INTERNATIONAL GREEN ROOF INSTITUTE

Do Extensive Green Roofs Reduce Noise?

Jens Lagström University of Malmö Eco-cycle Programme

Examination Project Environmental Science Spring 2004

www.greenroof.se

1

Publication No 010

2

Cover Photo: Louise Lundberg

INTERNATIONAL GREEN ROOF INSTITUTE

Do Extensive Green Roofs Reduce Noise? Jens Lagström University of Malmö Eco-cycle Programme

Financed by: MKB Fastighets AB City of Malmö in cooperation with EU-life Malmö University Examination Project Environmental Science Spring 2004

Augustenborgs Botaniska Takträdgård Ystadvägen 56, SE-214 45 Malmö, SWEDEN +46-40-94 85 20 e-mail: [email protected] www.greenroof.se

Publikation nr 010 3

ISBN 91-973489-9-6

4

Contents Foreword ....................................................................................................................................................... 7 Abstract ....................................................................................................................................................... 7

1 Introduction ....................................................................................................................................... 8 1.1 The Aim of the Investigation .................................................................................................... 8 1.2 Structure of the Study ................................................................................................................ 8 1.3 Clarification ..................................................................................................................................... 8

2 Theoretical Background ............................................................................................................ 9 2.1 What is Sound? .............................................................................................................................. 9 2.2 What is Noise? .............................................................................................................................. 12

3 The effect of sound on people ........................................................................................... 13 3.1 The Physiology of the Ear ........................................................................................................ 13 3.2 How Does Noise Affect People? ............................................................................................ 14

4 The Link Between Green Roofs and Noise ................................................................. 16 4.1 4.2 4.3 4.4 4.5

Why Reduce Noise with Green Roofs? ................................................................................ 16 What are Green Roofs? ............................................................................................................. 18 The History of Green Roofs ..................................................................................................... 20 Advantages of Green Roof Vegetation .............................................................................. 20 Problems with Green Roof Vegetation ................................................................................ 22

5 Methodology ................................................................................................................................... 23 5.1 Methodology for Noise Assessment ................................................................................... 23

6 Material ................................................................................................................................................26 6.1 Structural Description of the Cabin ..................................................................................... 26 6.2 Laying Vegetation on the Cabin Roof .................................................................................. 26 6.3 Positioning of the speaker on the cabin roof ................................................................... 27

7 The Results of the Noise Level Assessment .............................................................. 29 8 Discussion ......................................................................................................................................... 32 9 References ......................................................................................................................................... 34

5

6

Foreword This study has been carried out in part to contribute with empirical data to current ongoing research, and in part to satisfy my own interest for green roof vegetation. There are companies in the green roof business who claim that green roof vegetation really has a noise reducing effect. I have therefore chosen to seize the opportunity to contribute with what I believe to be new knowledge in the area. With appropriate equipment it is not difficult to investigate the noise reduction capacity of green roofs, but writing about noise is more complicated when the focus is environmental rather than physics or acoustics. Those who have assisted me with their knowledge and support in my ups and downs are in alphabetical order: Per Hillbur, Peter Lindhqvist, Louise Lundberg and Djamel Ouis and others close to me. Many thanks.

Abstract The human being controls her environment by the help of her sense of hearing, which can not be turned off in any circumstances. Since it is so important from a health perspective to have a decent sleep, noise disturbance can cause severe consequences for the state of health. Approximately 20 % of the population of Europe is exposed to levels of noise that are unacceptable which often leads to irritation, disturbance of sleep and a risk of negative effects on health. Noise involves an annual cost within the EC on fully 10 Billion €. The cost of noise for the Swedish society is estimated to 21 Million € each year. The question if extensive green roofs reduce noise or not is answered by measurements of noise levels on a reference building without a green roof, which is compared with the identical measurements of noise levels on the same building, but with a green roof. The method that has been used for the measurements of noise levels is the Schroeder method. The results of the impulse response from the comparative measurements of noise levels confirm that extensive green roofs do reduce noise, which is an additional positive effect supporting the development of green roofs.

7

1

Introduction

There are, in my opinion, many reasons to construct green roofs. That green roofs do not inhibit, but rather promote biodiversity is not difficult to understand, as newly laid roofing felt is virtually dead, whilst a green roof is living. There are of course better ways to improve biodiversity than creating a green roof, but remember that land in cities is expensive and whichever alternative is available, it should also function to minimise the load on sustainable urban drainage systems. In addition, the chosen solution should be aesthetically pleasing and simultaneously decrease dramatic temperature changes in the roof construction and thereby decrease energy consumption. Maybe it could even reduce noise levels in homes, offices and other buildings.

1.1 The Aim of the Investigation The aim is to assess if extensive moss-sedum green roofs have a noise reducing effect. This is carried out by two comparable noise assessments, one carried out with roof vegetation and the other without.

1.2 Structure of the Study The theoretical elements are presented first in order to create a greater understanding of the experiment. In this section the questions “What is sound?” and “What is noise?” are addressed. The reader is then informed of how sound affects people, focussing on the physiology of the ear, to be followed by the physical and psychological effects of noise on the human body. When the theoretical part is completed then the environmental aspects follow in order to link physics and acoustics to green roofs from an economic perspective. The nature of extensive green roof vegetation is explained in addition to the history of green roofs and the various advantages and disadvantages of creating a green roof. The methodology for the experiment is then presented, the equipment in the shape of a labourer’s cabin is described, along with the laying of the green roof vegetation and later installation of the speaker on the roof of the shed. Finally, almost, there is some discussion, but first the exciting results of the experiment are presented.

1.3 Clarification In this publication the terms green roofs, green roof vegetation and extensive green roof vegetation are used with the same meaning. However, green roofs can figuratively include a wide range of different roofs, both with and without vegetation. Green roof vegetation is specifically roof with vegetation, but the maintenance needs of this are not defined. The different definitions will be presented in more detail in section 4.2 “What are green roofs?”. These three expressions are consciously used synonymously in the text to create a more varied and readable style. In the document, vegetation with associated growing substrate (soil) is referred to as ”moss-sedum mats” which refers to prefabricated mats with growing substrate and vegetation consisting of different moss and sedum varieties. Sedum is the Latin name of a genus in the crassulaceae family 1. 1 Stenberg, 1992, p. 183

8

2

Theoretical Background

2.1 What is Sound? Sound can be described as variations in density and pressure that expand in an elastic medium which could be a gas, liquid or solid body. The precondition for variations in pressure to be defined as sound is furthermore that in addition to the source of the sound and the transmitting medium, there should also be a receiver of the sound. What we humans experience as sound are variations of pressure in the air causing the eardrum to vibrate 2. In order to assess sounds, our hearing uses the parameters of level, frequency and timing. When the variations in pressure reach the ear they create a certain sound level at the ear. This sound level is measured in decibels [dB]. The frequency is the number of cycles per second [Hz] and is experienced as the tonal pitch of the sound 3. The sound of a pure tone can be visualised as a sine curve and its strength can be seen in the relative distance between the top and bottom of the waves. If this distance is halved then the amplitude of the sound can be assessed. The amplitude is measured in decibel [dB] in a logarithmical scale which means that a halving of perceived sound corresponds to a decrease of 10dB. In the same way, an increase of 10dB in intensity is perceived as a doubling 4.

Figure 1, Source: Johansson, 2001, s. 14

We also receive a large amount of qualitative information about the sound by its duration. For example, strong impulse sounds with very short duration. The three factors – level, frequency and duration are used to identify the character of the source of the sound. Other factors that affect the sound are spatial, absorption and reflection which together constitute the acoustics of the space 5. The physical qualities of sound can be categorised as infrasound, audible sound and ultrasound. Infrasound is sound with a frequency of less than 20 Hz and ultrasound is

2 Andersson, 1998, p. 11 3 Magnusson et. al, 1989, pp. 39-40

4 Berg et. al, 1995, p. 153 5 Magnusson et. al, 1989, pp. 39-40

9

sound with a frequency of over 20 000 Hz. These two frequencies; 20 and 20 000 Hz, constitute the boundaries of audible sound for human beings 6. Our hearing, however is most attuned to frequencies between 500 and 8 000 Hz 7. Bats, by comparison, can hear sounds with a frequency up to 100 000 Hz 8. The speed of sound is approximately 340 metres/second given a temperature of 20°C and normal atmospheric pressure. The speed of sound is represented by [c] in mathematical calculations and formulae. The frequency of sound is measured in Hertz [Hz] and is the number of variations in pressure per second, which is represented by the figure [f ] in calculations. The wavelength of sound is measured in metres, and is represented by [λ] in calculations. The relationship between the speed of sound, frequency and wavelength can by shown as [ c = f . λ ] 9. The diagram to the right shows examples of sound generated by different activities in order to understand different sound levels. When measuring sound with a sound level metre it is usual to use a weighing filter which gives a frequency-dependant moderation of the microphone signal so that the human hearing can be replicated 10.

Figure 2, Source: Åkerlöf, 1999, p. 10

The simplest filter, and probably the most used, is the A-filter with which sound levels can be expressed as dB [A]. The A-filter adds a weighting related to the sensitivity of then ear at that particular frequency. An A-filter strengthens the frequencies between 20 and 1 000 Hz. The frequencies between 1 000 and 7 500 Hz are mildly moderated, whilst frequencies between 7 500 and 20 000 Hz are amplified as demonstrated in Figure 3. This filter is constructed to replicate the human sound threshold and is used for example for assessing the risk of damage to hearing 11. There is also a C-filter, expressed as dB [C], which is usually used to measure and assess impulse sound 12.

6 Ingemansson, 1977, pp. 13-14 7 Magnusson et. al, 1989, p. 39 8 Arbetarskyddsstyrelsen, 1990, p. 11 9 Johansson, 2001, pp. 13-15

10

10 Johansson, 2001, p. 26 11 Kinsler et. al, 1980, p. 280 12 Johansson, 2001, p. 26

Figure 3. Source: Kinsler et. al, 1980, p. 280

Frequencies are measured in a logarithmical scale which is split into octave bands, see Figure 4. These octave bands are named according to their geometric average frequencies: 31,5, 63, 125, 250, 500, 1 000, 2 000, 4 000, 8 000 and 16 000 Hz and range thus from 22 to 22 000 Hz. The octave band’s geometric average frequencies are determined by an international standard by ISO (International Organisations for Standardisation) 13.

Figure 4, Source: Andersson, 1998, p. 33

13 Andersson, 1998, p. 32-33

11

This figure provides a greater understanding as some of the results from the measurements are presented as octave bands. Note at the foot of the figure the buildings acoustic measurement range in accordance with the Swedish buildings standard whose measurement range is between 20 and 3 500 Hz. The lower threshold is that of human audibility and the upper limit is in accordance with the Swedish standard.

2.2 What is Noise? Sound normally consists of pure tones at different frequencies. In industrial environments sound includes all frequencies with a random range of intensity, called fuzz. What we perceive as speech is a mixture of pure tones and fuzz 14. The difference between fuzz and noise is the perceived sound level and the two concepts, sound and noise are usually characterised through the simple definition that noise is undesirable sound for the listener. The characterisation of noise is therefore dependant on a subjective evaluation of whether the sound is perceived as pleasant or unpleasant 15. There is a range of different kinds of sound which are perceived as noise. Certain types, such as electric motors, principally emit a constant noise where the combination of frequency and intensity does not vary. The usual kind of noise is fluctuating, where the combination of frequency and intensity vary, which is usual in the manufacturing industry. There is also impulse noise such as for example hitting noises and the clatter of machines 16. The upper figure shows pure tones from a tuning fork and the lower figure shows fluctuating noise from a mechanical lathe. The large difference in the appearance of the figures is because noise contains all the frequencies with a differing intensity range.

Figure 5, Source: Ingemansson, 1977, p. 13 14 Arbetarskyddsstyrelsen, 1990, pp. 10-12 15 Andersson, 1998, p. 11 16 Arbetarskyddsstyrelsen, 1990, pp. 10-12

12

3

The Effect of Sound on People

3.1 The Physiology of the Ear We can perceive and differentiate between different kinds of sound with the help of our hearing. Our sense of hearing works even when we are sleeping, which can lead to us being awoken by unexpected sounds and signals. With the help of our ears we have the ability to ascertain the direction to the source of a sound. The ear is a highly sensitive organ with functions to receive and transform sound impressions and to analyse the frequency of the impression 17. The sensitivity of the ear also has a fantastic range both regarding frequency and intensity. For example it is possible for us to hear frequencies, as mentioned earlier from 20 Hz to 20 000 Hz, a factor of 1 000. As a comparison the sensitivity of the eye to electromagnetic waves is around 400 to 700 nanometres, a factor of less than 2 18. From an anatomical perspective it is possible to separate the ear into three parts – the outer ear with the ear canal, the middle ear with the transmission mechanisms and the inner ear. Variations in pressure in the air reach the outer ear, are amplified in the auditory meatus and cause the eardrum to vibrate. These vibrations in the eardrum are transferred via the anvil in the middle ear to the oval window, which transmits the sound waves further to the inner ear and the inner ear fluid which in turn affects the hairs in the cochlea roughly like reeds in the waves. When the hair cells are stimulated they affect the auditory nerve in the form of electrical impulses which are transferred via the nerve cells to the brain 19. The basiliary membrane in the cochlea is 3.5 cm long and contains around 30 000 nerve cells, which are often called hair cells because of their physical appearance 20.

Figure 6, Source: Johansson, 2001, p. 51

17 Andersson, 1998, p. 41 18 Berg et. al, 1995, p. 144

19 Johansson, 2001, p. 49 20 Berg et. al, 1995, p. 147

13

3.2 How Does Noise Affect People? People are affected by noise in a number of different ways from insignificant noise which barely affects us to directly damaging noise that is both physically and mentally disturbing. Noise can from a physiological perspective include everything from undamaging to painful and damaging 21. The most usual ways that noise affects people is damage to hearing, sleep disturbance, decreased capacity and tiredness but noise can also affect blood circulation and cause stress symptoms 22. The sense of hearing cannot be disconnected either in a waking condition or in sleep, unlike the sense of sight which can be controlled by closing ones eyes. This means that people monitor their environment even in sleep. As good sleep is important for human health, then noise disturbance can lead to serious consequences for health. Individual sensitivity varies significantly and it is also possible to get used to noise 23. There are studies that indicate that sleep quality is reduced as a result of exposure to individual sounds of 45 dB [A] at the most for noise sensitive people, but it should be noted that this group of noise sensitive people includes one third of the population. Results of epidemiological studies in areas around airports with high noise levels (67 – 75 dB [A]) show that heart problems, visits to the doctor and purchase of medicines are more common than in quiet areas (46 – 55 dB [A]). Psychosocial well being in terms of reported depression was significantly lower amongst people living in a noisy environment 24. Performance at work is also affected negatively by noise, depending on how complex the task and how many sources of information need monitoring. This is particularly the case for noise of a very strong and varied character, as stress can lead to lowered performance through increased frequency of error. Surroundings with weaker and monotonous noise, particularly at a low frequency, can cause tiredness. This can lead to difficulties dealing with information and lowered speed of reaction 25. Our hearing capacity naturally lessens as we grow older, but at a significantly lower rate for those who are not exposed to noise in eg. the manufacturing industry. Hearing loss is often incremental and it can take many years before the damage is significant. Early hearing tests are important to identify incremental hearing loss and to act appropriately. Studies have shown major individual differences in the perception and susceptibility to noise 26. The ISO definition of hearing impairment is a level of hearing loss where the ability to follow everyday speech in a quiet environment starts to decline 27. Hearing loss is essentially an increase in the individual’s hearing threshold which means that it is not possible to hear weak sounds to the same extent as before exposure to noise. There are two different kinds of hearing impairment, temporary and permanent. The degree of temporary hearing loss and the time needed for the hearing threshold to regain its original value is primarily the result of the volume of the noise and the exposure time. Permanent hearing impairment is due to the damage of a number of hair cells in the inner ear and the larger their number, the greater the damage. This damage can unfortunately not be repaired and the damaged hair cells will never re-grow. There is also research that indicates that in addition to damage to the sensory cells that also nerve cells in the central ear canal can be damaged by noise 28. 21 22 23 24

14

Kinsler et. al, 1980, p. 279 Ingemansson, 1977, p. 6 Arbetarskyddsstyrelsen, 1990, p. 37 Naturvårdsverket, 1994, pp. 14-16

25 26 27 28

Johansson, 2001, p. 56 Carlsson, 1988, pp. 17-20 Möller et. al, 1978, p. 33 Möller et. al, 1978, pp. 9-13

There are 800 000 people in Sweden with hearing problems. 3% of the population in Sweden use hearing aids and each year 30 000 hearing aids are prescribed. 1.5 Billion Swedish Kronor have been paid out since 1974 as compensation payments for workrelated noise damage. This noise damage results in suffering and costs for the individual, but also large costs for society 29. Communication between people is also affected by noise as it becomes harder to hear what is said if the noise level is high which can increase the risk for shouts of warning not to be heard 30. This is referred to as Speech Interference Level and can be partially compensated for by decreasing the distance, increasing voice strength or using an electronic amplifier 31.

Figure 7, Source; Ingemansson, 1977

The primary reasons to decrease noise is to avoid direct damage, but also to increase comfort whilst sleeping, working and socialising. Good insulation is important, particularly in multi-family housing and open landscape offices, not purely to protect individual integrity, but also to avoid intrusion into the private sphere of others. The degree of isolation depends on sound insulation in the roof and between walls, but also on ambient noise levels. Insulation can be achieved in many ways; by building better walls or roof to decrease their ability to transmit noise, by increasing the amount of absorbing material in the room, which is the source of the noise and finally to increase the sound levels in the receiving room to counteract the incoming sound 32. In the context of widely varied individual sensitivity to noise, the case of the ventilation in an open plan office with such god sound insulation that the ambient sound level was unsatisfactory. It was simply too quiet in this office where great effort had been taken to improve the working environment. The solution was to increase the background noise level and the chosen option was to play back the sound of a fan from the internal speaker system. The result of this improvement, however, was that the staff, who were previously quite happy, started to complain of draughts from the speakers. Some members of staff even complained of stiff necks as a result of the draught 33. 29 Andersson, 1998, p. 59 30 Ingemansson, 1977, p. 6 31 Kinsler et. al, 1980, p. 283

32 Kinsler et. al, 1980, p. 283 33 Andersson, 1998, s. 186

15

4

The link Between Green Roofs and Noise

4.1 Why Reduce Noise with Green Roofs From a societal perspective it can be stated that a decrease in noise pollution will lead to a decrease in social problems. This, however, will just be a marginal gain as this decrease will be small in comparison to the large proportion of the population who will be more rested and alert as they will not have their sleep disturbed if noise levels are reduced 34. Noise can also cause economic problems through decreased efficiency amongst employees as well as lowered property values due to less demand 35. There are 170 million people living in the EU in areas where noise is a source of irritation in the daytime 36. In addition to these 170 million people, there is an estimated 80 million people, or 20% of the European population, exposed to noise levels that are unacceptable, noise levels which lead to irritation, sleep disturbance and risk for negative health impacts 37. A conservative estimate is that noise leads to an annual cost of approximately 10 Billion € each year in the EU. In Sweden the total cost of noise is calculated to be 21 Million € each year 38. A possible decrease in noise levels is just one of the possible benefits of laying green roofs. The advantages of green roofs are many and include aesthetic, technical and environmental benefits, which will be presented in more detail in section 4.4 “The advantages of green roof vegetation”. In order to calculate the potential noise reducing effect of extensive green roof vegetation, a speaker was installed faced down on the roof of the cabin that was used during monitoring of noise levels. This was done as only the noise that has passed through the roof is the subject of investigation. This will be described in more detail in chapter 6.3 “The positioning of the speaker on the cabin roof ”. The speaker’s position and direction can be compared to the noise impact of air traffic. Noise from air traffic has received more attention as there is a large number of people who are seriously affected by noise from aircraft and the costs to minimise effects on people of this kind of noise are very high. The significant attention is due to the complexity of the measurement system for air traffic noise monitoring as every different category and model of aeroplane must be assessed individually at both take-off and landing 39. According to previous studies a standard roof reduces noise by 33 dB. The estimated noise reduction of a standard roof with green roof vegetation, where the thickness of the vegetation and drainage layers are not specified, is 41dB when the roof is dry and 51dB when wet. This can be compared to the noise reducing capacity of a 100 millimetre concrete wall of 43dB. These estimations mean that the noise reducing capacity of green roof vegetation is the equivalent of 8dB or more, when compared to a conventional roof. It is assumed that it is the growing medium layer and drainage layer that are most efficient in noise reduction. The effective noise reducing barriers that are used today are often made of solid materials such as concrete, as opposed to thin layers of vegetation that are not efficient noise barriers 40.

34 35 36 37

16

Karlsson Hjort, 2000, p. 16 Kinsler et. al, 1980, p. 279 Karlsson Hjort, 2000, p. 16 Grant et. al, 2003, p. 24

38 Åkerlöf, 1999. p. 33-35 39 Kinsler, et. al, 1980, p. 295 40 Grant et. al, 2003, p. 24

This is reinforced by a study that has investigated the noise reduction effect green barriers. The results show that a shelterbelt with a width of 10 metres only decreases noise by 3 - 6dB. A reduction of 20dB is first achieved at a width of 50 metres. This study was undertaken at eight mines which were surrounded by greenery. The vegetation had different characteristics of density, ground cover and height which had an impact on the individual results from the study sites 41. Other studies show that indoor noise from air traffic can be reduced to 35dB in upstairs rooms in two storey houses by installing an extra window 25 centimetres inside the ordinary window or by installing a noise weakening ventilation unit in every un-insulated room. The total cost for each house would be £200, if carried out in a large number of properties. In order to further decrease noise levels it was in some cases necessary to improve noise insulation in the roof and also to seal the chimneys 42. It should be noted that this study was carried out in 1967 and that the costs are not adapted to the current economic situation and therefore are irrelevant. According to studies in central Dalby, the maximum levels for indoor noise from air traffic, 45 dB [A] 43, was exceeded several times each night. The measurements were made as a result of complaints from residents that they were suffering sleep disturbance as a result of air traffic from Sturup airport 44. The Swedish Environmental Code classes airports as environmentally hazardous activity. The Environmental Code includes the term “detrimental to human health”, which is activity which from a medicinal perspective can affect health 45. All environmentally hazardous activity should be carried out in a location that is appropriate so that protective measures, limitations and other precautions can be adopted to prevent or counteract damage or problems for human health. Human health problems are those that are not very minor or completely temporary with regard to both physical and mental health. The Environmental Code also mentions that issues such as noise are classed as a problem and that regard should be taken to more sensitive individuals 46. As soon as there is reason to assume that noise from an activity can cause damage or inconvenience for human health or the environment, then it is possible for action against that activity supported by the Environmental Code 47.

41 42 43 44

Pal et. al, 2000, p. 163 Scholes et. al, 1967, p. 37 Åkerlöf, 1999, p. 17 Glimberg, 2003-04-30

45 Åkerlöf, 1999, p. 27 46 Rubensson, 1999, p. 25-26 47 Åkerlöf, 1999, p. 27

17

4.2 What are Green Roofs? The definition of green roofs, or roof vegetation, is a roof with some kind of vegetation established either spontaneously or intentionally. This can then be subdivided into extensive and intensive roof vegetation, depending on the need or lack of it for water and nutrients. Intensive roof vegetation needs watering and fertilising and also needs attention to prevent the establishment of unwanted species, whilst extensive green roofs do not need artificial watering or fertilising 48. Turf roofs are therefore classed as extensive green roofs as they are naturally broken down and disappear, which is not the case for extensive green roofs in a longer time frame 49. Figure 8 shows an example of extensive green roofs in the housing area in Malmö’s Western Harbour.

Figure 8, Source: Referensobjekt, Veg Tech, 2003

Green roofs consist of several layers as seen in figure 9, the roof construction and waterproofing layer (roofing felt), root penetration protection, drainage layer, filtration layer, growing substrate, and sedum and moss plants 50. The growing substrate and drainage layers contain air that can be assumed to reduce noise, but the air also decreases the density of the layers which lowers its noise reducing capacity.

Figure 9, Source; Piga, 1995, s. 21

48 Söderblom, 1992, p. 5-6 49 Piga, 1995, p. 8 50 Piga, 1995, p. 21

18

In Sweden it is standard practice not to use root penetration protection when establishing extensive green roofs and therefore no such layer has been used in the noise level study. No roofing felt was used either as the cabin used in the study had a roof construction with a fibreglass waterproof seal. The drainage layer used in the noise level study was Grodan©, a mineral wool mat (figure 10) whose function is to absorb precipitation and drain off any excess 51. These are therefore effective in retaining water to even out run-off rates, and also have the advantage of being lighter than gravel. Mineral wool mats are a good alternative as drainage layers for green roofs on an incline as their friction prevents the sedum mats from sliding. This is not the case for flat roofs or roofs with a gentle incline as other materials such as gravel can be used 52.

Figure 10, Grodan©, Source: Veg Tech, 2003

Extensive green roofs can be established on all normal roof inclines, but it is beneficial of the roof has a slop of at least 2 degrees in order to prevent water collecting on the roof as retained water can provide the opportunity for undesired plants to root which necessitates a regular check for roofs in shady areas. At inclines above 25 degrees then the vegetation mats must be secured to the underlay by gluing for example 53. The maximum incline for a roof to function successfully is 30 degrees 54.

51 52 53 54

Veg Tech, 2000, pp. 7-10 Söderblom, 1992, p. 16-17 Söderblom, 1992, p. 21 Velazquez, 2002-10-14

19

4.3 The History of Green Roofs The first green roofs were probably at the hanging gardens of Babylon which were constructed between 700 and 500 B.C. Roof gardens have been found in the ruins of the Roman Herculaneum which was buried in an eruption of Vesuvius in 79 A.D. There are also examples of roof gardens from the early middle ages, such as the Mont Saint Michel in Normandy which was built in the 13th Century 55. Each level of the monastery has still today green layers in the form of lawns, herb gardens, vegetable gardens and meadows 56. The inspiration for today’s green roofs comes originally from rugged Iceland where grass and turf have been used for hundreds of years. Icelandic architecture is adapted to its lack of natural resources and uses what nature can provide 57. The turf roofs provided protection from damp and cold and protect the roof and walls of both homes and barns. As such there were only functional motives for the turf roofs 58. A layer of birch bark was laid under the turf as a waterproof layer. Birch bark contains a specific kind of oil that makes the rotting procedure very slow. The bark rolls up when it dries, therefore the layer of turf was used to keep the bark in place. This specific oil is very flammable and the problem was the risk of fire from the easily ignited grass in the spring that could lead to disaster if sparks from the chimney lit the turf. The solution to this was to lift a goat onto the roof in the spring to graze off the grass 59. Around thirty years ago the green roof techniques were re-adopted in Germany since which green roofs have become increasingly popular around Europe, largely due to their positive effect on the environment 60. In recent years, development has been towards thinner and lighter roofs, which has meant that large roofs with low load bearing can be covered with green roof vegetation. German industrial buildings dominate by area, due to legislation that proposes green roofs on new industrial buildings in most German cities 61.

4.4 Advantages of Green Roof Vegetation As cities expand and build on green space then there should be, or even must be, a general aim to maintain and establish as many as green surfaces as possible 62. Laying green roofs is part of this work. Their ability to delay run-off is also a functional instrument to decrease pressure on local storm water systems, as both the vegetation and the drainage layers absorb rain water until they are saturated, thereby delaying run-off and decreasing the volume which otherwise would have gone straight down the drain. Many large roofs, which are covered in roofing felt or gravel, are perceived as dull and sterile deserts and can be transformed to objects of life and beauty by roof vegetation whose aesthetic benefits are significant. The vegetation provides a display which never ceases to fascinate as the colours change through the course of the year. As the city becomes greener it becomes more attractive, as seen in many research results which show the importance of greenery for people, for relaxation, lowered blood pressure and lower pulse. Green roofs are softer and have a significantly larger surface area with their leaves and stalks than hard roofs, which means that green roofs have a greater capacity to bind dust and particles from the air blowing across the city. Green roofs are beneficial for 55 56 57 58

20

Grant, et. al, 2003, pp. 12-13 Personal observation, 1999-07-10 Velazquez, 2002 Piga, 1995, p. 9

59 60 61 62

Veg Tech, Takvegetationens historia, 2003 Velazquez, 2002 Piga, 1995, pp. 9-10 Piga, 1995, pp. 11-13

biodiversity as they increase the biological area compared to traditional hard roofs 63. As the green roofs are softer than hard roofs the vegetation protects the roof by absorbing blows from heavy hail or from things being dropped on the roof 64. Roof vegetation has a water regulating effect which means that when it rains, some of the water is bound in the earth layer from which it can be taken up by the plants or evaporate. The water not retained on the roofs is delayed which means that the run-off is evened out in comparison to hard roofs, decreasing pressure on the storm water system. 60% of annual precipitation can be retained, and thereby never reach the storm water system, by 2 cm thick sedum mats on a gravel drainage layer 65. A comparison of the temperature of the waterproof layer on a roof covered with roofing felt and one with roof vegetation shows significant differences. There are temperature variations of 100°C on the roof with just roofing felt as the surface temperature can vary from 80°C on hot summers days and -20°C on cold winter nights 66. Such temperature variations in a roof with roofing felt surface results in tensions and movement in the roof, which can eventually cause serious damage to both the seal and the roof construction and as such have a negative effect on the life expectancy of the roof 67. This should be compared with a roof with roof vegetation where the temperature on cold winter nights does not fall below -5°C and does not exceed 25°C on hot summer days. The temperature variation on a roof with a roof vegetation is therefore just 30°C, dramatically decreasing the stresses caused by temperature variations and extending the life expectancy of the roof, and in some cases doubling life expectancy. The green roof’s insulating capacity decreases temperature variation by 70°C. This means that a house with roof vegetation needs less energy to heat the house in the winter months than it would with a conventional roof which leads to economic savings for the house owner or tenant. The same is true of warm summer’s days when less energy is needed for cooling 68. Hard surfaced urban areas absorb solar heat during the day and radiate heat in the evening and at night. This means that the air in large cities warms faster the following day, increasing the average temperature in the city by about 6°C. This in turn doubles levels of low-level ozone which is formed in a photochemical reaction in which heat is a catalyst. This phenomenon of raised temperatures in large cities is called the Heat Island Effect and is clear in infrared heat-sensitive photographs taken from a high level. If these large areas of hard surfacing are instead covered with green roof vegetation, the situation will be reversed 69. The green roofs function to even out temperatures by soaking up water which later evaporates, cooling the surrounding air on hot days and warming the air on cold nights as the water evaporates 70.

63 64 65 66

Skärbäck, 2002-09-13 Söderblom, 2002-09-06 Bengtsson, 2002, pp. 245 – 250 Giesel, 2002-10-14

67 68 69 70

Piga, 1995, p. 15 Giesel, 2002-10-14 Velazquez, 2002-10-14 Söderblom, 2002-09-06

21

4.5 Problems with Green Roof Vegetation The winds dynamic upward force must be taken account of when establishing an extensive green roof, of particular importance when planning to use thin lightweight layering systems. All corners and edges are particularly susceptible to the winds effect, but all roof surfaces are more or less at exposed depending on their shape, size and height above the ground. Should the wind catch hold of thin vegetation mats, however, they will not be blown away, but will simply fall again by their own weight. This problem can easily be avoided by laying a strip of gravel along the outer edges of the vegetation mats and putting stones on top of the mats 71. Increasing the weight of the vegetation layer with stones decreases the risk of gusts of wind lifting the roof vegetation. The stones can also become an aesthetic adornment by varying the appearance of the roof and creating a more living impression as seen to the right. There are a number of different methods for creating moss and sedum roof vegetation. These are direct seeding, spreading shoots, planting and pre-fabricated vegetation mats. Direct seeding is the most cost efficient, but demands constant watering, but not so much as to risk washing them from the roof. It should be added that mosses cannot by reproduced by seed. Spreading shots can cover large areas quickly and rationally. Early spring, late summer and early autumn are the best times for establishing shoots which may fail and at any other

Figure 11, Source: Veg Tech, 2000

time of year. Planting plugs is a relatively labour intensive and time consuming and therefore costly process. The fastest method of establishment is laying prefabricated vegetation mats. These can be laid directly onto the roof with a gradient of up to 30 degrees. The advantage is that the roof is ready at once, as opposed to the other methods which are more demanding in their establishment 72, although shifting may occur if the gradient is greater than 30 degrees and the growing substrate is not properly anchored 73.

71 Piga, 1995, p. 30 72 Piga, 1995, pp. 44-46 73 Piga, 1995, p. 20

22

5

Methodology

5.1 Methodology for Noise Assessment A roof with a waterproof membrane, without roof vegetation will allow a given amount of noise to penetrate into the building. This noise can be easily measured. The current common method for establishment of green roofs is by laying a drainage layer with sedum mats on an existing waterproof layer. When this is done, the roof ’s thickness increases by the combined thickness of the sedum mat and drainage layer, and theoretically the insulating capacity increases. If the noise level in the building is monitored after the establishment of a green roof and it is found that the noise level has decreased, then it can be assumed that the noise reducing capacity can be attributed to the roof vegetation and drainage layer. The potential gain of this is an increased comfort level in the spaces where people spend time, by reducing noise levels. The precondition is for both assessments to be identical. When a sound signal is sent from a computer through an amplifier to a speaker it includes all audible frequencies. When the first sound impulses reach a condenser microphone 74 after passing through the building’s roof, the signal is amplified and registered in a computer. The aim of the monitoring is to compare two identical sets of data of this kind, one with green roof vegetation and one without. The measurement method applied in this study was the Schroeder method, see figure 12, which is normally used to assess spatial acoustics, where a room’s absorption capacity for sound is quantified. The Schroeder method is known as a way of measuring impulse responses, but is not known for assessing transmission sound as it can just offer a rough approximation of sound transmission. When using the Schroeder method and measuring the impulse response the result is presented as a function of time and linear sound pressure. In order to express this as a function of sound pressure levels (dB) and frequency, the computer calculates this by a Fast Fourier Transform (FFT) 75. In order to convert the result from a function of sound pressure level to a function of time, a Transfer Function, TF, is used 76.

74 For information on condenser microphones, see Andersson, 1998, pp. 153-154 75 For further information on FFT and TF, see Schroeder, 1965, pp. 409-412 76 For equations, see Kinsler et. al, 1980, pp. 23-25

23

Figure 12, Schematic figure of measurement system, Djamel Ouis, 2004.

When random noise is broadcast from a signal generator contains all frequencies between 0 and 20 000 Hz. The signal is sent to the speaker which then emits the sound. After a short time the condenser microphone receives the signal and forwards this to a crosscorrelater in the computer in which the incoming signal is compared with the broadcast signal, resulting in an impulse response 77. Other standardised methods are used, see below, to measure the transmission sound, the sound that has passed through a wall or roof. The Schroeder method, however, has been chosen for its speed and because it is not as expensive as other alternatives. Furthermore this method is replicable and does not necessitate that the experiment is carried out in an acoustic laboratory. The results from the noise level assessment are an average from sixteen measurements, both with and without green roof vegetation. These sixteen measurements from each experiment showed similar results, which shows that the results are reliable. Furthermore, as previously mentioned, the Schroeder method is replicable which means that equivalent results should be achieved if the equivalent experiment is repeated.

77 Schroeder, 1965, pp. 409-412

24

The standard method normally used for sound pressure measurements of transmission sound gives a reduction figure, which represents the insulation of the material. This kind of measurement should be carried out in an acoustic laboratory as the total absorbing area in the receiver room must be calculated. The method essentially compares the sound pressure in two rooms that are separated by the material to be assessed. The equation is presented as: TL = L1 - L2 - 10 log A S In this equation TL (Transmission Loss) represents the reduction figure, or sound insulation, of the material whose capacity is to be measured. L1 represents the sound pressure level in the broadcasting room and L2 the equivalent in the receiver room. A is the absorption figure for the absorbing surface in the receiver room and S is the area of the material whose noise insulating capacity is to be measured. The material should have a minimum area of 10m2 to provide reliable statistics 78.

78 Kristensen et. al, 1989, p. 28

25

6

Material

6.1 Structural Description of the Cabin The cabin that was used for the noise level monitoring is owned and used by the Direct Works Department at the City of Malmö for road maintenance work during the winter. The cabin is insulated for summer use but is fitted with a small gas fire to ensure that the temperature can be acceptable even in more severe winter weather.

Figure 13, Photo: Jens Lagström

The winter road maintenance cabin is of standard dimensions of 2.6m wide and 7.0m long. Figure 15 on page 28 offers an overview of the layout of the cabin. The two vents above the “Vinterberedskap” sign ventilate the toilet inside. The door to the left of the two is the entrance to the toilet and the door to the right is to the common area in the cabin. The roof is a total of 0.15m thick and consists of a pine panel on the inside, then a fibre glass wool insulation layer and finally the fibre-glass outer roof. Green roof vegetation was laid on the horizontal roof for the study and more details of the laying are found on the next page with a picture.

6.2 Laying Vegetation on the Cabin Roof The mats of roof vegetation that were available for the study had been moved before and when they had been moved they had lost some growing substrate round the edges, so the original mass had decreased. In order to reach the most accurate results, then mosssedum mats should have been established untouched for a year to grow together properly and become denser. The mats were intentionally overlapped to compensate for this and to overcome the problem of cutting the mats to fit exactly on the cabin roof. The cabin is 2.6m wide and the mats are 1.0m wide so in order for the mats to cover the whole roof and to avoid the loss of cover between the mats, then an overlap of 0.2m in two different places on the roof were used, see figure 14.

26

The water retentive mineral wool layer Grodan © had also been used which meant that it was not of original quality, which would have been preferred. It should not be moved after being laid if it is to maintain its quality, but in this case the mineral wool mats were moved when they were wet which meant that they easily broke under their own weight. These mats, however, were laid properly, ie edge to edge. The overlapping of the mosssedum layer can be seen to compensate the poor quality of the mineral wool layer.

Figure 14, Jens Lagström

6.3 Positioning of the Speaker on the Cabin Roof The speaker used for the noise study was positioned 1.0m from the edge of the cabin on its left side in the direction of travel, as there were more vents on the right side through which sound could travel interfering with the measurement of primary sound. The two windows in the cabin have steel shuttering that can be used to prevent vandalism. These shutters were closed for the study to prevent sound leakage. The speaker was positioned on the roof directly above the microphone so that the sound for the study came directly through the roof, see figure 15. The speaker was placed 0.2 m from the roof angled downwards towards the roof as the cabin had a number of windows and vents through which the sound could penetrate if the speaker had been placed so that the sound came towards the side of the cabin. This was due to the fact that it was just the sound that had penetrated the roof layer that was the object of the study.

27

Figure 15, Jens Lagström

28

7

The Results of the Noise Level Assessment

The data has been corrected for the results of the study without green roof vegetation in order to give comparable results. As the roof vegetation was removed the distance from the speaker to the condenser microphone decreased with the effect that the sound could reach the microphone in a shorter time than when the speaker stand was resting directly on the green roof, as the distance had lessened. The sound picked up would be louder with this shorter distance. The data without the green roof has therefore been corrected down by 2.58 dB for sound level and by the same proportion for the impulse response. 2.58 dB corresponds to a lowering of the speaker by 7cm, the total thickness of the green roof and drainage layer. Below are two comparable figures that show the impulse response from both studies. At the top is the impulse response from the study without the green roof and at the bottom is the study with the green roof.

Figure 16

29

There is a clear difference in the sound pressure of the two studies as seen in these two impulse responses. Note that the sound pressure is not the same as sound pressure level (dB). The computer does a Fast Fourier Transform in order to transform these units to dB and frequency.

Figure 17

A Fast Fourier Transform has been carried out to visualise this figure, showing only the direct impulse results. The direct impulse is the first impulse from the speaker that reaches the condenser microphone. The first impulse from each study is analysed and the frequencies are separated by the computer. The distance between the curves shows the noise reduction, which is between 5 and 20 dB, distributed between the different frequencies. At a frequency of 750 Hz the noise reduction is as much as 20 dB, but at a frequency of 1 400 Hz the reduction is just 5 dB. This is explained by the different noise reduction capacity at different frequencies of the individual materials used in the roof. Note that the highest frequency is 3 500 Hz in accordance with Swedish measurement standards.

30

Figure 18

This figure shows the results of the study for the whole impulse response when the monitoring has been going on for about one minute. The upper curve shows the study without the green roof and the lower curve shows the study with the green roof. The horizontal scale shows the frequency by octave band, which in the figure means a range from 20 to 3 500 Hz. The octave bands are furthermore presented in a linear fashion which means that the figure just shows an average of a highly varied curve. As there is a level difference between the two studies it can be concluded that green roof vegetation has a noise reducing effect.

31

8

Discussion

There are many reasons, in my opinion, to lay green roofs, but none of these reasons in itself is sufficiently convincing or economically viable until considered in the context of the overall benefits. As mentioned earlier, it is unlikely that anyone will create a green roof purely to promote biodiversity or for its insulating capacity as it is more cost efficient to use conventional insulating materials. A further positive effect is to be gained from storm water management, but this study shows that green roofs also have a sound insulating effect. It is the combination of all of these benefits which make green roofs more exciting. The noise insulation capacity of green roofs offers wider economic benefits. Who has not had their night sleep disturbed by noise? I believe that the answer is the same, regardless if you live in a small village or large city. Most people are woken sometimes by noise and the main difference is the frequency of this disturbance. Green roofs can contribute to decreasing the occurrence of noise disturbance in a way that not only is good for the individual, but also supports biodiversity and local storm water management. Such a reduction in noise levels would lead to a significant improvement in living conditions and health for the 80 million Europeans who are exposed to unacceptably high noise levels on a daily basis. The long term solution of decreasing noise through the use of green roofs would not merely lead to better living conditions, but a large number of other benefits. It would be interesting to compare with the cost of laying green roofs in target areas with the total cost of noise insulation, increased dimensions for storm water systems, aesthetic improvements and public health improvements. Houses with green roof noise insulation would, unfortunately, probably be more expensive to purchase than a noisier alternative without a green roof, although additional benefits of lower energy use and extended life expectancy of the roof materials and other benefits are included in the price. The results of this study show that earlier estimations of other researchers have been correct. Their assumption that roof vegetation could reduce noise levels by 8 to 18 dB depending on saturation level of the growing substrate and drainage layers, would appear to be correct. The monitoring carried out in this study show that roof vegetation and its drainage layer decrease noise levels by 5 to as much as 25 dB, depending on frequency range, with an average of around 10 dB, or a halving of noise levels. In order to give more reliable data, it would be appropriate to carry out this kind of study on a building with a conventional construction and insulation level. As the cabin used in this study does not have the normal level of insulation, these results cannot be seen to be fully accurate. If the study had been carried out in a reference building of standard dimensions, it would have been possible to have gained more relevant data. This study does, however, prove that green roofs do have an insulating effect. The study had convincing results on a lightweight roof. The noise insulating effect would presumably have been lower on a heavier roof such as a concrete roof. It can therefore be concluded that extensive green roofs are more appropriately used on lightweight roofs from a noise reduction perspective.

32

With more time available it would have been interesting to measure the noise reducing effects of green roof vegetation without the drainage layer. It would similarly have been informative to see if, and to which extent, the drainage layer has an insulating effect. It would have also been interesting to explore the relative insulating capacity of the different layers when both wet and dry. In the course of this study, southern Sweden suffered a series of downpours, so both the vegetative and drainage layers were saturated. It is probable that weight has a relationship to sound insulation, as the mass in the layers is greater when saturated than when dry. I hope to have the opportunity to seek an answer to these questions. It would be desirable to have the opportunity to carry out new noise level testing on an existing roof on which it is planned to lay vegetation, or preferably to carry out standard testing for transmission sound in a laboratory. In retrospect it is always possible to find different ways of doing things, as is the case here. It was a mistake not to adjust the height of the speaker after the roof vegetation was removed. That had simplified things for the reader and would improve the aesthetics of the work. I would also have liked to have a camera to document the work more clearly as a picture is worth a thousand words.

33

9

References

Literature Andersson, J., 1998; Akustik & Buller, fjärde upplagan, AB Svensk Byggtjänst, Stockholm. Arbetarskyddsstyrelsen, 1990; Buller och bullerbekämpning, Arbetarskyddsstyrelsens böcker. Bengtsson, L., 2002; Avrinning från gröna tak, Vatten 58, Lund Berg, R. & Stork, D., 1995; The physics of sound, second edition, Prentice-Hall Inc, Englewood Cliffs, New Jersey. Carlsson, R. & Hellström, P A., 1988; Bygg hörselvänligt, Bättre ljudmiljö för alla; Tekniska hjälpmedel för hörselskadade, Svensk Byggtjänst, Stockholm. Grant, G., Engleback, L., Nicholson, B., Gedge, D., Frith, M. & Harvey, P., 2003; Green Roofs: Their existing status and potential for conserving biodiversity in urban areas, Report Number 498, English Nature, Northminister House, Peterborough PE1 1UA, England. Ingemansson, S., 1977; Bullerbekämpning – Principer och tillämpning, Arbetarskyddsfonden, Stockholm. Johansson, B., 2001; Buller och bullerbekämpning, fjärde upplagan, Arbetsmiljöverket. Karlsson Hjort, H-O., 2000; Lågfrekvent buller i boendemiljön, Boverket, Karlskrona. Kinsler, L., Frey, A., Coppens, A. & Sanders, J., 1980; Fundamentals of Acoustics, third edition, John Wiley & Sons, New York. Kristensen, J. & Rindel, J. H., 1989; Bygningsakustik- Teori & Praksis, Statens Byggeforskningsinstitut, Hörsholm, Danmark. Magnusson, L. & Qvist, L., 1989; Inomhusklimat för människan, Liber, Stockholm. Möller, A R. & Bjurö – Möller, M., 1978; Människan och bullret – Medicinska aspekter på buller och bullerskador, Studentlitteratur, Lund. Naturvårdsverket, 1994; Hälsoeffekter av samhällsbuller – Sammanfattning och uppdatering 1993 – 1994, Redaktör: Hygge, S., Naturvårdsverket, Solna. Piga, C., 1995; Grönare tak – Extensiv vegetation på tak, Stad & Land nr 134, Movium, Sveriges Lantbruksuniversitet, Alnarp. Rubensson, S., 1999; Miljöbalken – Den nya miljörätten, andra upplagan, Norstedts Juridik AB, Stockholm.

34

Stenberg, L., 1992; Den Nordiska Floran, andra tryckningen, Wahlström och Widstrand, Brepols, Turnhout, Belgien. Söderblom, P., 1992; Sedumtak, Lätta gröna tak av sedumväxter, Stiftelsen ARKUS och Byggförlaget, Stockholm. Åkerlöf, L., 1999; Skönheten och oljudet – Handbok i trafikbullerskydd, andra upplagan, Svenska Kommunförbundet, Stockholm

Articles Pal, A.K., Kumar, V. & Saxena, N.C., 2000; Noise attenuation by green belts, Centre of Mining Environment, Indian School of Mines, Dhanbad, India. Journal of Sound and Vibration, 2000, 234(1), Scholes, W.E. & Parkin, P.H., 1967; The insulation of houses against noise from aircraft in flight, Building Research Station, Garston, Watford, Herts, Great Britain. Applied Acoustics, Volume 1, Issue 1, January 1968. Schroeder M. R., 1965. New method of determining reverberation time. Journal of the Acoustical Society of America 37, 409-412.

Publications Veg Tech. 2000; Takvegetation, Mångfunktionella levande tak, Vislanda.

Press Glimberg, M., 2003-04-30; Bullerstörda får vänta på beslut, Sydsvenska Dagbladet.

Personal communications Giesel, D., 2002-10-14; International Green Roof Conference, Augustenborg’s Roof Gardens, Malmö Skärbäck, E., 2002-09-13; Lecture on the advantages of green roofs, Augustenborg’s Roof Gardens, Malmö Söderblom, P., 2002-09-06; Lecture on the motives for green roof development, Augustenborg’s Roof Gardens, Malmö Velazquez, L. S., 2002-10-14; International Green Roof Conference, Augustenborg’s Roof Gardens, Malmö

35

Internet Exploring the ecology of organic greenroof architecture; History, www.greenroofs.com/ history.htm, Publisher Velazquez, L. S., last updated 2002, date of reference 2002-1201. Veg Tech, Grodan©, Tillbehör till takvegetation, bild 8, www.vegtech.se, Publisher: Porselius, T., last updated 2003-11-28, date of reference 2004-01-20 Veg Tech, Referensobjekt, Takvegetation och bjälklag, bild 10, www.vegtech.se, Publisher: Porselius, T., last updated 2003-11-28, date of reference 2004-01-20 Veg Tech, Takvegetationens historia och nutid, www.vegtech.se, Publisher: Porselius, T., last updated 2003-11-28, date of reference 2004-01-20

Mentors Per Hillbur Djamel Ouis

36

37

38

39

INTERNATIONAL GREEN ROOF INSTITUTE

Do Extensive Green Roofs Reduce Noise? The human being controls her environment by the help of her sense of hearing, which can not be turned off in any circumstances. Since it is so important from a health perspective to have a decent sleep, noise disturbance can cause severe consequences for the state of health. Approximately 20 % of the population of Europe is exposed to levels of noise that are unacceptable which often leads to irritation, disturbance of sleep and a risk of negative effects on health. Noise involves an annual cost within the EC on fully10 Billion €. The cost of noise for the Swedish society is estimated to 21 Million € each year. The question if extensive green roofs reduce noise or not is answered by measurements of noise levels on a reference building without a green roof, which is compared with the identical measurements of noise levels on the same building, but with a green roof. The method that has been used for the measurements of noise levels is the Schroeder method. The results of the impulse response from the comparative measurements of noise levels confirm that extensive green roofs do reduce noise, which is an additional positive effect supporting the development of green roofs.

40 ISBN 91-973489-9-6

Augustenborg’s Botanical Roof Garden This publication is a part of a bigger project aiming for greener, more sustainable cities. Augustenborg’s Botanical Roof Garden, in Malmö, Sweden, is a centre for research, information and inspiration about living green roofs. On top of a group of industrial buildings in the area of Augustenborg, the City of Malmö, supported by the EU and the Swedish Ministry for the Environment has created a unique establishment to inform the public about the advantages of green roofs. Living green roofs have positive effects on the environment in many different ways: • Storm water management • Energy efficiency • Better local climate • Noise reduction • Habitats for wildlife • Aesthetics and well-being Augustenborg’s Botanical Roof Garden is made to inspire with its beauty, at the same time as it is an important research site where a number of different universities are involved in research around for example green roof technologies, plant materials, nutrient balance, storm water management, run-off quality, noise reduction, energy saving, increased membrane life, users’ aspects, biodiversity and development of new habitats. The botanical roof garden can be visited live, or on the website. International Green Roof Institute (IGRI) In connection with the botanical roof garden, a green roof institute, IGRI, is working to promote the use of green roofs. The most important task of the Institute is to co-ordinate the research around living green roofs, as well as educating and spreading information. IGRI co-operates with Malmö University on a university course on green roofs. Yearly international and national research seminars are arranged to share knowledge and to further the aim of IGRI – to contribute to a sustainable future through an increased use of green roofs. More about IGRI and Augustenborg’s Botanical Roof Garden on the web: www.greenroof.se