Assessing the influence of street canyon shape on aircraft noise: results from measurements in courtyards near Amsterdam Schiphol Airport.

Martijn Lugten 1 2

1 Department of Architectural Engineering + Technology, Delft University of Technology, The Netherlands; E-Mail: m.c.lugten@tudelft.nl, P .O. Box 5043, 2600 GA Delft, The Netherlands 2 Amsterdam Institute for Advanced Metropolitan Solutions, Amsterdam, The Netherlands

ABSTRACT

Aircraft noise prediction models traditionally omit buildings to optimize speed. Buildings and vertical surfaces scatter and reflect incident sound, affecting sound levels around buildings and within streets. Previous studies showed the impact of buildings on aircraft noise, based on a small number of measurements. Based on additional numerical models, it was found that the shape of buildings, i.e. a slanted or overhanging roof, lead to lower sound levels compared to streets comprised of vertical and flat surfaces. To examine these findings, a full- scale field lab was built near Amsterdam Schiphol Airport, comprised of shipping containers. The experiment consists of three courtyards, in which ten microphones measure sound levels from aircraft flyovers, near facades facing towards and away from the sound source (airplanes). Measurements are matched with meteorological and radar data, which gives information about the position of aircraft and the local weather conditions. The measurements show substantial differences depending on the position, ranging between 8/9dB(A) for a courtyard with straight facades, up to 14dB(A) for facades in a courtyard with a slanted facade and a building inset. Results can be used to rethink the way areas near airports are designed.

1 INTRODUCTION

Aircraft noise causes stress-related complaints and has a negative impact on the well-being of residents living close to airports 1,2 . To protect people from excessive noise exposure, noise contours are a commonly used policy instrument, which restrict and regulate the expansion of urban areas where the noise levels exceed the legal thresholds 1 . Within the EU, contours are calculated on the basis of the European Noise Directive (END), which maintains that noise levels are based on the weighed equivalent sound levels ( L den and L night ) 3–5 . However, sound exposure does not automatically lead to annoyance, but rather is the consequence of reciprocal processes between exposure, context and response 6–9 .

inter.noise 2-24 august scornsuevearcanmus )())D

Figure 1 A) Two source positions with schematic sound paths: being reflected, diffracted around buildings. B) Schematized effect of refraction versus sound propagation without atmospheric refraction.

Literature on the annoyance-reducing effects of quiet building sides show the importance of context for the level of noise annoyance. For road traffic noise, a quiet building side is defined as a facade without a direct line of sight to the noise source ( n LOS from now on), where the L Aeq is <45dB(A), or where the relative difference with the exposed facade (Δ L Aeq ) is >10dB(A) 10–15 .

The shape and surface characteristics of buildings and streets can abate sound levels around or between buildings to meet the criteria for quiet facades (for example see e.g. 16–18 ). Roof shape, (green) cladding, urban density and building dimensions scatter, diffract or absorb sound, as the sound energy decays due to reflections and diffraction over ridges and protrusions.

Since the noise reducing capacity of smart building designs are largely studied in relation to road and rail traffic, it is uncertain to what extent buildings can reduce aircraft noise as well. Theoretically, the position of aircraft means that sound is dispersed from above, limiting the sound abatement by building edges (see Figure 1) (see e.g. 26–28 ). Furthermore, the direction from which the source emits noise, in combination with refraction, changes the angle of incidence of the sound waves as they hit a building 29–31 . This can negate or greatly reduce the noise abating potential of buildings and tall barriers (see Figure 1b) 29,30,32 .

For buildings close to the ground track of a flight path, buildings and streets scarcely attenuate any aircraft noise, which means that the sound level near d LOS and n LOS facades are almost equivallent 33–35 . For buildings at a greater distance from flight paths, buildings seem to act like tall barriers, which leads to a difference between n LOS and d LOS facades. However, the design of the street canyons around a façade may affect the sound levels. Reflections between buildings can amplify the sound level (i.e. L Amax levels) within streets with buildings on both sides 33,35 .

This is different for sites at the flanks of a flight path, where the horizontal distance between the ground track of an aircraft flyover and a building is larger. For example, a computational study comparing twenty urban locations located less than 1000 metres from a flight path (altitude: 100-200ft) found differences up to 4.6dB between the individual locations 27 . In a different computational study, it was found that canyons with slanted facades and building insets yield a greater sound reduction near n LOS facades compared to canyons with straight

shadow zone al shadow zone a2 source position --- propagation path from an overhead source © receiver position = =~ propagation path in case of refraction --- propagation path from an overhead source — propagation path for a source near the ground

facades 36 . The results from both studies suggest that urban and architectural shape may reduce aircraft noise in such areas. The results are not backed by measurements or follow-up studies, and atmospheric effects were not, or only rudimentarily, considered. This raises the question as to what extent the design of street canyons influences aircraft noise based on measurements.

This article presents the first results of a study which examines this question, based on experiments near Schiphol airport in Amsterdam, the Netherlands. The aim of the project is to identify the impact of the shape of buildings on aircraft noise for low-rise residential areas. In the context of this study, low-rise residential areas correspond to buildings which are up to three stories tall.

2 METHOD

2.1 Receiver positions To examine the impact of building geometry on the propagation of aircraft noise, a full scale ‘field lab’ was built close to Amsterdam Schiphol Airport. The lab is formed by three courtyards which are enclosed by walls comprised of shipping containers. Each courtyard has a different geometry, as shown in Figure 2.

Figure 2 Top view and sections of the field lab

The field lab is located near Schiphol’s runways (Kaagbaan), which is one of the airport’s most frequently used runways. Depending on wind direction and engineering works, the

runway is mainly used for departures. Although that flight procedures continue throughout the day, without intermission during the night, most flights depart or arrive between 7am-11pm. The location of the field lab, runway, and ground tracks of flights for a representative day are shown in Figure 3.

Figure 3 Map showing ground tracks of departures from Schiphol's Kaagbaan runway (in blue) on March 1th 2022 and the location of the field lab (in orange).

2.2 Equipment In the field lab, sound levels are measured with ten microphones all near facades facing either towards or away from the nearest flight route. The position of the microphones are shown in Figure 2. The microphones are placed 0.2m away from the facades, each 1.5m above the ground surface, except for microphone 2 and 6 which are each 3.9m above the ground surface. This height corresponds to the position of a window on the first building story. Class II microphones were used (NP2 series), provided by Munisense, equipped with a porous water repellent wind screen. Microphones sit in a thermoplastic waterproof box, which is connected to the electricity grid. The microphones also have a built-in battery, which can provide electricity in case of power cuts. Acoustic data is stored as WAV files on a flash drive on site, and remotely on a cloud server through 4G. Sound pressure levels (SPL) in third octave bands are recorded every 0.125 seconds and uploaded on the cloud server. The acoustic data is matched with a time stamp, linked to a clock at the server.

2.3 Sound sources The microphones record the SPL continuously, which is the accumulated sound from various sound sources around the field lab. The field lab is situated near a road and a motorway, and

surrounded by three farmhouses and warehouses. The road, farms and warehouses mainly produce sound during daytime, while passing cars on the motorway emit a constant hum. To reduce the risk that the sound data for the aircraft flyovers is contaminated, the field lab’s courtyards are fenced off on all sides. This reduces the sound levels which are recorded for the road traffic, which makes it easier to detect sound emitted by aircraft flyovers. The location of the field lab is relatively close to a runway, which means that the SPL is substantially higher during an aircraft flyover than the normal SPL in between aircraft flyovers. Schiphol’s radar system was used to retrieve flight data. Besides information about the flight number, aircraft type and routing, the radar data contains the altitude and geographic coordinates with a resolution of four seconds.

2.4 Meteorological data Weather data was retrieved from a weather mast on Schiphol airport, operated by KNMI (Royal Dutch Meteorological Institute). The distance between the field lab and the weather mast is approximately 5500m. The local surroundings near the weather mast are comparable with the area around the field lab in terms of the ground surfaces (grass fields) and density (limited buildings). This means that the surface roughness will likely be comparable (see also 37 ). The weather data publicly accessible, with the wind and the temperature sensors at a height of 10m and 1.5m above the ground surface. The wind velocity, temperature, humidity, and pressure levels are averaged for each hour.

2.5 Data analysis Aircraft flyovers were detected based on four conditions, which are also shown in Figure 4. Data analysis was carried out through a flow chart written in MATLAB R2018b.

Figure 4 Diagram showing the different steps of the analyses.

Start define dates for which data is analysed Obtain data from micro- phones in SPLs between Spm - 7am Calculate 7, forthe night (9pm “Fam for ‘each microphone Calculate Lj Lag for all 1/3th octave-bands for each microphone Obtain data from KNMI weather mast Obtain data from radar system Schiphol spun) For each microphone, detect peaks in L,..,, Aircraft flyover are selected based on the following conditions: 1. threshold level > L,., + 10dB(A) 2. 58> yy < 1208 3. no rain or strong wind gusts 4 1 is detected within radius of 1500m around the field lab If conditions 1-4 are met, time interval of +/-30s around Z,.. is selected for | analysis For each aircraft flyover /... and ASEL is calculated for each microphone and matched with meteorological and flight information

During a first step, sound events were detected based on acoustic criteria, following ISO 20906. First, the L Aeq for each microphone was calculated, for all data collected between 9pm and 7am. Second, a sound event was defined based on the condition that SPLs had to remain 10dB above the L Aeq for more than 5s and not longer than 120s. This means that passing cars or chirping birds are not accidentally selected as sound events by the algorithm. In some cases, SPLs can exceed the threshold value for a prolonged period, e.g. in case of engineering works or rain. To avoid that such sound events are mistakenly seen as aircraft flyovers, a cut- off value of 120s was chosen. This means that events which duration exceeded 120s were filtered out.

Third, meteorological data from the nearby weather mast at Schiphol airport was linked to the acoustic measurements. The database contains information about precipitation and wind velocity, which was used to omit sound events coinciding with rainfall and very strong wind speeds (>17m/s).

Fourth, information about the local weather conditions and about the flight were matched with the acoustic measurements. If a sound event made it through the the first step, the maximum SPL during the sound event was calculated. Based on the time stamp linked to the L Amax , it is possible to screen the radar data for airplanes in the proximity of the field lab close to the time of the recording. To perform this step, the resolution of the flight and acoustic datasets was matched, by applying a linear extrapolation of the data between points from 4s to 1s. Only flights within a radius of 1500m around the field lab were selected for further analysis, flying at an altitude higher than 150m and lower than 1000m. Also, a square around the field lab was drawn based on geo-coordinates. Only flights flying within this window during the sound event were selected.

If all conditions were met, for each sound event, a window of +/-30s around the L Amax was drawn. Only the data falling inside this window were analysed, and for each microphone the L Amax and ASEL was calculated. In addition, the SEL value for two 1/3-octave bands were calculated (50Hz and 630Hz). Based on the altitude and geographic coordinates, a slant angle was computed based on the time stamp of the L Amax . The slant angle combines the distance between the field lab and the ground track of a flyover with the altitude. This means that the slant angle indicates at which angle sound waves hit the shipping containers captured by a single value. Besides the slant angle, the wind velocity, wind direction, humidity, atmospheric pressure, aircraft type, flight route, and flight procedure (arrival/departure) were combined with the acoustic data.

3 RESULTS

The data which was analysed for this paper was collected between 27 th October 2021 and 2 th November 2021. In total 259 aircraft flyovers were detected and analysed, based on the conditions as set out in the previous section of this article. The wind direction varied between 140 degrees and 210 degrees and 4m/s and 8m/s, depending on the hour and day.

Figure 5 distribution of maximum sound pressure level per microphone (A-weighted).

Distribution of maximum SPL per microphone 105 100 i 'Lamax mic 1 95 i 'LAmax mic 2' 90 ° ° BH 'tAmax mic 3 3 3 3 85 i : BH 'tamax mic 4' $38 é = ° 3 = e B'tamaxmic5' 3 80 ° 2 3 s ° BH 'LAmax mic 6 _ 3 i Wi 'LAmax mic 7' ° ° 70 = ° HH 'LAmax mic 8' Bi 'LAmax mic 9' 65 BH 'tAmax mic 10° 60 55 Microphone

Figure 6 distribution of (A-weighted) sound exposure level (ASEL) per microphone.

Figure 5 and Figure 6 show the distribution of the maximum sound pressure level (SPL) and ASEL for each flight. The whisker boxplots show the mean values, quartiles, and outliers. Figure 5 shows that the range between both quartiles is smaller for microphones 1,2,3

HB 'ASEL mic 1' HB 'ASEL mic 2' BASEL mic 3' HB 'ASEL mic 4' HB 'ASEL mic 5' HB 'ASEL mic 6' HB UASEL mic 7' HB 'ASEL mic 8' HB 'ASEL mic 9' WH 'ASEL mic 10° dB(A) 105 100 ss: 90 85 80 7s 70 65 55 Distribution of ASEL per microphone Microphone ee

compared to 5,6,7 and 9. These microphones are positioned near the facades facing away from the airplanes. The distribution and mean SPL values for microphones 4,8 and 10 are similar. These microphones all face towards the flight path and are therefore fully exposed to sound from airplanes. The results in Figure 6 show that the variance between Q2-Q3 and Q1- Q4 in the boxplots is similar across the microphones near facades facing away from the flight paths. Based on the maximum SPL, buildings seem the yield a shielding effect varying between 7dB(A) for microphone 6 and 7, and 14dB(A) for microphones 1 and 3. Looking at the ASEL, the differences are slightly smaller, but still vary between 7db(A) and 12dB(A).

Figure 7 distribution of sound exposure level (SEL) per microphone for the 50Hz 1/3-octave band.

Distribution of SEL in dB - 50Hz 1/3-octave band 105 8 8 100 3 . ° sash mien + i! HH ASEL mic 1 95 t:. 3 8 3 BE 'ASEL mic 2! 833 3 i 8 90 8 8 BASEL mic 3' BASEL mic 4' a B'aseimics' S$ 80 BB 'ASEL mic 6' 7 ° MASEL mic 7' 70 HB 'ASEL mic 8' BASEL mic 9" 65 WH 'ASEL mic 10' 60 55 Microphone

Figure 8 distribution of sound exposure level (SEL) per microphone for the 630Hz 1/3-octave band.

The shielding effect was further analysed for two 1/3-octave bands, i.e., 50Hz and 630Hz. The distribution of the SEL values for each microphone are shown in Figure 7 and Figure 8. The graphs show that the differences between courtyards and microphones are greatest for higher frequencies compared to lower frequencies. In fact, for 50Hz, the building insets and the slanted wall do not seem to contribute to an additional reduction of the SPL near facades facing away from the airplanes, apart from the shielding provided by the height of the barriers. Although that the shielding provided by the shipping containers is lower for 50Hz than 630Hz, still the differences between d LOS and n LOS facades vary between 6dB and 8dB.

4 CONCLUSIONS

The results show that the sound levels inside the courtyards depend on the shape of the surrounding walls. The following conclusions can be drawn from the data that was analysed for this article:

• The shipping containers shield off the microphones near facades facing away from the airplanes in all three courtyards. Based on ASEL values, the difference between facades facing towards and away from the airplanes range between 7dB(A) and 12dB(A). Based on maximum SPLs, these differences range between 8dB(A) and 14dB(A). • Compared to other microphones, the range between Q1 and Q4 for maximum SPLs in the whisker boxplots is smaller for the microphones near facades facing away from the airplanes in the courtyard with a slanted façade and a building inset. This suggests that

Distribution of SEL in dB - 630Hz 1/3-octave band 105 100 HB 'ASEL mic 1' 95 BASEL mic 2' ° Hy 90 i BASEL mic 3! 8 ° TB 'ASEL mic 4' 85 of ° 8 H'asetmics' S$ 80 o$ HH 'ASEL mic 6' 3° mic _ 8 3 : HE 'ASEL mic 7' . 70 HB ASEL mic 8' BASEL mic 9' 65 WH ASEL mic 10° 60 ° ° 55 Microphone

the maximum sound levels are more stable in this courtyard compared to the other two courtyards. • The level of shielding is greatest in the courtyard with a slanted façade and a building inset, and smallest in the courtyard with straight facades. The difference between both courtyards is approximately 4dB(A), based on the L A,max and ASEL values. • The shielding effects induced by the shape of the buildings seem to disappear for lower frequencies (50Hz). This means that, based on these results, buildings simply act as tall barriers for these frequencies.

In this article, the data hasn’t been correlated to wind velocity and wind direction. These analyses will be carried out in the following months. It is expected that wind effects will influence diffraction and atmospheric refraction, and therefore influence the results. The results show that a carefully designed shape and orientation of buildings can help to reduce aircraft noise near facades facing away from flight paths. The results show that it is most effective to use tilt facades or roofs to reflect incident sound towards the skies. However, also straight walls have a shielding potential, especially for (very) low frequencies.

5 ACKNOWLEGMENTS

The research is funded by the municipality of Haarlemmermeer, the Dutch ministry for Infrastructure and Water Management, the Dutch ministry for housing and domestic affairs, and the Stichting Leefomgeving Schiphol.

6 REFERENCES

1. Netjasov, F. Contemporary measures for noise reduction in airport surroundings. Appl. Acoust. 73 , 1076–1085 (2012). 2. Zaporozhets, O., Tokarev, V. & Attenborough, K. Aircraft Noise: Assessment, Prediction and Control . (Taylor & Francis, 2011). 3. Licitra, G. Noise mapping in the EU: models and procedures . (CRC Publishers, 2012). 4. European Environment Agency. European Noise Directive (2002/49/EC). http://ec.europa.eu/environment/noise/directive_en.htm (2016). 5. Hornikx, M. Ten questions concerning computational urban acoustics. Build. Environ. 106 , 409–421 (2016). 6. Andringa, T. C. & Lanser, J. J. L. How pleasant sounds promote and annoying sounds impede health: A cognitive approach. Int. J. Environ. Res. Public Health 10 , 1439–1461 (2013). 7. Bartels, S., Márki, F. & Müller, U. The influence of acoustical and non-acoustical factors on short-term annoyance due to aircraft noise in the field - The COSMA study. Sci. Total Environ. 538 , 834–843 (2015). 8. Kroesen, M., Molin, E. J. E. & van Wee, B. Testing a theory of aircraft noise annoyance: A structural equation analysis. J. Acoust. Soc. Am. 123 , 4250–4260 (2008). 9. Stallen, P. J. M. A theoretical framework for environmental noise annoyance. Noise Health 1 , 69–80 (1999). 10. de Kluizenaar, Y. et al. Road traffic noise and annoyance: a quantification of the effect of quiet side exposure at dwellings. Int. J. Environ. Res. Public Health 10 , 2258–2270 (2013). 11. de Kluizenaar, Y., Salomons, E. M. & Janssen, S. A. Traffic noise and annoyance : The effect of quiet facades and quiet areas. Euronoise 2012 281–284 (2012). 12. de Kluizenaar, Y. et al. Urban road traffic noise and annoyance: The effect of a quiet façade. J. Acoust. Soc. Am. 130 , 1936 (2011). 13. Öhrström, E., Skånberg, A., Svensson, H. & Gidlöf-Gunnarsson, A. Effects of road traffic noise and the benefit of access to quietness. J. Sound Vib. 295 , 40–59 (2006). 14. Van Renterghem, T. & Botteldooren, D. Focused study on the quiet side effect in dwellings highly exposed to road traffic noise. Int. J. Environ. Res. Public Health 9 , 4292–4310 (2012).

15. Bodin, T., Björk, J., Ardö, J. & Albin, M. Annoyance, sleep and concentration problems due to combined traffic noise and the benefit of quiet Side. Int. J. Environ. Res. Public Health 12 , 1612–1628 (2015). 16. Van Renterghem, T. et al. Road traffic noise reduction by vegetated low noise barriers in urban streets. Proc. - Eur. Conf. Noise Control (2012). 17. Van Renterghem, T., Hornikx, M., Forssen, J. & Botteldooren, D. The potential of building envelope greening to achieve quietness. Build. Environ. 61 , 34–44 (2013). 18. Echevarria Sanchez, G. M., Van Renterghem, T., Thomas, P. & Botteldooren, D. The effect of street canyon design on traffic noise exposure along roads. Build. Environ. 97 , 96–110 (2016). 19. Van Renterghem, T., Salomons, E. & Botteldooren, D. Parameter study of sound propagation between city canyons with a coupled FDTD-PE model. Appl. Acoust. 67 , 487–510 (2006). 20. Van Renterghem, T. & Botteldooren, D. The importance of roof shape for road traffic noise shielding in the urban environment. J. Sound Vib. 329 , 1422–1434 (2010). 21. Hornikx, M. & Forssén, J. Modelling of sound propagation to three-dimensional urban courtyards using the extended fourier PSTD method. Appl. Acoust. 72 , 665–676 (2011). 22. Hornikx, M. & Forssén, J. Noise abatement schemes for shielded canyons. Appl. Acoust. 70 , 267–283 (2009). 23. Lee, P. J., Kim, Y. H., Jeon, J. Y. & Song, K. D. Effects of apartment building façade and balcony design on the reduction of exterior noise. Build. Environ. 42 , 3517–3528 (2007). 24. Hothersall, D. C., Horoshenkov, K. V. & Mercy, S. E. Numerical modelling of the sound field near a tall building with balconies near a road. J. Sound Vib. 198 , 507–515 (1996). 25. Hossam El Dien, H. & Woloszyn, P. Prediction of the sound field into high-rise building facades due to its balcony ceiling form. Appl. Acoust. 65 , 431–440 (2004). 26. Maekawa, Z. Noise reduction by screens. Appl. Acoust. 1 , 157–173 (1968). 27. Hao, Y. & Kang, J. Influence of mesoscale urban morphology on the spatial noise attenuation of flyover aircrafts. Appl. Acoust. 84 , 73–82 (2014). 28. Maekawa, Z. Noise reduction by distance from sources of various shapes. Appl. Acoust. 3 , 225–238 (1970). 29. Arntzen, M. & Simons, D. G. Modeling and synthesis of aircraft flyover noise. Appl. Acoust. 84 , 99–106 (2014). 30. Salomons, E. M. Computational atmospheric acoustics . (Kluwer Academic Publishers, 2001). 31. Attenborough, K., Li, K. M. & Horoshenkov, K. Predicting outdoor sound . (CRC Publishers, 2006). 32. Arntzen, M. Aircraft noise calculation and synthesis in a non-standard atmosphere. (TU Delft, 2014). 33. Donavan, P. R. Model study of the propagation of sound from V/STOL aircraft into urban environments. (MIT, 1973). 34. Flores, R., Gagliardi, P., Asensio, C. & Licitra, G. A Case Study of the influence of urban morphology on aircraft noise. Acoust. Aust. 389–401 (2017). 35. Ismail, M. & Oldham, D. The effect of the urban street canyon on the noise from low flying aircraft. Build. Acoust. 9 , 233–251 (2002). 36. Lugten, M. Schiphol verankert, een stedenbouwkundig archetype voor de luchthavenregio. in De adaptieve luchthaven Schiphol naar de toekomst (eds. van Faassen, J. & de Jong, B.) 168– 176 (nai010, 2016). 37. Trikootam, S. C. & Hornikx, M. The wind effect on sound propagation over urban areas: Experimental approach with an uncontrolled sound source. Build. Environ. (2019) doi:10.1016/j.buildenv.2018.11.037.