scenarios considered in this paper is shown in Table 1. Note that the n = 0.5 case is included to illustrate the effect on DRF when n < 1. In reality the n < 1 case is rare; most of the toxic chemicals that have been studied by experts have toxic load exponents between 1 and 3 (Chan, 2006; HSE, 2013).

2.3 Interpreting Dose and Predicting Health Effects: The Importance of Dose-Response Relationships

Figures 2 and 4 assume a linear dose-response relationship. However, chemicals have different dose-response relationships; these must be accounted for when considering the effects of ventilation on dose and the likelihood of adverse health effects. Haber’s Law does not apply to all chemicals and toxicological endpoints. It is widely accepted that for some chemicals time-integrated concentrations are not a good indicator of effects such as mortality (Ten Berge et al., 1986; Chan et al., 2007b; Chan et al., 2007a). Dose, and the effect produced by that dose, is still the most important variable, but this effect is often non-linear. To address this, a metric called the ‘toxic load’ (TL) recognises that chemicals elicit different responses over different concentrations (c) and timescales (t), and it is often used to estimate adverse health impacts from acute exposures. The following modification to Haber’s Law describes the toxic effect (or TL): cn x t. In this simple equation, the n value or ‘toxic-load exponent’ is a chemical-specific parameter that characterises the dose-response relationship. It is a measure of the degree to which a given chemical’s toxic effect is dependent on exposure concentration or exposure time, and is dependent on a material’s toxicological properties. It reflects the fact that, for many chemicals, the exposure concentration will eventually determine the effect, and this is not simply a combination of c x t.

Table 1 shows that the different ventilation scenarios’ effect on dose reduction is generally less significant for peak chemicals (n = 2 and n = 3), for which the principal aim is to reduce the exposure concentration by minimising ventilation during the period when an outdoor hazard is present. Nonetheless, scenarios 2 and 4 show how the higher indoor concentrations associated with higher ventilation can lead to markedly reduced protection (i.e., higher DRF values than scenario 1). Delayed purge ventilation after an outdoor hazard has passed has a far less significant effect on DRF for peak chemicals than for linear (n = 1) and cumulative chemicals (n = 0.5), for which maximising ventilation after an outdoor hazard has passed is important to minimise the final dose (illustrated by comparing the DRF at 241 minutes in scenarios 1 and 3). In the hypothetical n = 0.5 scenario, the dose experienced by people sheltering indoors can, in fact, be higher than the outdoor dose in situations when a chemical is present and people continue to shelter. When dose-response is linear (n = 1), over time the indoor dose approaches the outdoor dose in scenarios 2 and 3 (i.e., the DRF tends towards 1). A pragmatic option may be for people to ventilate the building and to move outdoors as soon as the outdoor hazard has passed and outdoor concentrations become lower than indoor concentrations. Calculated DRFs allows one to compare relative levels of protection, but in order to decide whether adverse health effects are likely to occur, a calculated TL must be compared to a toxic load

The effectiveness of sheltering varies significantly with the toxic-load exponent. For chemicals for which n < 1, cumulative exposure is important

________________________________________________________________________________________________________________________

7

J Stewart-Evans ________________________________________________________________________________________________________________________

Table 1. The effect of ventilation and dose-response on dose reduction factor (120 min release).

The 2002 edition of Approved Document Part L1 of the Building Regulations incorporated, for the first time, an explicit air leakage target (Department of Transport Local Government and the Regions, 2002); in recent years, national efforts to improve energy efficiency and reduce carbon dioxide emissions have led to greater airtightness of new houses and use of mechanical ventilation systems, rather than natural ventilation. The European HealthVent project, which aims to develop healthbased ventilation guidelines, has examined ventilation system types in some EU countries, finding a similar trend of decreasing use of natural ventilation systems in favour of mechanical ventilation systems (Litiu, 2012).

limit (TLL). Whilst this concept is not discussed in more detail in this paper, it is fundamental to guidelines developed specifically for acute exposure during emergencies, and it can be used to inform decisions about the acceptability of exposure and to estimate predicted impacts at the personal or population level. 2.4 Ventilation Systems in UK Residential Properties: Their Implications for the Effectiveness of Sheltering Indoors from an Outdoor Airborne Chemical Hazard

The majority of UK buildings are naturally ventilated. Stephen (1998) noted that “the great majority of UK dwellings are naturally ventilated most of the time, by a combination of air infiltration through air leakage paths and ventilation through purpose-provided openings such as trickle vents”. Past air leakage tests conducted by the Building Research Establishment (BRE) indicated that air infiltration played by far the greatest role in most dwellings (i.e., leakiness was the greatest contributor to air exchange), with purpose-provided openings playing a secondary role in all but the most airtight dwellings (Stephen, 1998).

In the UK, buildings constructed between 1980 and 2010 assured ventilation mostly through increasing use of fan-assisted natural ventilation, but also mechanical supply or extract ventilation. Litiu (2012) estimated that mechanical ventilation systems accounted for half of ventilation systems in new houses in 2011. Approved Document F (Ventilation) of the Building Regulations (HM Government, 2010) currently

________________________________________________________________________________________________________________________

8


This could have significant implications if people are unable to shelter effectively: as discussed earlier in this paper, sheltering indoors is less effective at reducing exposure if ventilation remains constant throughout an incident, and delays in minimising ventilation after a release has occurred or maximising ventilation after an outdoor hazard has passed can lead to significantly higher exposures than if the ventilation strategy had been optimised.

describes four possible systems: background ventilators and intermittent extract fans; passive stack ventilation (PSV); continuous mechanical extract (MEV); and continuous mechanical supply and extract with heat recovery (MVHR). The numbers of new MHVR systems are very low, relative to the existing (mostly naturally ventilated) housing stock, and not all new homes are currently fitted with MVHR – 18,000 MHVR units were sold in 2010/11 (Zero Carbon Hub and Ventilation Indoor Air Quality Task Group, 2012); the annual housing supply in England over the same time period amounted to 117,700 new-build homes (Department for Communities and Local Government, 2011). However, a recent report compiled by a group of industry practitioners and academics considers that the current trend in the UK towards MVHR is likely to continue, with it becoming the dominant form of ventilation in new homes (Zero Carbon Hub and Ventilation Indoor Air Quality Task Group, 2012).

Many commercial buildings vary their ventilation rates according to occupancy level and season: when the outdoor temperature is mild, an “economiser mode” brings in large quantities of outdoor air to keep the building at a comfortable temperature in an energy efficient manner. When the outdoor temperature is too warm or too cold to be used this way, the rate of outdoor air supply is reduced to save energy (Chan et al., 2008). Ventilation rates are expected to be highest in summer months and lowest in winter months. In common with residential properties, to maximise shelter effectiveness during a chemical emergency it is important that building managers are able to rapidly change the ventilation rates of their building when required.

The trend towards mechanical ventilation in newbuilds has a number of implications for sheltering during chemical emergencies. Internal recirculation (where supply air is shut off completely) has the potential to increase protection by minimising air exchange and the ingress of outdoor pollutants due to ventilation. Mechanical ventilation may be fitted with intake filters that lower the concentration of particulate matter (including nitrates and sulphates) in supplied air (NIOSH, 2003) and filtering of external or internally-recirculated air using high efficiency particulate arrestment (HEPA) or carbon filters can greatly improve day-to-day indoor air quality.

The usefulness of mechanical ventilation filters in emergency scenarios is unclear. Properly designed, installed and maintained air-filtration and aircleaning systems could be effective at removing certain types of contaminants (such as particulate matter and aerosols) from a building’s air supply and thereby provide improved shelter effectiveness (NIOSH, 2003). Studies involving portable filters have generally examined protection at high ambient levels, rather than the elevated levels expected during incidents such as fires, and have focussed on biological releases (Muller, 2004; Ward et al., 2005). Fradella (2005) found that the effectiveness of HEPA filtration as a measure to reduce particle concentration depended upon contaminant size, filtration duration, and filter flow rate relative to the infiltration flow rate of the building. Their results indicated that, if given sufficient time, HEPA filters with a high clean air flow rate relative to the building infiltration flow rate may provide protection to building occupants from outdoor particles. Since HEPA filters did not eliminate all ambient particles, it cannot be assumed that filtration is an effective sheltering technique to protect against particulate matter that is a concern at very low concentrations or for brief exposure times. However, HEPA filtration, while not completely removing particulate contaminants, may allow for enough of a reduction in peak exposure to prevent

Filter efficiency is affected by the standard of installation and frequency of maintenance; there is no formal accredited installations scheme for installation and current evidence suggests that maintenance is likely to be poor or non-existent in residential properties (NHBC, 2009b; Zero Carbon Hub and Ventilation Indoor Air Quality Task Group, 2012). This means that many residential filters may not operate effectively. There is evidence that ventilation behaviour and air exchange due to ventilation can be unpredictable. Some people may leave their ventilation systems permanently on boost, whilst others may turn off mechanical ventilation entirely (Zero Carbon Hub and Ventilation Indoor Air Quality Task Group, 2012). There may also be issues around householders’ understanding and use of the controls of ventilation systems (Zero Carbon Hub and Ventilation Indoor Air Quality Task Group, 2012).

________________________________________________________________________________________________________________________

9

J Stewart-Evans ________________________________________________________________________________________________________________________

For chemicals for which the likelihood of adverse health effects is driven by exposure concentration (peak exposure), sheltering in place for short periods can be an effective protective measure, because the dilution of outdoor air mixing with indoor air reduces the maximum concentration indoors. It is important to minimise ventilation before an outdoor hazard arrives. Delays in maximising ventilation after it has passed are less likely to cause extra harm, since most protection has already resulted from the lower peak exposure concentration indoors during the hazard’s passage outdoors.

occupant injury; some studies have found maximum removal rates as high as 90% using three standalone portable filters (Ward et al., 2005). HEPA filters are unlikely to be effective at removing toxic gases from supplied air; hence in many typical chemical release scenarios, including fires, they are unlikely to offer substantial additional protection. Nonetheless, the potential for different types of commonly-used fitted, rather than portable, filters to offer additional protection to sheltering populations is worthy of further attention.

For chemicals for which the likelihood of adverse health effects is driven by both time and concentration (cumulative exposure), which include those that exhibit a linear dose-response relationship, it is particularly important for people to both promptly minimise ventilation before an outdoor hazard arrives and maximise ventilation after it has passed, in order to minimise their cumulative exposure.

Lastly, the location of the supply intake is an important consideration when considering exposure: in a large building the intake may be located in a raised position where the concentration of an outdoor hazard differs from that at ground level. A study modelling the release of chlorine, a dense gas, found that most fatalities indoors were people on the first few floors of buildings: if air intakes are at roof level instead of on each floor, this could greatly reduce the hazard for people sheltering indoors (Barrett and Adams, 2011); this is most relevant when considering large multi-storey buildings or incidents when airborne hazards are close to the ground.

The majority of UK residential properties are naturally ventilated; opening and closing windows can significantly change air change rates and other measures, such as closing controllable vents and openings and taking other steps to reduce infiltration, also affect air exchange. There is an increasing trend towards the use of mechanical ventilation in newbuild properties that are more airtight than their predecessors. The potential for ventilation filters to offer additional protection to sheltering populations when outdoor particulate and chemical concentrations are very high is worthy of further attention. It is also important to consider the future implications of recirculation and filtration in the event of biological and radiological releases, when exposure to particulate matter is of particular significance.

3. Conclusions Ventilation behaviour directly influences indoor exposure during chemical emergencies and it is an important determinant of the effectiveness of sheltering indoors from an outdoor chemical hazard. This paper has used theoretical incident and ventilation scenarios to show how air exchange rates and chemical dose-response characteristics determine indoor dose and effects on health. In practice, it is vital to account for incident-specific factors such as meteorological conditions (particularly the wind pressure and indoor-outdoor temperature differential), outdoor concentrations (which are likely to be variable, rather than constant), the internal configuration of the shelter building and physical attenuation processes (such as filtration, deposition and sorbtion) that lower the rate of increase of indoor concentrations and increase the effectiveness of sheltering, which are not considered in this paper. During an incident, rapid and effective public communication is critical so that people are able to take prompt protective action; the 15 minute ventilation delays presented in this paper have the potential to be very much longer in an incident if communication and implementation of sheltering advice are slow.

Acknowledgements This paper develops concepts that are discussed in an original report published by Public Health England. The author wishes to acknowledge the contributions of the co-authors of that report: N Brooke, J Isaac, A Kibble, L Mitchem, J Russell, L Izon-Cooper, A Nisbet, P Bedwell and J Wellings. References Argonne National Laboratory: (2006) “Shelter-inPlace Protective Action Guide Book, Chemical

________________________________________________________________________________________________________________________

10


https://www.gov.uk/government/uploads/system/upl oads/attachment_data/file/7009/2021176.pdf.

Stockpile Emergency Preparedness Program (CSEPP)”,. Available from: http://orise.orau.gov/csepp/documents/planning/repo rts/argonne-reports/SIP-Guide-Book-May-2006.pdf. Barrett AM and Adams PJ: (2011) “Chlorine truck attack consequences and mitigation”, Risk Analysis, 31, pp1243-1259.

Department for Communities and Local Government: (2012). “English Housing Survey: HOMES 2010”,. Available from: https://www.gov.uk/government/uploads/system/upl oads/attachment_data/file/6748/2173483.pdf.

Biwer BM, Kamboj S, Arnish J, Yu C and Chen SY: (2002) “Technical Basis for Calculating Radiation Doses for the Building Occupancy Scenario Using the Probabilistic RESRAD-BUILD 3.0 Code”. Available from: http://www.orau.org/ptp/PTP%20 Library/library/Subject/Decommissioning/6755.pdf.

Department of Transport Local Government and the Regions: (2002). “The Building Regulations 2000 Approved Document L1: Conservation of fuel and power”. Available from: http://www.planningportal.gov.uk/uploads/br/BR_P DF_ADL1_2002.pdf.

Brelih N: (2012) “Ventilation rates and IAQ in national regulations”, REHVA Journal. Available from: http://www.rehva.eu/fileadmin/hvac-dictio/012012/ventilation-rates-and-iaq-in-nationalregulations_rj1201.pdf.

Dunning AE and Oswalt JL: (2007 ) “Train Wreck and Chlorine Spill in Graniteville, South Carolina: Transportation Effects and Lessons in Small-Town Capacity for No-Notice Evacuation”, Journal of the Transportation Research Board, pp130–135.

Brown SK: (1997) “Indoor Air Quality”, Australia: State of the Environment Technical Paper Series (Atmosphere). Available from: http://www.environment.gov.au/soe/1996/publicatio ns/technical/pubs/12indora.pdf.

Environment Agency: (2005) “Review of Building Parameters for Development of a Soil Vapour Intrusion Model”. Available from: http://publications.environmentagency.gov.uk/PDF/SCHO0605BJEB-E-E.pdf.

Chan WR: (2006) “Assessing the Effectiveness of Shelter-in-Place as an Emergency Response to Large-Scale Outdoor Chemical Releases”. Available from: http://orise.orau.gov/csepp/documents/planning/repo rts/berkeley-reports/60184.pdf.

EPA: (2011) “Exposure Factors Handbook: 2011 Edition”. Available from: http://www.epa.gov/ncea/efh/pdfs/efh-complete.pdf. Fradella J and Siegel JA: (2005) "An evaluation of shelter-in-place strategies in four industrial buildings”. In: Indoor Air, 2005.

Chan WR, Nazaroff WW, Price PN and Gadgil AJ: (2007a). “Effectiveness of urban shelter-in-place-I: Idealized conditions”, Atmospheric Environment, 41, pp4962-4976.

Health and Safety Executive: (2013). “Assessment of the Dangerous Toxic Load (DTL) for Specified Level of Toxicity (SLOT) and Significant Likelihood of Death (SLOD)”. Website: http://www.hse.gov.uk/chemicals/haztox.htm.

Chan WR, Nazaroff WW, Price PN and Gadgil AJ: (2007b). “Effectiveness of urban shelter-in-place-II: Residential districts”, Atmospheric Environment, 41, pp7082-7095.

HM Government: (2010). “The Building Regulations 2010 Approved Document F: Means of Ventilation”. Available from: http://www.planningportal.gov.uk/uploads/br/BR_P DF_AD_F_2010_V2.pdf.

Chan WR, Nazaroff WW, Price PN and Gadgil AJ: (2008), “Effectiveness of Urban Shelter-in-Place. III: Commercial Districts”, Building Simulation, 1, pp114-157.

Howard-Reed C, Wallace LA and Ott WR: (2002). “The effect of opening windows on air change rates in two homes”, Journal of the Air and Waste Management Association, 52, pp147-159.

Chan, WR, Nazaroff WW, Price PN, Sohn MD and Gadgil AJ: (2005) “Analyzing a database of residential air leakage in the United States”, Atmospheric Environment, 39, pp3445-3455.

Johnson T and Long T: (2005). “Determining the frequency of open windows in residences: A pilot study in Durham, North Carolina during varying temperature conditions”, Journal of Exposure Analysis and Environmental Epidemiology, 15, pp329-349.

Department for Communities and Local Government: (2011) “Net supply of housing: 2010-11, England”, Housing Statistical Release. Available from:

________________________________________________________________________________________________________________________

11

J Stewart-Evans ________________________________________________________________________________________________________________________

Litiu A: (2012). “Ventilation system types in some EU countries”, REHVA Journal. Available from: http://www.rehva.eu/fileadmin/hvac-dictio/012012/ventilation-system-types-in-some-eucountries_rj1201.pdf.

Evacuate or shelter-in-place”, Journal of Hazardous Materials, 109, pp1-11.

LMU: (2009). “Low Carbon Housing Learning Zone (web)”. Website: http://www.leedsmet.ac.uk/teaching/vsite/low_carbo n_housing/airtightness/introduction/index.htm.

Stephen RK: (1998). “Airtightness in UK Dwellings: BRE's text results and their significance”, BRE Report 359.

Stephen R: (2000). “Airtightness in UK dwellings”, BRE information paper IP 1/00 2000.

Stewart-Evans J, Brooke N, Isaac J, Kibble A, Mitchem L, Russell J, Izon-Cooper L, Nisbet A, Bedwell P and Wellings J: (2013). “Protective actions in acute chemical and radiological incidents: evacuate or shelter-in-place?”, (in press). Available from: In press.

Muller CO: (2004). “Protecting against chemical agents of war: Applying internal filtration”. In: Sustainable Development: Gearing Up for the Challenge, Indianapolis, IN. pp973-984. NHBC: (2009a). “A practical guide to building airtight dwellings”. Available from: http://www.nhbcfoundation.org/LinkClick.aspx? fileticket=Id3W%2BcSdRak %3D&tabid=339&mid=774&language=en-GB.

Ten Berge WF, Zwart A and Appelman LM: (1986) “Concentration–time mortality response relationship of irritant and systemically acting vapours and gases”, Journal of Hazardous Materials, 13, pp301309.

NHBC: (2009b). “Indoor air quality in highly energy efficient homes - a review”. Available from: http://europeanparliamentgypsumforum.eu/wpcontent/uploads/2012/04/Indoor-air-quality-inhighly-energy-efficient-homes.pdf.

The British Standards Institution: (2005). “BS 5250:2002 Incorporating Amendment No. 1: Code of practice for control of condensation in buildings”. Available from: http://www.commercialconnections.co.uk/Content/b uilding_standards/BS5250%20-%20Code%20of %20Practise%20for%20condensation%20in %20buildings.pdf.

NIOSH: (2003). “Guidance for Filtration and AirCleaning Systems to Protect Building Environments from Airborne Chemical, Biological, or Radiological Attacks”. Available from: http://www.cdc.gov/niosh/docs/2003136/pdfs/2003-136.pdf.

Wallace LA, Emmerich SJ and Howard-Reed C: (2002). “Continuous measurements of air change rates in an occupied house for 1 year: The effect of temperature, wind, fans, and windows”, Journal of Exposure Analysis and Environmental Epidemiology, 12, pp296-306.

Price PN and Sherman, MH: (2006). “Ventilation Behavior and Household Characteristics in New California Houses”. Available from: http://escholarship.org/uc/item/8gx9v5fb.

Ward M, Siegel JA and Corsi RL: (2005). “The effectiveness of stand alone air cleaners for shelterin-place”, Indoor Air, 15, pp127-134.

Sorensen JH: (1988). “Evaluation of warning and protective action implementation times for chemical weapons accidents”. Available from: http://web.ornl.gov/~webworks/cppr/y2001/rpt/3412 .pdf.

Zero Carbon Hub & Ventilation Indoor Air Quality Task Group: (2012). “Mechanical ventilation with heat recovery in new homes”. Available from: http://www.zerocarbonhub.org/resourcefiles/ViaqRe port_web.pdf.

Sorensen JH, Shumpert BL and Vogt BM: (2004). “Planning for protective action decision making:

________________________________________________________________________________________________________________________

12