Aircraft Noise

Human Response to Noise

There is very wide range of response to noise. While one person sleeps soundly at night with aircraft 300 feet overhead producing noise at ground level which is 8 times as loud as that which is scientifically rated as likely to awaken, another can tell you the time at which the same aircraft passed over his house at 3,000 feet.

Measurement of sound levels

Sound is measured in decibels (dB) which follow a logarithmic scale because our perception fits such a representation – a sound which we rate as twice as loud as another is measured 10 decibels higher. In aircraft terms, a ‘plane which produces 80 dB on flyover will be perceived as twice as noisy as one which produces only 70 dB. A decibel is one tenth of a bel, named for Alexander Graham Bell, and is used in its place because it represents the smallest change we can detect under laboratory conditions; a difference of 3dB between aircraft is generally regarded as the minimum we will be aware of.

Sound is usually measured on an “A-weighted” basis, dB(A), which means that the sound pressure waveform picked up by a microphone is put through a filter which weights it to mimic the variations in sensitivity of the human ear to the different frequencies within it. However, this ignores low frequency vibration which is annoying and can generate audible noise by, for example, rattling windows and doors.

The figure below shows the sound pressure waveform (blue) and meter readings (red) for an aircraft taking off at about 1 mile from the end of a runway.

Characterising aircraft noise exposure around airports

Noise around airports is most often presented in terms of the equivalent average noise energy, L_eq, in dB(A) units for aircraft movements within specified hours of the day and over a specified number of days. The resulting noise indicator is, therefore, averaging the intense noise events which occur as ‘planes fly over and the longer periods of quiet between the events.

The most common such noise indicators are for the 16 hours of between 07:00 and 23:00 (daytime) and 23:00 to 07:00 (night), both for an average operational day based on the 92-day summer period 16 June to 15 September. The official night-time noise index in the UK is the 8-hour L_eq between 2300 and 0700. Contours are plotted at 3 dB(A) intervals from 57 to 72 (daytime) and 48 to 72 (night).

However, the European Commission 2002 community noise directives now require new presentations for the purposes of noise mapping and airport noise controls (although there are currently no plans to change the official index for other purposes such as planning): L_den (day, evening, night) is a full 24-hour, 365 day indicator in which, to reflect the public’s greater sensitivity to noise at these times, flights in the evening period (19:00 to 23:00) are weighted by having 5 dB(A) added to the noise they are assumed to generate and those at night (23:00 to 07:00) 10 dB(A). L^night _{is the same as the 8-hour Leq except that it is produced for the full year, not 92 summer days.}

Computer Models

It is important (but difficult) to avoid the word “measured” in this context as L_eq is seldom actually measured. It is usually calculated using computer programs which take as input the flights, aircraft types and journey lengths (from which fuel loads and hence take-off weights are calculated). Where predictions are being made, it is assumed that the direction of runway operation follows historic averages.

When modelling the noise at airports, it is usually assumed that aircraft follow the local preferred tracks, often known as Noise Preferential Routes (NPRs). Where routes involve turns, there is significant variation in the actual tracks flown and efforts must be made to model the resulting dispersion about the centre line if the resulting noise contours are to be accurate.

The computer programs which generate the noise estimates use the actual noise measurements for take-off, landing and sideline which were made for each aircraft type for the purposes of noise certification when it was introduced to service. However, the certificated noise measurements are obtained under test conditions: studies of actual noise measurements at airports have shown that these sometimes underestimate the noise experienced.

There are two computer programs commonly in use in the UK: ANCON2, which is maintained and run by the Civil Aviation Authority (CAA), and INM from the US Federal Aviation Authority (FAA). INM is publicly available and is therefore used widely by airport operators, consultants, public authorities, etc. It is understood that the CAA validates ANCON2 through a programme of measurement conducted at Heathrow which in part deals with the weaknesses mentioned above.

Given the simplification inherent in the process of “simulating” the generation of noise, the process is bound to be subject to significant error. However, where contours are deemed to represent the noise impact of tens of thousands of flights, it is likely that the cumulative error is small unless there is bias in the input parameters. Computer-generated contours are usually claimed to be accurate within 1 dB(A).

However, where predictions of future noise impacts are being made, it is important to scrutinise the assumptions which are being made to generate them. The map below shows how a noise contour, in this case a 66 dB(A) daytime L_eq contour, can vary with the underlying assumptions. All assume a similar number of aircraft movements but the types of traffic vary significantly. We have no comment on the fact that the contour published by the airport operator happens to be the smallest.

Alternative noise indicators

The use of L_eq, the equivalent average noise exposure over a significant time period is not an intuitively obvious indicator of disturbance from noise which actually occurs as short bursts of intense noise separated by long periods of quiet. It is more suited to noise from motorways where, for long periods of the day, the noise level can be heard as a wavering level of noise.

Its proponents argue that, because of the logarithmic dB scale, it weights the noisiest events heavily so that it puts pressure on operators to eliminate the noisiest aircraft from the fleet. This is true: campaigners against Heathrow calculated that one flight of Concorde generated noise equivalent to 35 flights by Boeing 747-400s. But the effect of this has been that airports claim, on the basis of average noise indicators, that their operations are getting quieter at the same time as the number of flights is increasing hugely. A simple example shows this to be nonsense: if all aircraft were to be replaced by models which were 3 dB(A) quieter, most people on the ground would be hard pressed to notice the change. Yet L_eq noise indicators would not return to the same levels until the number of flights was doubled, a change which would be very difficult for those living below to ignore.

Most people perceive noise as increasing with increasing frequency of flights whether the aircraft are quieter or not so average noise does not represent what they experience. In addition, average noise is not easy to measure. This has led to interest in alternative ways of characterising noise around airports.

One such indicator is N₇₀, a simple count of the number of overflights at a location for which the peak noise on the ground exceeds 70 dB(A). It is usually calculated for the “average day” in a time period. N₇₀ has been used extensively in Australia to supplement average indicators. N₆₀ is sometimes used for the night period.

Before L_eq was adopted in the UK, the standard indicator was the Noise and Number Index (NNI) which attempted to reflect both the frequency of noise events and their loudness. It may, therefore, be better suited to indicate disturbance from aircraft noise than either L_eq or N₇₀. A recently completed study for the UK Government known as ANASE has provided some support for this view. It concluded that either flight numbers are under-represented by L_eq indicators or that, if L_eq is a valid indicator of annoyance, people are now more sensitive to aircraft noise than they were in the early 1980s when a similar study was completed.

Individual events

In addition to the indicators which attempt to aggregate the impacts of multiple flights, there are others which address the effects of individual flights. These are most often quoted in relation to sleep disturbance where one noise event is enough.

L_Amax is the most easily understood of these as it simply indicates the peak A-weighted noise generated on the ground as the noise rises and falls with the passing of the aircraft. A more complex indicator, and one which is not intuitively useful but which is favoured by the industry because it can be related back to L_eq, is the Sound Equivalent Level (SEL), sometimes known as the Single Event Level. This is the equivalent noise level in dB(A) which would have been reached if all of the sound energy generated by the event had been concentrated in a single pulse of one second’s duration.

When used as a predictors, these indicators are usually calculated for the “typical” aircraft expected to operate at night as it is difficult to predict the absolute noisiest ever likely to visit. As an example, the SERAS study produced noise “footprints” based on an SEL of 90 dBA (“SEL90”) for the noisiest aircraft in the Quota Count 2 (QC/2) category.

The Quota Count (QC) is a noise scale operating at the three airports at which night noise is regulated by the Secretary of State for Transport. Each aircraft’s Quota Count is based on its noise classification and, for departures, destination and the airports, Heathrow, Gatwick and Stansted, each have a nightly noise quota as well as a cap on total movements.

Noise Limits

There are no accepted limits or targets for noise exposure beyond the boundary fences of airports. The Government’s Planning Policy Guidance (PPG24) specifies constraints for building where aircraft noise is likely to be a problem but these do not apply in reverse: “You can build your airport next to my house but I cannot build my house next to your airport”. PPG24 expresses various levels of constraint on development based on L_eq day and night indicators and states that development should not be permitted where noise events regularly exceed 82 dB(A) L_Amax at night.

The World Health Organisation (WHO) recommends that daytime L_eq should not exceed 55 dB(A) to avoid significant annoyance. It also recommends that outside maximum noise (L_Amax) from aircraft over-flights should not exceed 60 dB(A) if sleep disturbance is to be avoided in rooms with open windows.  British Standard BS8233, which specifies noise insulation standards for buildings, specifies similar noise limits and, in particular, an upper target of 55 dB(A) L_eq for outside areas.

Luton & District Association for the Control of Aircraft Noise - site updated 22 August, 2008