Odours typically arise from complex mixtures of gases, which stimulate the olfactory, or smelling sense. Human response to odour is not easily predicted, however, because there are many factors which can affect the determination of whether an odour is detected and/or recognised, and whether it is deemed to be pleasant or unpleasant.
Bias, cultural influences, exposure history, awareness and individual expectations can combine to confound the manner in which individuals react to an odour. The context in which we first encounter an odour and our reaction to it can become associated with that odour and subsequent exposures can trigger strong emotional responses.
In fact, recent research has shown that newborn babies, detect, discriminate, and even have preferences for certain odours and tastes based on earlier foetal exposure. In addition, some of the ‘association learning’ related to pleasant and unpleasant odour experiences can influence the babies’ behaviour throughout their lifetime (Schaal, et al, 2004).
Parallels between odour and noise
Interestingly, there are many similarities in the way that humans react to odours and noise. Scales and acceptance criteria are often influenced by the context in which the stimulus is experienced and there can be a profound interaction between the psychological response and our physical response to odours and noise. Thus, some individuals demonstrate significant reactions at odour concentrations which others may not even detect. Similar to noise, our reaction to odour is non-linear and the human nose perceives the stimulus of smell in a logarithmic scale of intensity. While humans have evolved an incredibly sensitive response to odour, our reliance on this sense in today’s world is much less than our forebears’, who’s very survival was affected to a large extent by their ability to smell.
What odour communicates
According to Reed (1990), one of the reasons for the sensitivity of the human nose is the close coupling of molecular odorant recognition events to neural signalling, and this allows the human nose to detect a few parts per trillion of some odorants. While an ability to detect the early signs of rancidity in food as well as other potential dangers, such as fire and potential exposure to chemicals, is still undoubtedly helpful, our response to odours can be completely out of kilter with the degree of risk. Some odours may cause strong physical reactions like nausea and vomiting without being present at toxicologically significant concentrations.
A case in point is the odour associated with protein hydrolysis and the putrefaction of animal tissue. Two extremely odorous compounds, putrecine (1,4-diaminobutane) and cadaverine (1,5-pentanediamine), are involved, the odour of which will make most people vomit; however, their toxicological significance is limited (acute oral toxicity of more than 2,000 mg/kg body weight in rats). According to a US Environmental Protection Agency (US EPA) study (1992), the lack of correlation between risk and odour is simply because the mechanisms that appear to be involved with odour detection have very little to do with the mechanisms involved in chemical induced toxicity and carcinogenesis.
Despite this, however, detection of odour in combination with information regarding chemical identity and toxic potency can be useful information, especially in those cases where the odour threshold concentration is known and can be compared to health-based ambient criteria (US EPA,1992).
Odour thresholds and measurement
There are essentially two types of odour thresholds – the detection threshold and the recognition threshold. Detection can be defined as the concentration at which the average panel member notices an odour – over and above an odourless background – but cannot necessarily identify it. The recognition threshold is the lowest concentration at which the average panellist can identify a definite character of the odour.
According to the US EPA, “The difference in concentration between detection and recognition thresholds can vary from approximately twofold to tenfold and significant inconsistencies have arisen with comparable data.” Some of the earlier threshold compilations such as Verschueren (1977), and Fazzalari (1978) contain threshold values from data published in the early 1900s and in some cases, reported threshold values vary by a factor of a million or more for one compound (US EPA, 1992). In the same way that subjective assessments of noise levels have been substantially enhanced by technological developments and the widespread availability of acoustic measurement equipment, there is a trend to move away from using the judgement of an individual (environmental health officer or sanitary officer) to determine whether an odour is likely to cause nuisance. In addition, controls now tend to focus on predicting the likelihood of odour problems or complaints more proactively. Since the early 1970s indices to rate the potential of a material to cause odours have been developed.
The Odour Index as initially proposed by Hellman and Small in 1973 was used to predict whether complaints were likely to arise from the use of chemicals under certain conditions, e.g. emissions associated with spillages, leaks and evaporation. Major developments have arisen since then, however, including the discovery of the ‘gene family’ encoding human olfactory receptors in the early 1990s (Buck and Axel, 1991).
Nowadays odour monitoring and assessment is widely adopted as a critical tool in the management and regulation of ‘odorous’ industries and processes. As with many environmental issues, policy and regulatory developments are often influenced by trends and technological developments, and our tolerance to odours has varied over the centuries – as has the appetite and framework for control.
The profound odour emissions associated with poor sanitation and the problems with (or complete absence of) sewerage networks – and reliance on horses for transport – meant that the urban environment of the early, Twentieth Century was profoundly unpleasant to the nose. Municipal authorities were slow to focus on the odour impacts of poor sanitation and sewerage, however, and our early public health legislation focused on ‘Offensive Trades’, many of which no longer exist in modern urban environments. This is undoubtedly related to the perception of risk and political and socioeconomic factors, but for many centuries there was an assumption that ‘foul smelling air’ was the cause of great distress and illness.
In most instances, odour assessment predominantly addresses the annoyance potential of odorants in the ambient air as opposed to the health aspects; however, a full appreciation of the World Health Organization (WHO) definition of health – “a state of complete physical, mental and social well-being…” – would clearly require an absence of objectionable odours.
Typically, the methods for analysing airborne odours involve two main approaches – olfactometry and chemical/physical measurements of chemical components which give rise to odours. In some industries there may be hundreds of individual compounds involved and olfactometry and odour modelling have evolved as the tools of choice. An olfactometer is a device in which a sample of odorous gas is diluted with clean or neutral gas in a defined way and presented to an odour panel under reproducible conditions. Thus, olfactometry is the measurement of the response of assessors to the ‘odorous gas’.
The odour concentration at the detection threshold is defined to be 1 oum-3. The more odorous the sample, the greater will be the number of dilutions required. For example, if an odour sample has been diluted by a factor of 1,000 to reach the detection threshold, then the concentration of the gas sample is deemed to be 1,000 oum-3.
Standards for guidance
While certain member states had done much of the pioneering work and even had their own criteria, EU odour standards have been under development since the early 1970s. In 1990, a European Committee on Standardisation (CEN) produced a draft standard and by 1999 inter-laboratory tests had been carried out and the European odour unit, the European Odour Reference Mass (EROM) and the performance criteria for measurement had become well defined.
While it took some years before its widespread adoption throughout the EU, the trend is to now use a standardised European odour unit, as follows: European odour unit (ouE m-3) is defined as “that amount of odorant that, when evaporated into one cubic metre of neutral gas at standard conditions, elicits a physiological response from a panel (detection threshold) equivalent to that elicited by one European Reference Odour Mass (EROM), evaporated in one cubic metre of neutral gas at standard conditions.” (The subscript E is used to denote the linkage with the European unit.) European Reference Odour Mass (EROM): this is “the accepted reference value for the European Odour Unit, equal to a defined mass of a certified reference material. One EROM is equivalent to 123 µg n-butanol (CAS No. 71-36-3). Evaporated in one cubic metre of neutral gas this produces a concentration of 0.040 µmol/mol.”
The 2003 European Standard (EN 13725) defines a method for the objective determination of the odour concentration of a gaseous sample using dynamic olfactometry with human assessors and the emission rate of odours emanating from point sources and area sources. The primary application is to provide a common basis for evaluation of odorant emissions in the member states of the European Union. While guideline values designed to limit odour annoyance have been published by the WHO, these are only for a small number of individual compounds, as opposed to groups of compounds or mixtures which are normally associated with industrial and waste/waste water emissions. Nonetheless, as part of the planning, permitting and environmental management procedures for plant and industries with significant odour potential, extensive guidance and guidelines have evolved and a substantial body of these utilise the European odour unit and the EN 13725 methodology.
Thus, odour exposure criteria have evolved as a statistical means of associating odour emissions from a source with the impact (airborne concentration) at ground level, in terms of probability of occurrence. The criteria are probability-based and therefore are not absolute ‘limits’. Strictly speaking, the criteria are based upon ‘average concentrations’ that are likely to occur for a specified percentage of the time over a year.
Factors of influence
Ambient odour problems are most likely to arise in adverse meteorological conditions, e.g. temperature inversions or low wind speeds, when relatively little dilution arises. For this reason, most computer dispersion modelling will incorporate a minimum of three years’ met data sourced from a local station. In essence, odour impact assessment utilises odour source concentrations, ventilation rates and emission strength (odour emission rates), topographical information together with meteorological data and air dispersion software to model odour dispersion circa the source.
Approach to assessment
Odour impact areas can be defined by plotting isopleths of odour concentration corresponding to selected values for odour impact criteria, and/or ground level concentrations (GLCs) can be predicted at specific locations, e.g. site boundaries or neighbouring properties. According to Yang and Hobson (2002), a quantitative approach is necessary for odour control because a rational and consistent means is needed for the justification, evaluation and specification of solutions for preventing and reducing odour nuisance. Thus, predicted GLCs can be assessed with reference to odour criteria which are typically expressed in terms of percentile values.
A 98th percentile value ‘x’ of a year of hourly averaged concentrations means that hourly averaged concentrations will be less than or equal to ‘x’ for 98% of the year. For 2% of the year, hourly averaged concentrations will be higher than or equal to ‘x’. In other words, the 98th percentile is the predicted concentration for the 176th worst hour of the year. Similarly, the 99.9th percentile is the predicted concentration for the eighth worst hour of the year; the 100th percentile is the predicted concentration for the worst hour.
Guidance for industrial applications
In Ireland, odour concentrations arising from industrial sources should generally be below a maximum of 6 ouE m-3 for the 98th percentile of the time in one standard meteorological year (e.g. 2%, or 175 hours per annum) in order to prevent odour complaints arising. By contrast, in the UK assessment criteria have developed to serve nuisance assessment and offensive odour assessment. The assessment of nuisance draws on a 5 ouE m-3 98th percentile criterion, which was first used at a planning inquiry for the Newbiggins-on-Sea waste water treatment works.
Since then the criterion has become a de-facto nuisance level, and the UK has gone on to develop risk assessment procedures with specific industries and sectors being targeted. For the Integrated Pollution Prevention and Control (IPPC) sector, extensive guidance has been published in the UK, and the 2011 Environment Agency’s Guidance provides a quantitative approach, based on assessment criteria which range from the 98th percentile value of 1.5 to 6 ouE m-3 (98th percentile). These criteria are based upon one hour averaging times and it is recognised that emission limit values imposed in any particular case will depend ‘upon the installation-specific circumstances and what is achievable through the application of Best Available Technology (BAT)’.
The latest version of the guidance expresses the criteria in terms of benchmarks based on the 98th percentile of hourly average concentrations of odour modelled over a year at the site/installation boundary. The benchmarks are: • 1.5 odour units for most offensive odours • 3 odour units for moderately offensive odours • 6 odour units for less offensive odours
An absolute threshold or not?
These UK criteria are clearly ‘indicative values’ which are designed to avoid ‘justifiable complaints’ and, similar to noise criteria, they have been derived from dose effect studies. Intensive livestock industry development in the Netherlands, in close proximity to the relatively dense human population, led to some of the pioneering work being undertaken in that jurisdiction.
This experience in the Netherlands has been used to inform policy and guidance throughout other member states, and the Irish EPA adopted criterion for a new continuous source in populated areas is 1 ouE/m3 as the 99.5th percentile of one hour averages, with a target value of 1.5 ouE m-3 in 98% of all hours. Strictly speaking, this criterion applies to intensive agriculture, but the target value provides a general level of protection against odour annoyance for the general public. The Irish odour exposure criteria aim to define ‘acceptable odour exposure’ that should not be exceeded at ‘sensitive receptors’, e.g. dwellings or schools. According to the latest UK guidance on waste water treatment plants (DEFRA, 2006), it is not possible to define an absolute threshold level of complaints that will be indicative of statutory nuisance.
Any such determination will be dependent upon the evidence gathered and an assessment of relevant facts and complaints, which are designed to evaluate whether an odour is prejudicial to health, or a nuisance, taking into account the ‘FIDOL factors’, and those other criteria used for assessing statutory nuisance. According to DEFRA, the FIDOL factors are frequency, intensity (and therefore concentration), duration, relative offensiveness (hedonic tone/character) and the location, along with any aggravating characteristics. Typically, factors used in assessing statutory nuisance in the UK include the sensitivity of sufferers. Interestingly, DEFRA acknowledges that for a particular odour, approximately 2% of the population are likely to be hypersensitive and 2% anosmic; that is, unable to detect any odour.
Thus, there is a reluctance to impose mandatory numerical standards set in the form of odour concentrations in ambient air, although some definitive guideline values as well as case law have arisen.
Far from being an absolute and unconditional methodological approach which is used in isolation, odour measurement and modelling is used in the planning, design and management of facilities and in the interpretation of Best Practicable Means and in the application of Best Available Technology. While standardisation across Europe has done a lot to advance the science, technical advances have recently allowed the development of electronic, or ‘e-nose’, continuous odour monitoring techniques.
Data from these devices are sent to proprietary software, which models the atmospheric dispersion and quantifies real-time odour plume dispersion to facilitate automatic odour alerts. Thus, immediate responses and controls can be applied in anticipation of the human olfactory response. While odour measurement technology and methodology has made major advances in recent years, care must be taken to minimise the uncertainties and to strike a reasonable balance in the interpretation of the data.
In the same way that noise will always be a feature of industrial endeavour, so too will odour emissions, and as a result of variations in individual sensitivity some people will unfortunately continue to be affected more than others.
Published: 01st Sep 2012 in AWE International