A method to quantify the noise annoyance in an airport community Truls Gjestland 1 SINTEF Digital P O Box 4760 Torgarden, N-7465 Trondheim, Norway Idar L N Granøien 2 SINTEF Digital P O Box 4760 Torgarden, N-7465 Trondheim, Norway

ABSTRACT Most methods for quantifying annoyance from aircraft noise only consider residents that are highly annoyed. The much larger population that are annoyed to a lesser degree are usually neglected in such calculations. A method has been proposed that combines the noise level and the number of exposed people in order to give a single-number quantification of the total annoyance experienced by an airport community. The method is particularly useful for assessing the effect of changes in airport operations.

1. INTRODUCTION

Noise from aviation impacts the community around an airport in various ways (World Health Organization, 2018). The most prominent impact, i.e., the impact that is experienced by most residents, is "annoyance", and, if the airport is open for night-time operations, also sleep disturbance. Several methods have been launched to predict the annoyance impact based on measurements or predictions of the physical noise environment (World Health Organization, 2018), (Miedema & Vos, 1998). Such methods, however, can only be used to get a broad overview of the general situation. They fail to give a detailed description, especially of the effect of changes in the airport operations. The main reason for this is that the annoyance response is only partly dependent on the cumulative noise exposure. About two-thirds of the variance in the annoyance response can be explained by other non-acoustical factors (Basner, et al., 2017). The only way of getting a reliable description of the annoyance situation is to conduct a social survey (Fidell, et al., 2011), preferable using standardized methods for easy inter-survey comparisons (ISO, 2021).

The result of an aircraft noise annoyance survey is typically a dose-response curve, valid for that particular airport, showing the prevalence of highly annoyed people among the noise exposed population. This is the preferred way of characterizing the annoyance situation (FICON, 1992), but this dose-response curve gives no information about residents that are annoyed to a lesser degree than highly annoyed, nor does it give any information about the real extent of the annoyance in terms of number of people that are actually affected.

1 Truls.gjestland@sintef.no 2 Idar.granoien@sintef.no

The World Health Organization has published an estimate of the burden of disease of environmental noise (2011). These calculations also rely on highly annoyed residents only, and WHO fails to quantify the negative effects of noise experienced by those that are annoyed to a lesser degree than highly.

This paper discusses an impact model that was originally developed as part of the EU-funded MIME project (Gjestland & Granøien, 2010), (Hullah, 2011). The model quantifies the annoyance combining the number of people affected and their individual degree of annoyance. The model includes all noise-affected residents and not only those that are highly annoyed. The proposed model is particularly useful for monitoring changes in the noise situation around individual airports.

2. ANNOYANCE SCORE

The results from annoyance surveys are usually presented as dose-response functions that show the percentage of the population highly annoyed when exposed to a certain noise level. However, the primary result from an annoyance survey is different. There are standardized questions (ISO, 2021) (Fields, et al., 2001) where a respondent is asked to assess his/hers annoyance on a numerical scale, e.g. 0 – 10, where the end points have been defined as "not at all annoyed" and "extremely annoyed". Verbal scales are also being used, but in most cases the annoyance categories, e.g., "very annoyed” or “moderately annoyed", are subsequently scored unto a numerical scale for the data analysis. The annoyance experienced by a person exposed to a certain noise level can thus be expressed by a single number: the annoyance score, for instance as a value between "0" (zero) and "100" (one hundred).

For the final presentation of the survey data a person whose response is located within the upper 28 percent of the annoyance scale, i.e., annoyance score larger than 72, is considered highly annoyed.

The annoyance score function is sigmoid-shaped with an asymptotic approach to the endpoints. It has been shown (Miedema & Oudshoorn, 2001) that this function is nearly linear in the exposure range about Lden 40 dB to Lden 85 dB. The curve in Figure 1 is based on an analysis done by Miedema & Vos (1998) in a background paper for the European Noise Directive (EU, 2002). For all practical purposes the annoyance score can be described by a linear function in the exposure range most relevant for regulatory purposes.

Annoyance score for aircraft noise

100

80

Annoyance score [%]

60

40

20

0

30 40 50 60 70 80 90 100

Noise exposure level, Lden [dB]

Figure 1. Annoyance score (blue line) as a function of the noise level for aircraft noise and a linear approximation (black line) (Miedema & Vos, 1998)

For transportation noise sources (road, rail, air) the annoyance score function can be described by the following equation (within given exposure limits): AS = k (L den – X) [annoyance score] Eq. 1

The slope of the linear function, k, is similar for all transportation noise sources. A mean value for k for road, rail and aircraft noise is k = 0.0158. For aircraft noise alone the slope has been found to be slightly steeper, k = 0.016.

The linear function indicates that a certain increase in the noise level results in a corresponding increase in the annoyance score regardless of the absolute noise level. This confirms the assumption by Schultz (1978) that the annoyance increases as does the loudness function.

The value of the annoyance score is dependent on the zero-crossing (X dBA) of the linear function. Values for X from different aircraft noise surveys can vary ± 10 dB or more. This can be attributed to the effect of non-acoustic factors (Fidell, et al., 2011). Miedema and Vos (1998) based their dose-response curve for aircraft noise on 20 different surveys. The average value for the zero crossing for these surveys is X = 33.4 dB. However, if the task is to monitor changes in the noise annoyance situation at one particular airport, the absolute value of the annoyance score is not critical. The annoyance score is a quantification of the annoyance experience by a person exposed to noise. Similarly, the sum of the annoyance scores from all the residents around an airport, can be used as a quantification of the total annoyance experienced by that community. This quantity is therefore useful for noise management and control.

When the total annoyance impact on the community can be expressed by a single number, the following noise management measures is possible:

• The airport and the individual airlines may be granted permission to operate only within quantifiable impact limits • Landing rights may be awarded and possibly priced based on individual impacts • Limits for future expansions may be easily quantified • Targets for noise mitigation measures may be quantified

The total annoyance quantity can be used for regulatory purposes. Permission to operate an airport can, for instance, be granted on the condition that the total annoyance impact on the neighborhood community is kept within agreed limits, expressed by the total annoyance quantity. An expansion may be accepted provided that the increase in the total annoyance quantity does not exceed a certain level, and targets for impact mitigations can be expressed in numbers that are easily understood, etc.

In the following discussion the annoyance quantity will be expressed in units of MIMEs (after the initial EU project). The quantity "1 MIME" is equal to one person extremely annoyed by aircraft noise, (annoyance score 1.0) or two persons moderately annoyed (annoyance score 0.5), etc. 3. CALCULATION EXAMPLE

The total annoyance quantity for an airport community can be found by following this simple procedure:

1. Establish a "grid" of cells around the airport, for instance 100 m by 100 m. The grid must include all the impacted residents in the community 2. Measure or predict a noise level that is representative for each grid cell, for instance the noise level in the middle of the cell

3. Find the annoyance score for each cell using Eq 1, k = 0.016, X = 33.4 dB (unless the exact zero-crossing has been established through a previous survey) 4. Find the annoyance quantity, AQ c , per cell by multiplying the annoyance scores with the number of residents per cell 5. Find total annoyance quantity, AQ t , for the community by summation across all cells

The value for the zero crossing, average 33.4 dB, indicates the theoretical limit for annoyance. All persons exposed above this level can be considered annoyed. However, it is not practical to include all of these in the total annoyance quantity. The aircraft noise must exceed a certain level to be heard above the general background noise. It may seem realistic in most cases to include areas within the Ldn 40 dB or 45 dB contour. A report on background noise level in Europe (Gjestland T. , 2008) prepared for EASA gives a justification for this choice.

A simplified version of the procedure is shown in Figure 2. Consider an airport with a single runway running east-west for reference purposes. The impacted area around this airport is defined by a 2.5 x 10 kilometers rectangle. This area has been divided in cells, 250 x 500 meters, and the noise level in the center of each cell has been calculated, see panel 1. (This example is for illustration only. The size of the cells is too large for practical purposes and areas outside the Ldn 50 dB contour should also have been included.)

50 52 54 56 58 60 63 65 65 60 60 65 65 63 60 58 56 54 52 50

52 55 60 62 63 65 68 70 70 65 65 70 70 68 65 63 62 60 55 52

55 60 63 65 67 70 73 75 RWY 75 73 70 67 65 63 60 55

52 55 60 62 63 65 68 70 70 65 65 70 70 68 65 63 62 60 55 52

50 52 54 56 58 60 63 65 65 60 60 65 65 63 60 58 56 54 52 50 Panel 1

500 500 400 400 400 400 500 500 250 250 250 250 500 500 400 400 400 200 200 300

500 500 500 300 400 400 300 200 250 250 250 250 400 600 500 500 250 250 200 300

600 400 500 500 400 400 400 100 RWY 200 300 300 200 250 250 250 250

500 500 600 600 600 400 500 300 250 250 250 250 500 500 500 700 600 200 300 200

600 400 700 500 600 600 400 400 250 250 250 250 300 700 600 400 400 400 300 200 Panel 2

133 149 132 145 157 170 237 253 126 106 106 126 253 237 170 157 145 66 60 80

149 173 213 137 189 202 166 117 146 126 126 146 234 332 253 237 114 106 69 89

207 170 237 253 215 234 253 67 RWY 133 190 176 108 126 118 106 86

149 173 255 275 284 202 277 176 146 126 126 146 293 277 253 332 275 85 104 60

159 119 231 181 236 255 189 202 126 106 106 126 152 332 255 157 145 132 89 53 Panel 3

Figure 2. Calculation of total annoyance quantity. Panel 1 shows the noise level per grid cell. Panel 2 shows the number of residents per cell, and panel 3 shows the annoyance quantity, MIMES, per cell, AQ c (annoyance score x number of residents).

Panel 2 shows the number of residents per cell. The total population of the impacted area is 36,900 people. This information can usually be found from public census data. Panel 3 shows the annoyance quantity per cell, AQ c , calculated from eq. 1 and the population per cell. Summation across all 96 cells gives the total annoyance quantity for this airport: AQ t =16,376 MIMEs. In other words,

on average the impacted population around this airport is a little less than moderately annoyed (AS = 0.44). 3. PRACTICAL USE OF THE TOTAL ANNOYANCE MODEL

The model is particularly useful for assessing the effect of various changes in airport operations as illustrated in the examples below. .

3.1. Redirecting traffic

A visual inspection of the data in Figure 2 shows that the noise situation seems quite symmetrical with an equal amount of traffic in each direction, see panel 1. The population on the other hand, is higher on the west side than on the east side of the runway. The population on the west side of the symmetry line amounts to 20,200, and on the east side 16,700. The concept of the total annoyance quantity, AQ t , can be used to see how a re-directing of the traffic will affect the annoyance situation.

Increasing the noise levels in less populated areas while reducing it in densely populated areas will reduce AQ t . Let us assume that the traffic can be redirected so that all noise levels on the west side is reduced by 3 dB and at the same time the levels on the east side will increase by 3 dB. This will yield the following change in AQ t : Reduced Δ AQ t, west = 0.016 * 3 * 20,200 = 970 MIMEs Increased Δ AQ t, east = 0.016 * 3 * 16,700 = 802 MIMEs Net reduction Δ AQ t = 970 – 802 = 168 MIMEs

In the original situation we had AQ t = 16,382 MIMEs. The net reduction in the total annoyance quantity amounts to only 1.0 percent. In this particular case redirecting traffic has very little effect with regard to the noise situation and is probably not worth considering.

Similarly, a visual inspection of the population density reveals that there are less people living north of the runway than on the south side. What would be the change if the traffic could be redirected so that the noise levels in the cells in direct extension of the runway (east-west) are kept constant, but all the levels north of this line are increased by 3 dB and the levels on the south side are reduced by the same amount? A calculation like the above example shows that the net decrease will be only 115 MIMES or about 0.7 percent.

In this example redirecting traffic will have little impact on the total annoyance. However, in case there are large unpopulated or sparsely populated areas near the airport, a redirecting of the flight paths may reduce the total annoyance considerably.

3.2. Replacing the aircraft fleet

The noise emission from different aircraft within the same category (similar passenger capacity, range, etc.) may vary. So, what would be the effect of replacing the entire aircraft fleet with more quiet planes, that on average will reduce the noise level in each cell by 2.5 dB? The reduction can be calculated as follows: Reduction Δ AQ t = 0.016 * 2.5 * 36,900 = 1476 MIMEs

This reduction amounts to an almost ten percent change from the original situation with 15,408 MIMEs, which is likely to have a significant impact.

3.3. Increasing airport operations

The effect of an increase in the traffic volume can be readily quantified. If one assumes a proportional increase (same type of traffic but more of everything) the noise levels in the neighbourhood will increase accordingly. If the number of operations increases by 25 percent, the equivalent level will be 10*log(1.25 )= 0.97 dB higher, corresponding to an increase in the total annoyance: Increase: Δ AQ t = 0.016 * 0.97 * 37,200 = 577 MIMEs

This quantity represents an increase in the total annoyance of about 3.5 percent. Similarly, a 50 percent increase or a 100 percent increase (a doubling of the traffic) will have an effect equivalent to a 6.4 percent or 10.9 percent increase in the total annoyance, respectively.

The change in the total annoyance due to an increase in the operations can be accurately predicted and quantified and can be used in discussions between the airport authorities and the surrounding community prior to a possible expansion of the airport operations. 3.4. Specially protected areas

With the described method an area with low population density will contribute little to the total annoyance. In the initial traffic planning process such areas may therefore be favoured for high noise exposure. However, there may be reasons to protect certain areas from excessive noise. Examples are wilderness areas, recreational parks, locations for health care facilities, etc. Such areas can be assigned a large virtual population so that noise at these locations would have a corresponding large impact on the annoyance budget. 4 . CONCLUSIONS

The proposed method for assessing the annoyance from aircraft noise quantifies the noise impact on the whole exposed community and not only on those that are highly annoyed. People that are annoyed to a lesser degree are also taken into account. The method combines the noise level and the number of exposed residents. The method provides an easy-to-understand quantification of the effect of proposed changes in the operation pattern for a specific airport and can be used as an illustrative tool in discussions and negotiations between an airport and the surrounding community.

6. REFERENCES

Basner, M., Clark, C., Hansell, A., Hileman, J. I., Janssen, S., Shepherd, K., & Sparrow, V. (2017). Aviation noise impacts: State of the Science. Noise & Health, 19: 41-50.

EU. (2002). Environmental Noise Directive, 2002/49/EC. Retrieved from http://ec.europa.eu/environment/noise/directive_en.htm

FICON. (1992). Federeal Agency review of selected airport noise analysis issues. Washington, DC: Report for the Department of Defense.

Fidell, S., Mestre, V., Schomer, P., Berry, B., Gjestland, T., Vallet, M., & Reid, T. (2011). A first principles model for estimating the prevalence of annoyance with aircraft noise exposure. J Acoust Soc Am, 130(2), 791-806.

Fields, Jong, d., Brown, Flindell, Gjestland, Job, . . . Yano. (1997). Guidelines for reporting core information from community noise reaction surveys. J Sound and Vibration, 685-695.

Fields, Jong, d., Gjestland, Flindell, Job, Kurra, . . . Felscher-Suhr. (2001). Standardized noise-reaction questions for community noise surveys: research and a recommendation. J Sound Vib, 641-679.

Gjestland, T. (2008). Background noise levels in Europe. Retrieved from EASA document library: https://www.easa.europa.eu/document-library/research-reports/90e10277

Gjestland, T., & Granøien, I. (2010). MIME - market-based impact mitigation for the environment. Baltic-Nordic Acoustic Meeting. Bergen.

Hullah, P. e. (2011). Final Report of the MIME project. Brussels: European Commission.

ISO. ( 2021). ISO/TS 15666 Acoustics - Assessment of noise annoyance by means of social and socio- acoustic surveys. Geneva Switzerland: International Standards Organization.

ISO. (2003). TS 15666: Acoustics - Assessment of noise annoyance by means of social and socio-acoustic surveys. International Standardization Organization.

Miedema, H. M., & Oudshoorn, C. G. (2001). Annoyance from transportation noise: Relationships with exposure metrics DNL and DENL and their confidence intervals. Environmantal Health, 109, 409-416.

Miedema, H. M., & Vos, H. (1998). Exposure-response relationships for transportation noise. J Acoust Soc Am, 3432-3445.

Schultz, T. J. (1978). Synthesis of social surveys on noise annoyance. J Acoust Soc Am, 377-405.

World Health Organization. (2011). Burden of disease from environmental noise. Copenhagen, Denmark: WHO Regional Office for Europe.

World Health Organization. (2018). Environmental noise guidelines for the European region. Copenhagen, Denmark: WHO Regional Office for Europe.