3.2  Risk, Variability and Uncertainty

Appreciating from the previous section how environmental noise models are constructed and used, the next stage is to discuss the challenges associated with use of the modelling in environmental noise assessment and identify how these challenges may translate into assessment risk.

It is essential that the users have an appreciation of variability in environmental noise and equally important to recognise the distinction between variability, uncertainty and risk.

This section of the guidelines commences with a discussion of environmental noise variability and the challenges it presents to any attempt to objectively rate a noise field. Understanding variability in this way provides a basis for identifying the types of factors that a noise modelling exercise should take into account.

The section then concludes with a discussion of the compounding factors related to the use of predictions that can introduce a further risk of incorrect assessment outcome if not properly understood.

By far the most important limitation to the use of models is the fact that it is necessary in exercising them to make some selection of the environmental parameters. Often, a model will be required to output a single figure and it is crucial to realise that this figure is dependent on assumptions about, for example, weather conditions that will rarely be realised in practice. Since the variation of output values with input condition values is so large, a model may very well therefore give an answer far different to what is experienced in reality.

Environmental noise fields exhibit very large variability in space and time. In the UK, environmental noise assessment of commercial or industrial installations relies heavily on comparisons between the specific noise (the noise attributable to the installation in question) and the background noise (the underlying noise in the absence of influence from the installation in question). Thus, these forms of assessment are burdened by the challenges of dealing with variability in two distinct sound fields.

In terms of both the background and specific noise, changing source and propagation conditions will give rise to changes in both predicted and actual sound fields. The reflection of these changes in predictions or measurements is thus a representation of real trends exhibited in the sound field. Table 3 gives examples of how background and specific noise levels can vary:

Table 3:  Significant causes of variation in environmental noise sound fields

The variability of environmental noise fields presents the most critical challenge to any attempt to rate the field by prediction. It raises the question of how an objective rating of a sound field in one set of conditions may relate to those occurring in other conditions. Further, the variation patterns of a sound field may be a very important consideration to the assessment. The use of well-intended measures to suppress the degree of variability exhibited in objective ratings may in fact undermine the validity of an assessment finding by failing to recognise how actual sound variability may influence the decision making process. For example, two separate sound fields each characterised by identical average noise levels could invoke very different assessment outcomes if one of the fields was characterised by constant levels consistent with the average value, whilst the other field was characterised by widely varying levels with maximum levels substantially above the average at certain times. This highlights the importance of understanding the relative timing of the highs and lows exhibited by the background and specific noise sources, e.g. does the highest background noise only occur when the highest specific noise occurs? This is an important consideration for objective sound field ratings that are often produced for downwind atmospheric conditions that favour sound propagation from the source in question. Such an assessment condition cannot always be assumed to be the condition in which the greatest difference between specific and background noise levels occurs if the controlling background noise sources are also affected by atmospheric propagation conditions.

Whether the specific noise of the installation in question is being compared against a background related limit or fixed value limit, unidentified or misunderstood variability can create significant uncertainty as to whether an objective sound field rating produced for one set of conditions will provide a complete representation of the sound field. Subsequently, if unknown variability or uncertainty overlaps some threshold value at which different assessment outcomes are triggered, there is a significant risk of an incorrect assessment being made.

A related issue is that available prediction methods are limited in the range of conditions which they can model: for example, some engineering methods only enable the calculation of noise levels which occur under downwind conditions.

Given the reliance on knowledge of the background and specific noise for the assessment of commercial or industrial installations, it is important to recognise how each sound field may vary. Importantly, when evaluating appropriate conditions for which an assessment should take account of, it is necessary to identify any relationships that may exist between the variabilities of the background and specific noise fields.

The importance of such relationships can be seen from Figure 3. This demonstrates how the combined effects of variability in both sound fields can lead to critical regions where the ability to discern the causes of variability becomes significant. In this example assessment, where the priority is to determine if the specific noise exceeds the background noise, the critical region occurs when the range of specific and background noise levels overlap. The situation therefore arises as a result of the combination of two factors; the relative magnitudes of the sound levels and the extent to which each may vary. An important consequence is that, outside the critical zone, the specific and/or background noise levels may exhibit even higher degrees of variability but, due to the increased relative difference between background and specific noise levels, this increased variability may have no impact on the assessment outcome. In summary, outside of the critical zone, there is no risk of incorrect assessment outcome despite potentially high levels of variability, even if the sources of this variability are not known. However, within the critical zone the risk of incorrect assessment outcome becomes significant as long as the sources of variability remain unknown.

Figure 3:  Indicative sound level versus distance chart depicting increasing variability with distance from source

It is important to emphasise the importance of understanding the relative timing of the highs and lows exhibited by the background and specific noise sources, e.g. does the highest background noise only occur when the highest specific noise occurs? This is an important consideration for objective sound field ratings that are often produced for downwind atmospheric conditions that favour sound propagation from the source in question. Such an assessment condition cannot always be assumed to be the condition in which the greatest difference between specific and background noise levels occurs if the controlling background noise sources are also affected by atmospheric propagation conditions.

Whether the specific noise of the installation in question is being compared against a background related limit or fixed value limit, unidentified or misunderstood variability can create significant uncertainty as to whether an objective sound field rating produced for one set of conditions will provide a complete representation of the sound field. Subsequently, if unknown variability or uncertainty overlaps some threshold value at which different assessment outcomes are triggered, there is a significant risk of an incorrect assessment being made.