Environmental Noise Data from Dublin City Council Designated ‘Quiet’ Areas to Establish Compliance with the Target Noise Levels

Chapter 1

Introduction

1.1 Environmental noise

Environmental noise is any unwanted outdoor sound created by human activities that is detrimental to the quality of life of individuals. For people living in urban centres, environmental noise has become an unavoidable part of their everyday lives. The type, the level, location and time of the sound occurrence, plays a major role in how people perceive the impact of sound. Environmental noise particularly relates to sources of noise which originate from industry and transport frameworks. The concern is that by this type of noise becoming overfamiliar to us, we are less aware of it and of the damage it is doing to our health and wellbeing.

1.2 The European Noise Directive

Statistics presented by the European Commission [13], estimate that around 20 percent of the Union’s population or close to 80 million people suffer from noise levels that scientists and health experts consider to be unacceptable. In this environment most people become agitated, their sleep is disturbed and adverse health effects may be experienced such as an inability to relax resulting in anxiety, frustration, anger and even mental health problems.  An additional 170 million citizens are living in so-called “grey areas” where the noise levels are such to cause serious annoyance during the daytime. Thanks to legislation and technological progress significant reductions of noise from individual sources have been achieved. For example the noise from individual cars has been reduced by 85% since 1970 and the noise from Lorries by 90%. Likewise the noise footprint created by the modern jet around an airport has been reduced by a factor of 9 compared to an aircraft with 1970s technology. It has been estimated that since the 1960’s, the average perceived noise level of engine noise has been reduced by 20dB.

However data covering the past 15 years does not show significant improvements in exposure to environmental noise especially road traffic noise. The benefits of technological advancements have been reversed by the huge increase in the volume and activity of road traffic, air traffic and the expansion of high speed rail. In the case of motor vehicles other factors are also important such as the dominance of tyre noise above quite low speeds (50km/h) and the absence of regular noise inspection and maintenance procedures. For some sources such as railways and a wide range of noisy equipment used outdoors there are no Community or international standards setting emission limits. A number of Member States are planning national legislation for these products, which could cause problems for the functioning of the single market.

The European Noise Directive (END) was established in 2002 to create a common standard for EU member states in how they deal with and manage noise pollution. The directive’s stated aim was to:

“Define a common approach intended to avoid, prevent or reduce on a prioritised basis the harmful effects, including annoyance, due to the exposure to environmental noise.”

(European Parliament and Council, 2002)

This directive obliges each nation to assign the relevant local authorities to create noise maps every five years, inform and consult the public about noise exposure and investigate methods of noise management and mitigation.

1.3 Dublin City Council ‘Quiet’ Areas

In response to this directive, Dublin City Council was given the responsibility of establishing a noise monitoring network across the city and they have also designated eight ‘quiet’ areas to provide the citizens of Dublin with an escape from the high noise levels in the city. It has been stressed that although these areas have been identified and publicized as having low environmental sound levels, they are not by any means free from all of the common types of sound caused by social and recreational activities as well as construction or maintenance works [15]. The goal of the council is “to protect relatively quiet areas against an increase in noise” and the main purpose of these areas is presented below:

“The aim is for these areas to be used by the public, so in addition to all the other benefits of using the facilities of a park, there is also the added benefit of recreating in a tranquil and quiet setting.”

(Dublin City Council, 2013)

Environmental noise levels in accordance with the END, are to be maintained below 45 dBA during night time hours and below 55 dBA during daytime hours at these sites. These ‘quiet’ areas, which include public parks and green spaces, are listed below:

1. Blessington Street Basin, Phibsborough
2. Edenmore Park, Raheny
3. Mount Bernard Park, Phibsborough
4. Dollymount SAA
5. St Anne’s Park, Raheny
6. Palmerstown Park, Dartry
7. Ranelagh Gardens, Ranelagh
8. The Cabbage Gardens, Dublin 2

Figure 1-1: Map of DCC ‘Quiet’ Area locations

At two of the ‘quiet’ areas, permanent noise monitors are in operation, as part of the long term environmental noise monitoring system, created by DCC, in collaboration with Sonitus Systems [14]. The noise monitoring system includes monitors deployed at 14 sites around the Dublin City Council region. The noise monitor used is the Sonitus systems EM2010 and it has been reported that these particular units have achieved in excess of 95% of the perfect ‘score’ of 105,210 readings per annum with most units recording data more than 97% of the time [16].

1.4 Research Aims

The aim of this research project is to collect and analyse environmental noise data from some of the Dublin City Council designated ‘Quiet’ areas and to establish whether these areas are complying with the target noise levels. This degree of compliance will be measured using the L_DEN, the Day Evening Night Sound Level, which is the average sound level over a 24 hour period. These calculated levels will give an indication of the amount of time the sound levels are keeping within the noise guidelines at these locations.

As the ‘quiet’ areas may differ in landscape, features and surroundings, it will also be necessary to perform a site analysis of some of the areas where sound levels are being investigated. This research will be useful in identifying possible sources of noise, which may offer possible explanation for sound levels at the sites.

To identify the periods of time for which the target levels are being exceeded, different statistical techniques will be used. Yearlong noise datasets of the sites will be analysed to identify any variation in sound level with different hours of the day, weekdays and seasons. To measure the relationship between sound levels and an environmental factor such as weather, data such as rainfall, wind speed and wind direction from Met Éireann’s Dublin synoptic station, will be compared to the noise datasets available.

An additional objective which has over the course of the project is to investigate the park user’s perceptions of the sound levels in public parks. These perceptions will be examined using surveys, both a preliminary online survey for a large study group and later in-situ surveys for utilization in the designated ‘quiet’ areas for a smaller study group. The reasoning behind first researching the topic with an online survey is it provides the opportunity to gather numerous responses from a large and varied group and assemble a broader perspective of the topic. Lessons learned from the preliminary survey can then be used for application in the in-situ survey, which will provide a more explicit impression of the sound levels in a specific ‘quiet’ area. The in-situ survey may also introduce and highlight other influencing factors which impact the park user’s perception of sound levels.

1.5 Summary

The overall goal of this project is to determine whether the sound levels in the designated ‘quiet’ areas are keeping within the target noise levels as set by the END. In addition, the project intends to establish the perceptions of individuals using the ‘quiet’ areas and compare these perceptions with the actual noise levels. No similar research has been previously undertaken in Ireland and the methods used would be consistent with international practice. The results and recommendations of this research may be used to inform future decisions on noise mitigation measures.

Chapter 2 of this dissertation presents a Literature review which gives details of studies relevant to this research. Chapter 3 presents the Methods section which focuses on the different techniques used during the research. Chapter 4 presents the results of the research. Chapter 5 presents the discussion of the results and Chapter 6 outlines the conclusions of the project and the recommendations for any future work undertaken on the topic.

 

 

 

 

 

 

 

 

 

 

 

Chapter 2

Literature Review

2.1 Introduction

As this project is established on guidelines from the European Noise Directive (END), 2002, search of the scientific literature was conducted on noise studies carried out in Europe. The different aspects of the research contain details of studies involving both sampled and continuous noise monitoring systems and they are grouped in sections below. The majority of the literature was sourced from Science Direct.

2.2 Irish noise projects

Search into the relevant literature began by investigating whether any similar projects to this one had been previously carried out in Dublin or in Ireland. These would have related to environmental assessments of noise levels in public parks around the country and the noise level perceptions of park users.

There was no evidence of studies such as these in the academic literature. Many Irish noise studies and assessments have indeed been implemented; however these mostly concern the impact of road traffic noise alone on the population and the environment. The need for noise management strategies is suggested through the results of these studies. For example, a study estimating the extent of human exposure to traffic noise in central Dublin showed these exposure levels are considerable, and in relative terms, it is worse for the night-time period [2]. This study really highlighted the problem with high noise levels during the night-time in Dublin as this is the time of the day where noise levels are known to have the most harmful effects on human health [3], [6].

The study used the Harmonoise noise calculation method for calculating noise at the most exposed building façade. Noise measurements were recorded every 10 s during the measurement period of fifteen minutes. LDEN and LNIGHT determined residential exposure while LDEN alone determined worker exposure. The exposure estimates suggest that 53.6% and 28.0% of workers and residents respectively are exposed to noise levels exceeding 70 dBA during day-time while the corresponding figure for night-time noise exposure greater than 45 dBA is 90.2%. The study states that if the present exposure levels were extrapolated for the entire city, the magnitude of the problem is considerable. King et Al, 2009 state that “The impact that these levels of exposure are having on individual health and the quality of life as well as the economic impact in terms of reduced productively of workers is likely to be considerable also.”

A study by King et al. (2009) into a boardwalk development in Dublin, built to reduce pedestrian exposure to environmental pollutants including noise, was somewhat similar to this current study [1]. These boardwalks were designed and constructed to provide shelter for pedestrians from the surrounding noise and air pollution and the results found this goal was being achieved. Twenty-seven pairs of noise measurements were taken along the boardwalk and along the footpath simultaneously to quantify differences in pedestrian exposure to noise. The results showed that the level of noise on the footpath would be approximately 6 dBA greater than the noise on the boardwalk. Further reductions in noise levels were noted when the height of the boundary wall between the boardwalk and the road was increased. To achieve this, a noise barrier located on the top of the quay wall was modelled at a number of different heights and a possible relocation of the boardwalk 1 m below its current position was also investigated. Considerable potential was found to exist for increasing the attenuation of the noise in the study area. This could be achieved by dropping the boardwalk by 1 m and extending the height of the boundary by 1 m, giving an extra attenuation effect of 3 dBA while even greater attenuation could be achieved with the introduction of a 2 m high barrier.

Urban parks have a purpose akin to these boardwalks, as they are created to provide escape for people from the stress and crowding of urban centres, and they demonstrate a positive environmental impact on the health and wellbeing of the people of Dublin.

2.3 European Studies

Search into the literature was then extended into studies conducted in other European countries where the END is also effective. Environmental noise has been monitored and analysed in some of the biggest cities in the world, metropolitan cities such as Rome [4]. In addition, studies undertaken as part of the QUADMAP project included the cities of Florence, Italy, Bilbao, Spain and Rotterdam, Holland [7]. What was most interesting about the Roman study is that it included not only the classic quantitative analysis of A-weighted sound pressure levels but also a more qualitative approach in its measurement of the effect of environmental noise. It is a socio-acoustic survey which focused on three popular urban parks in Rome. These parks were selected from a preliminary online survey which identified the parks most popular and frequented by Romans.

A method of binaural audio recordings using headphones and a digital recorder was implemented in Rome which allowed the researcher to observe and notice particular sound events and analyse them later. For a more subjective angle on the sound levels in the parks, users were interviewed at random. The questions concerned the user’s age, gender, how often and for what duration they visited the park, the main purpose of their visit and their perception of the sound environment of the park. This study was also attentive to the user’s impression of the facilities and features of the park such as toilets, playground, wildlife, flowers, cleanliness, security, trees, tranquillity, silence and natural sounds.

Results from the Roman survey proved the sonic environment of the three parks is considered “good” or “excellent”, due to the influence of environmental features mentioned above. However in reality the sound pressure levels in these urban areas is nearly always higher than the survey respondents expected. The daytime LAeq limit is 50 dBA for daytime according to Italian legalisation however, two out of the three parks studied were above this level for the majority of the time (87% and 93% of the time respectively). The overall conclusion of the study was that the surrounding environment of the park had positive influence on how the noise levels in the park were perceived rather than the actual sound levels themselves.

2.4 Urban Quiet Area Projects

An EU project on Quiet Urban areas called “QUADMAP” is ongoing in the cities of Florence, Italy Bilbao, Spain and Rotterdam, Holland [7]. The main objective of the QUADMAP project is to develop a harmonized methodology for selection, assessment and management of Urban Quiet Areas. The assessment methods considered combine both quantitative and qualitative parameters. The management of the quiet areas is achieved though noise mitigation to increase the usability of the areas and user’s satisfaction. The project also aims to help to increase the understanding of the definition of an Urban Quiet Area, the meaning and the added value for the city and its citizens in terms of health, social safety and lowering stress levels.

To test whether the provisional methodology would work, the selected areas (pilots) were examined on:

• Occurring noise levels, by conducting short and long term noise measurements.
• Visitors and users were questioned on perception and appreciation of the selected area.
• Expert analysis on non-acoustic factors and general characteristics of the pilot areas were executed.

A presentation given at the 2013 Internoise congress in Innsbruck, Austria [8], highlighted some of the key findings and observations of the pilot study in Florence so far. For this study, the candidate quiet areas were six school gardens and the questionnaires were distributed to a combination of students, teachers, school staff and parents. The researchers noticed that it was not possible to use the same questionnaire for the entire sample group as children have more difficulty in understanding the questions than adults. This was particularly the case for questions regarding the user’s perception of the sound levels around them and the usability of the younger user’s responses to these questions was uncertain. In addition, they also noted that it is not possible to use the same questionnaire for all urban quiet areas as the typologies will differ, for instance the Florence study areas are school gardens while Rotterdam’s study areas are public parks.

A part of the methodology of QUADMAP was to divide the area into sub-homogeneous areas, defined HUA (Homogeneous Units of Analysis) in which to make a further acoustic and non-acoustic detailed analysis. HUAs are selected, based on three considerations, landscape, use and distance and presence of acoustic sources. Each HUA must have similar landscape, a specific use and the same visual and acoustic impact of noise sources in order for many HUAs to be compared to each other effectively.

In terms of noise measurements, both short and long term strategies were implemented concurrently in order to perform a more comprehensive analysis of the area. The short term measurements were performed during a 15 minute interview with the individual user. In the background the long term measurement was also in progress and ran for a week.

Each pilot study area in Florence had been identified to be mainly affected by road traffic. The results presented show the LAeq levels for the weekly long term measurements to be over the Italian guideline of 55 dBA. 24% of users suggested using sound barriers and 23% suggested reducing traffic noise to improve the acoustic environment in the quiet area. 44% of users suggested introducing more vegetation to the area to improve it visually.

A similar presentation as part of QUADMAP was also given at this congress for one of the pilot study areas of Bilbao, an inner city square [9]. Expert analysis identified the main noise source as being traffic noise which is both audible and visible at the location. When asked, the interviewees specified traffic noise, considered to be unpleasant and birds, considered to be pleasant as the dominant sound sources in the quiet area. 44% of the users were satisfied overall with the area. For short term measurements, the L_Aeq levels were found to be over 60 dBA for both the morning and evening. For the long term measurements, the L_DEN, L_DAY and L_EVENING levels were recorded as being between 61 to 66 dBA over the course of one week. The L_NIGHT level was recorded as being between 53 to 57 dBA over the week. These levels all exceed the END thresholds for acceptable day and night time sound levels of 55 dBA and 45 dBA respectively. Possible interventions suggested to reduce traffic noise and its effects include traffic reorganization, more fluency of traffic, creating a pedestrian preference area and urban barriers for traffic noise.

Rotterdam selected two city parks as pilot study areas [10], one is a small park enclosed by urban streets and buildings and the other is the largest urban park in the Netherlands. The outcomes of the Rotterdam pilots showed that respondents appreciate when the quiet area is well kept, accessible and when natural and visual elements are present. The acoustic environment was found to be of less importance than these environmental features and this was similar to the findings of the survey in Florence.

In addition, in Rotterdam the acoustic environment appeared to have less importance due to the fact noise levels were rather low. The L_Aeq was calculated to be between 52-57 dBA. However, in the case of the schoolyards in Florence, the reason for the inattention given to the sound levels was attributed towards the activities undertaken in the areas. Expectations of users and visitors of these schoolyards are not especially focused on noise but more on safety and natural elements. Many of the users being interviewed about the area were children. It can be assumed that they are not particularly interested in noise, especially when playing.

2.5 Continuous and Sampled Monitoring

A study initiated in Toronto, Canada, as a result of the studies being implemented in Europe in response to the END, was completed to assess population-level exposure to traffic noise [5]. Little had previously been known about this topic due to the lack of noise assessments undertaken in Canada. This is a useful study as it includes results from both sampled and continuous noise monitoring methods. Two strategies were used to capture both the short-term and long-term noise exposure among the population of Toronto. The short-term strategy involved selecting locations with a high probability of raised noise levels depending on population density and adjacency to busy roadways. Two cycles of 30 minute long, real-time noise measurements of traffic noise during the daytime were conducted at these locations. The long-term strategy involved continuous measurements of noise recorded for seven days at ten sites.

For the continuous study, no significant day-to-day variation was observed among the week-long sites. The results identified the quietest time of any given day within a week to be around 4 a.m. Noise levels gradually increased over time until 11 a.m., and remained relatively constant until 4 p.m. Noise levels increased again from 4 p.m., and reached their daily maximum around 7:30 p.m. Higher levels of noise were observed during weekdays compared to weekends although the difference was not significant. Overall, daytime noise averages were consistently and significantly higher than night-time noise averages. The median daytime noise level was 54.9 dBA, and night-time was 52.4 dBA. The results showed that the night-time average sound levels for all of the week-long samples exceeded WHO’s night-time noise recommendation for health protection of 40 dBA [6].

The main finding of this sampled study was that variability of traffic-related noise was primarily spatial in nature (59.6%). Traffic volume, length of arterial road, and industrial area were the three most important variables identified to explain the majority of the spatial variability of noise, after accounting for temporal variations. Also, the repeated short-term measurements collected at identical locations exhibited little seasonal variation between cycles, suggesting that noise measurements collected during one season may closely approximate the noise patterns of another season. Zuo et al. states “In comparison to the 16-h equivalent sound level guideline for outdoor locations set out by the Ministry of the Environment of the Province of Ontario, 80% of our sampled locations exceeded this guideline (i.e. 55 dBA, 16 h).”

Another study using a continuous noise monitoring system is one implemented in Gdansk, Poland [11]. It presents the comparison of noise assessments based on classical, static computationally produced noise mapping on one hand and dynamic noise mapping, which relies on real data from continuous measurements from a monitoring network, on the other. In response to the END, a noise monitoring network of 40 stations has been operating in the city of Gdansk since 2008, similar to the network established by Dublin City Council. The noise indicators presented in this study were calculated on the basis of the data acquired from 14 selected monitoring stations which gather mostly road traffic noise but also a selection of road, tramway and railway noise, depending on the location. The results of the measurements were compared with data subtracted from the noise maps.

The main findings were that the measured values from the noise monitoring stations were higher than the calculated values from the noise maps for the majority of the time. This was especially in case for  the L_NIGHT indicator and it is speculated, that at least some of the noise underestimation is because the maps neglect local sound sources positioned near the receiver like car parking manoeuvres and all socially related noises. The railway noise predicted in the map was about 6 dB lower than measured using a noise monitoring station for all indicators. The study also analysed the influence of seasons and weekdays on noise indicators which were both found to have a significant impact. For railway noise, the differences between winter and summer are significant, up to 3 dB while for road traffic noise the difference generally doesn’t exceed 1 dB. For L_DAY summer is the quietest season, spring and autumn are the noisiest. Spring is also the noisiest season considering L_EVENING indicator. The noise indicators for Sundays are up to 3 dB lower than for workdays. For Saturdays the levels are about 1 dB higher than on Sundays. Considering L_NIGHT noise indicator, the nights of Thursday/Friday and Wednesday/Thursday are the noisiest in the week.

2.6 Summary

The QUADMAP project and Roman study in particular stand out as best practice in the noise monitoring of quiet areas in urban settings and has helped to inform this study in terms of its method and approach.

 

 

 

Chapter 3

Methods

Inspired by methodology used by the QUADMAP projects in European cities, introduced in Section 2.4, the steps involved in the assessment of the main study area Blessington Street Basin include:

1. Site analysis – Identification of the main noise sources, acoustical factors and general characteristics of the study area. The installation setup of the permanent noise monitor at the site was also investigated.
2. Data analysis – Investigation of long term noise measurements at the study area
3. Preliminary Online survey
4. In-situ Survey – Visitors of the park were questioned on their use and assessment of the study area, in particular their perception of the sound levels and soundscape

Site and Data analyses were also carried out on two other sites, a second designated ‘quiet’ area Bull Island, Clontarf and a proposed ‘quiet’ area site, Woodstock Gardens, Dublin 6. Permanent noise monitors are installed at all these locations.

3.1 Site Analysis

Tools used for the Site analysis include Google Maps and Google Compass. Google maps allowed the surrounding areas of the sites to be analysed, noise sources to be identified and the distances between noise sources and the noise monitors to be measured. Google Compass was used to determine which directions noise could be potentially blocked at the sites by buildings, landscape features etc. The Dublin City Sound levels website and site visits were used to identify the locations of the monitors and microphones.

3.2 Data Analysis

The platforms used to analyse both noise and survey response data were IBM SPSS Statistics 22 and Microsoft Excel.

3.2.1 L_DEN Calculations

Before any statistical analysis could be performed, the sound pressure levels in decibels had to be converted into Pascal’s for each site using equations (1) and (2) below. The reason for this is since SPLs are a logarithmic scale, decibel values cannot be used for mathematical operations until they have been converted to their Pascal equivalent. The below formulae were obtained from [20].

SPL dB=10*log10P                          (1)

Sound Pressure Pa= 10SPL10                (2)

Where P is the sound pressure measured in Pascal’s and SPL is the sound pressure level measured in decibels.

To measure compliance at the sites with the target noise levels set for ‘quiet’ areas, the L_DEN was calculated. L_DEN is a 24-hour equivalent continuous level in dBA where 5 dB is added to the evening noise level, L_EVENING measured from 19.00 hours to 22.59 hours and 10 dB is added to night-time noise level, L_NIGHT, measured from 23.00 hours to 06.59 hours. These additions to the levels are included to account for the fact that certain noises may be perceived as more annoying at different times of the day.

The formula for the L_DEN sound level is shown below:

LDEN=10log12412*10LDAY10+4*10LEVENING+510+8*10LNIGHT+1010           (3)

3.2.2 Statistical tests

In order to analyse relationships between the datasets available and the different survey responses, a number of different statistical tests were used with SPSS, outlined below.

Chi-Square test

The chi-square test for association investigates whether two categorical variables are associated. It tests for the association/independence between two nominal variables. Nominal variables are categorical variables which have categories with no natural order to them (e.g. males/females). A Chi-Square test is run as part of a cross tabulation, which is presented in a contingency table with observed and expected frequencies for each cell of the design.

The results of many statistical tests, including the chi-square test for association, are presented in the Chi-Square Tests table, an example of which is shown below. The most useful statistic for this research was the Pearson Chi-Square which measures the significance of the association between the variables being analysed. The Chi-Square test evaluates the validity of the null hypothesis, which states the relationship between the variables is not significant. In statistics, the typical level of significance adopted to reject the null hypothesis is p < 0.05. If p < 0.05, the null hypothesis can be rejected and an association between the variables is confirmed.

Table 3-1: Chi-Square Tests example

	Value	df	Asymp. Sig (2-sided)
Pearson Chi-Square	81.906a	8	0.000
Likelihood Ratio	83.075	8	0.000
N of Valid Cases	858

1. 2 cells (13.3%) have expected count less than 5. The minimum expected count is 4.28.

The subscript (a.) to the Chi-Square Tests table also provides information on how many cells have an expected cell count less than five. This refers to the fact that the Chi-square test is nonparametric, which means the strict assumption of population distribution is relaxed however, there is still a requirement to fulfil the expected frequency in each cell which must be 5 or more. The Chi-square test is not reliable if the proportion of cells with expected frequency less than 5 is as high as 25% or more. This means this particular case of 13.3% is in accordance with this rule.

The main problem with the chi-square test for association is that although it informs whether the null hypothesis of no association can be rejected, it does not provide information on the strength/magnitude of any association. Two measures that do provide measures of effect size are presented in the Symmetric Measures table, as shown below:

Table 3-2: Symmetric Measures example

Phi (φ) and Cramer’s V are both measures of the strength of association of a nominal by nominal relationship. Phi is only suitable when you have two dichotomous variables, which are nominal variables with only two categories. In all other cases and for this research in particular, Cramer’s V is used. Both these measures can be interpreted in the same manner as a correlation. Cramer’s V ranges from 0 to +1, whilst Phi’s range can exceed +1.

Spearman’s Correlation

The Spearman rank-order correlation calculates a coefficient, ρ Spearman’s rho, also known as r_s, which is a measure of the strength and direction of the association between two continuous variables.

The magnitude of Spearman’s correlation coefficient determines the strength of the correlation. Some general guidelines for assigning strength of association to particular values are provided by Cohen (1988) [17]:

Table 3-3: Assigning Strength of association

Coefficient value	Strength of Association
0.1 < | ρ | < 0.3	Small correlation
0.3 < | ρ | < 0.5	Medium correlation
| ρ | > .5	Large/Strong correlation

Where | ρ | means the absolute value of ρ (e.g., | ρ | > 0.5 means ρ > 0.5 and ρ < -0.5).

Coefficient of determination: The coefficient of determination is the proportion of variance in one variable that is “explained” by the other variable and is calculated as the square of the Spearman correlation coefficient (ρ ²). This “explained” refers to being explained statistically, not causally.

One-Way ANOVA

A One-Way ANOVA is used to determine whether there are any statistically significant differences between the means of two or more independent groups. In order to interpret the one-way ANOVA, the assumption of homogeneity of variances for the data being analysed must be tested. The assumption of homogeneity of variance is that the variance within each of the populations is equal. It is tested using Levene’s test of equality of variances, which is one way of determining whether the variances between groups of the dependent variable are different.

If Levene’s test is statistically significant (p < .05), you do not have equal variances and have violated the assumption of homogeneity of variances. If Levene’s test is not statistically significant, (p > .05), you have equal variances and you have not violated the assumption of homogeneity of variances.

In this research, the assumption of homogeneity of variance was always violated which the standard one-way ANOVA could not be interpreted and a modified version of the ANOVA had to be used instead. The Welch ANOVA was used.

Table 3-4: Welch ANOVA example

Robust Tests of Equality of Means
Leq (Pa)
	Statistica	df1	df2	Sig.
Welch	15.460	23	31315.311	.000
a. Asymptotically F distributed.

The meaning of each part of the above table is shown below:

Table 3-5: Meaning of Welch ANOVA table

	Part	Meaning
df1	3 in (3, 14.574)	Indicates the Between Groups degrees of freedom (“df1”)
df2	14.574 in (3, 14.574)	Indicates the Within Groups [Error] degrees of freedom (“df2”)
Statistica	14.821	Indicates the obtained value of the F-statistic (obtained F-value)
Sig.	p < .0005	Indicates the probability of obtaining the observed F-value if the null hypothesis is true

The results are presented as:

Welch Fdf1, df2= Statistica, p=Sig.

 

 

3.3 Online Survey

A preliminary online survey was designed using SurveyMonkey to assess people’s impressions of sound levels in Dublin public parks and quiet areas in general. SurveyMonkey is a free platform for creating customizable surveys which can be made available to possible respondents via a web link, email, website or social media. The survey URL was distributed via email to all undergraduates, academic and administrative staff of Trinity College Dublin, inviting them to participate.

 

 

 

 

 

 

 

 

Figure 3-1: SurveyMonkey Analyse Results tab

Figure 3-1 shows the Analyse Results section of the survey. SurveyMonkey allows the user to view and analyse their results at any time during the survey collection process. Here the user can see a summary view of the data, browse individual responses, create and export dynamic charts, view and categorize open-ended responses, and easily download the results in multiple formats.  These exports can then be used for further analysis, with Excel and SPSS.

For open-ended responses, the Text Analysis feature could be used to categorize and filter the respondents’ important words and phrases.  The results can either be displayed in Cloud or List view, as illustrated in the figures below. In Cloud view, the most frequently used words in the responses are represented as the words with the largest font.

Figure 3-2: Text analysis Cloud and List views

The survey was designed specifically for park users.  Fifteen questions were asked to collect information on the following topics:

1. Respondent’s personal information (age, gender)
2. Presence of the interviewees in the park (frequency, days of the week, time of day) in terms of potential multiple answers to be chosen among proposed option
3. How close does the respondent live to their chosen park, a choice of within walking distance, a short driving distance (up to 10 minutes) and more than a 10 minute drive
4. The main reason for frequenting the park (walking, jogging, participation in sport, relaxation, social and other reason which they could themselves)
5. Assessment of the sound levels in the park, expressed on a scale from 1 (very quiet) to 5 (very noisy)
6. Assessment of whether they agree the park could be considered valuable as a “quiet area”, expressed on a scale from 1 (strongly agree) to 5 (strongly disagree)
7. Assessment of a list of sounds to determine the ones they find annoying in a park, expressed from 1 (most annoying) to 10 (least annoying)
8. Indication of the sounds they enjoy to hear in a park from a list (children playing, birdsong, natural sounds, people talking, people playing sport, music, other)
9. Measures they thought could reduce the noise levels which annoy them (open question)

3.4 In-situ survey

For the in-situ survey, participants were Dublin public park users. Potential participants were approached in the setting of a public park and invited to participate.  Participants were selected randomly from a wide range of age groups but people under 18 years of age were not recruited, as children in particular may have difficulty in understanding some of the questions, noted by previous research and discussed in Section 2.4.

The paper-based questionnaire was designed to elicit park users’ perception of soundscape and to identify the sounds which they found enjoyable and annoying. The paper-based questionnaire had less but almost identical questions to the online survey.  The reason for shortening the survey to twelve questions for the paper-based survey was that recruiting respondents for face-to-face interviews was expected to be more difficult.  In order to get people to participate, the survey couldn’t take up too much of their time, less than five minutes, and it had to be concise.

To capture the peak flow of visitors, surveys were planned to be conducted in the afternoon during a weekday and weekend. The reason for interviewing people on both a weekend and weekday was to see if there was any difference between the two in terms of number of people in the park, willingness to participate, differences in responses etc.

 

Chapter 4

Results

4.1 Site Analysis

Blessington Street Basin

This site is a designated ‘quiet’ area and urban park in the city centre. The major noise source at this site is traffic noise caused by the many roads located around the site. The microphone at this site is located on the roof of the gatehouse building. It is exposed to noise from all directions except noise from the east, which may be blocked somewhat by the building wall adjacent to the gatehouse. This wall would also tend to reflect and reinforce noise directed from a westerly direction. The distance of the microphone from the ground is approximately 4.4m.

Possible noise sources at this site:

1. Phibsborough Road – a main regional road out of the city centre
2. Junction between Blessington Street and Berkeley Street/Mountjoy Street
3. Dorset Street/N1 – a national road
4. Scrapyard

Noise from the Phibsborough Road (245m away from the monitor) is mostly blocked by residential housing between it and the monitor. Noise from Dorset Street on the right is also a less significant noise source due to the greater distance between it and the noise monitor (331m). Noise from the junction at Blessington Street will be the most picked up by the microphone due to the closer distance (132m).

Figure 4-1: Blessington Street Basin Site analysis

Bull Island

This site is a designated ‘quiet’ area, a nature conservation area and wildlife habitat located on Dublin Bay. The site is a grassy/sandy area which is very exposed to the elements and sound levels are expected to be mostly influenced by weather and waves. The microphone at this site is located on the south-westerly edge of the interpretative centre, shown below, and is blocked from wind noise traveling from northeast, north and northwest directions. The microphone is approximately 5m above ground level.

Possible noise sources at this site:

1. Weather and Waves from Dublin Bay
2. Link road leading to causeway

Figure 4-2: Bull Island Site analysis

Woodstock Gardens

This site is based in a retirement village in a highly residential area, with high volumes of slow traffic moving nearby. The microphone was not able to be precisely located at this site due to it being installed in a residential area within private property.

Possible noise sources at this site:

1. R117 – a busy regional road into the city
2. Marlborough Road
3. Residential area
4. Hospital

The most significant noise source is the R117 which is a busy regional road, located 110m from monitor. The Marlborough Road, with lesser volumes of traffic than the R117 and at a greater distance away from the monitor (137m), may affect noise levels, though to a lesser extent. Some noise may also be picked up by the microphone from the surrounding residential area.

Figure 4-3: Woodstock Gardens Site analysis

4.2 Data Analysis

4.2.1 L_DEN Calculations

The sites analysed for these calculations include two ‘quiet’ areas, Blessington Street Basin, Bull Island and a proposed ‘quiet area’ Woodstock Gardens in Dublin 6. A noisy site on the Dublin noise monitoring network, Ballymun Library, was also analysed to determine whether these sites experience low noise levels compared to other sites around the city. L_DEN, the average noise level over a 24 hour period, and L_NIGHT, the average night time noise level, were calculated for these sites.

Table 4-1: L_DEN Sample Calculation for Blessington Street Basin

Noise level average	Mean (Pa)	dBA	Noise level average	dBA
LDAY (07.00 – 18.59)	637819.25	58.1	LEVENING+5	57.8
LEVENING (19.00 – 22.59)	191342.43	52.8	LNIGHT+10	62.6
LNIGHT (23.00 – 06.59)	183729.97	52.6	LDEN	60.1

Table 4-2: L_DEN and L_NIGHT values for the four sites

Site	LDEN (dBA)	LNIGHT (dBA)
Blessington Street Basin	60.1	52.6
Bull Island	70.5	63.5
Woodstock Gardens	54.9	47.6
Ballymun Library	67.0	58.4

 

4.2.2 Compliance with targets

Another aspect of the project is to analyse the Blessington Basin dataset and those available for any other sites to establish the degree of compliance with the target ‘quiet’ area noise levels, < 55 dBA during the day and < 45 dBA during the night. Using SPSS, the yearlong datasets were manipulated to get the number of daytime/night time readings overall and the number of readings when the target levels were exceeded. Using percentages and these values, the amount of time for which the targets are being exceeded for each site was established.

Sample Calculation: Daytime at Blessington Basin

%Exceeding level= #Readings which exceed targetTotal # Readings*100%

%Exceeding level= 1688956717*100%=29.78%

∴Degree of Compliance=70.22%

Table 4-3: Compliance levels at Blessington Street Basin

Blessington Basin	Daytime levels	Night time levels
# Total Readings	56717	37458
# Readings equal to or exceeding the target level	16889	26959
% Exceeding level	29.78	71.97
Degree of compliance with target levels	70.2%	28.0%

Table 4-4: Compliance Levels at all sites

Site	Degree of compliance with Daytime target levels	Degree of compliance with Night time levels
Blessington Basin	70.2%	28.0%
Bull Island	51.7%	29.2%
Woodstock Gardens	91.7%	57.8%

Therefore the degree of compliance with the target noise levels for daytime is 70.2% at Blessington Street Basin, while the degree of compliance with the target noise levels for night time is lower at 28%. At Bull Island, the level of compliance with night time target levels is quite low at 29.2%. The daytime levels comply better with the target at 51.7%. Although not a designated “Quiet area”, the noise levels at the site Woodstock Gardens have also been calculated and this particular site has an extremely high level of compliance with the daytime target levels of almost 92% and a level of compliance of 57.8% for the night time target levels.