Proceedings of the Institution of Civil Engineers -

Urban Design and Planning

ISSN 1755-0793 | E-ISSN 1755-0807
Volume 173 Issue 2, April, 2020, pp. 62-73


Open access content Subscribed content Free content Trial content

Full Text

This paper presents a study on individual exposure of cyclists in Cork, Republic of Ireland, to air and sound pollution, along with vibration from cycling using relatively simple sensors. Cork is currently experiencing an increasing interest in cycling, and different types of cycling transects were covered in this regard with varied infrastructure and exposure conditions for air pollution, sound and vibration. Audio-visual records were maintained throughout this cycling campaign and to connect exposure observations with available infrastructure and traffic conditions. The study highlights how the variations of exposure conditions can be obtained with relatively easy instrumentation and how such observations can be related to decisions around cycling infrastructure even when the measurements are not of the highest standard and/or are sometimes missing. The method, observations and conclusions are relevant for understanding how low-cost, lower-quality sensors can still be relevant for assessing exposure conditions for cyclists.

Cork is the second-largest city in the Republic of Ireland. Over time, it has experienced higher road traffic combined with increased air and noise pollution levels. In recent years, the Republic of Ireland has become a prominent location where cycling is evolving as a preferred mode of transport, and there is strong support from the Government through various schemes, infrastructure and policies to support the increase in the mode share of cycling in Ireland's mixed-mode network. While this increase was initially Dublin-centric, other major cities like Cork, Galway and Limerick (1%, 2% and 1% mode share, respectively) are starting to focus on cycling as an important commuting mode through investment in cycling infrastructure despite low modal share as per 2011 data (Caulfield, 2014). For example, Cork city has started constructing dedicated bicycle lanes in the city centre and has bicycle hire facilities available. While cycling has its obvious health (Wegman et al., 2012) and environmental benefits (de Nazelle et al., 2010) at a societal level, there are several challenges and human factors (Deegan and Parkin, 2011) that can deter an individual from choosing cycling as a preferred mode of transport. The challenges and human factors include perceived (Lawson et al., 2013, 2015) and real safety issues on shared or segregated cycle paths and exposure to pollution levels. It has been recently shown that negative effects from pollution and other risks like collision can lead to negative benefits for individuals (Doorley et al., 2015, 2017) even when the societal benefits are positive (Rojas-Rueda et al., 2011). There is unresolved perceived safety for cycling, but overall cycling remains an extremely safe activity with substantial health benefits.

To better understand the individual benefits, it is important to obtain exposure data for individuals for new locations and especially for cities where the number of cyclists is on the rise or such a rise has been planned. Combined exposure data for cyclists are difficult to obtain or usually not present. New campaigns in different cities suffer from the lack of high-quality and sensitive instruments, along with the problem of many such instruments being static at a single location. However, recent studies have shown that off-the-shelf instruments with lower resolution can also be effective in providing important information around exposure conditions in a city (Kumar et al., 2015). While ‘safety in numbers’ (Jacobsen, 2003) is important for cycling as a preferred mode, it can only be effectively achieved when the provision of infrastructure (Deegan, 2016), policies and decisions are taken by considering the exposure of individual cyclists, along with societal benefits.

Campaigns in cycling-friendly cities, investigating combined exposure conditions can be extremely beneficial in this regard, despite limitations around sensors and quality of data, as compared to gold standards. There is a need for new demonstrative studies in small-to-medium-sized cities in terms of individual exposures of cyclists using methodologies that are practical, easy to implement and cognizant of limitations associated with interacting with public infrastructure.

While it is acknowledged that cyclists are exposed to air pollution, studies do show that the health benefits almost always outweigh the pollution effects (de Hartog et al., 2010) and are higher than that experienced in a car (Karanasiou et al., 2014; Rank et al., 2001). Minute ventilation, defined as the amount of air a person moves in 1 min, is 4 times higher while cycling than that of someone driving a car (Panis et al., 2010), and thus exposure estimates are sometimes more relevant when understanding health effects than concentrations alone. Simultaneous exposure to noise and pollution (Apparicio et al., 2016) has also led to the observation of low negative correlation (R2 = −0.07, p = 0.005) between the two exposures in Montreal, Canada. A recent systematic review of over 4000 studies found that (Cepeda et al., 2017) benefits of active commuting from physical activity are larger than the risk from an increased inhaled dose of fine particles. In fact, an earlier study (Jungnickel and Aldred, 2014) considers cycling as a sensory practice, and their exposure to the urban environment is mediated by their sensory strategies related to relaxation, motivation and location.

Despite these extensive studies, there is still a public debate around perceived environmental exposures influencing policy decisions in various cities around cycling. Quantification of such exposures in an urban environment and a comparison of such exposure conditions at various commuting areas is thus important. Simultaneous measurements (even with limitations of such measurements) of environmental exposures of cycling can lead to evidence bases that are not available for such cities, especially those with potential for growth in cycling. These evidence bases can positively influence the increased mode share of cycling over time.

This paper directly addresses the aspect of the ability to measure exposure variations using simple sensors with limited abilities and takes up Cork, the second-largest city in Ireland, as a case study. The study measures air pollution, noise and vibration exposure conditions for cyclists in the city centre and the university area and explores how such simple sensors can still lead to relevant information about relative exposure conditions. Such an approach can provide a first estimate of exposure conditions of cyclists and inform policy instruments. It can also shed light on the causes behind variations of such exposures, leading to future decisions on cycling patterns and the development of infrastructure. Off-the-shelf, simple sensors, despite their limitations, will continue to see significant growth in the future, and approaches similar to this paper can inform the design and planning of urban mobility patterns.

Air pollution, whole-body vibration (WBV) due to cyclist–road interaction and noise pollution are some of the key exposure conditions that affect individual cyclists, apart from collision risks, which cannot be measured directly.

2.1 Air pollution

Air pollution is related to pulmonary and cardiovascular diseases and associated with markers and contributors like particulate matter (PM) PM10, PM2.5, volatile organic compounds (VOCs), nitrogen oxides (NOx) and carbon monoxide (CO). An extensive body of literature on this topic exists (Cepeda et al., 2017; Karanasiou et al., 2014). A cyclist could see a 17% reduction in adsorption of pollutants by cycling faster, and a pedestrian could see reductions of up to 26% by walking faster (McNabola et al., 2007). Traffic-induced air pollution can be an important factor around morbidity and mortality estimates globally (Hoek et al., 2000; Tsai et al., 2010). World Health Organisation (WHO, 2011) provides guidance for allowable levels of pollution in air for informed policies and decision making, and this paper has referred to such guidance where necessary. Boogaard et al. (2009) studied 11 cities in the Netherlands and collected extensive particulate matter and PM2.5 samples in real time, correlating them eventually with noise and was one of the first studies to carry out such co-located tests, while de Nazelle et al. (2012) emphasise the need for location-specific measurements following their air pollution exposure studies in Barcelona. The comprehensive review by Oja et al. (2011) has also indicated the need for more evidence to link exposure and benefits, as activity overcomes any negative effect of air pollution (Woodward and Samet, 2016). There is a need to measure exposure conditions in several locations in an urban area, and it is also observed that while there have been extensive studies on the measurement of particulate matter, the variation of other gases in different urban areas have not been subject to the same scrutiny or detail. Partially, this has been due to the static and expensive nature of high-quality instruments and the lack of limited capability but cheaper mobile sensors for gases. Consequently, investigating such simple sensors with limited capabilities have the potential to distinguish variations of such gases in different urban locations, leading to effective decisions on cycling, estimates of health benefits (Doorley et al., 2019a) and even urban cycling network design (Doorley et al., 2020).

2.2 Noise pollution

Traffic noise pollution is a long-term issue in the European Union (Berglund et al., 1999). The noise between vehicle tyre and the road surface can exceed engine noise at speeds greater than 50 km/h for passenger cars and 80 km/h for lorries (Muzet, 2007). In Cork, the actual speed in the city centre is typically as low as 30 km/h but can increase to 60–100 km/h (100 km/h for national roads and 80 km/h in regional roads) zone just outside the city but within commuting distance, depending on the time of the day. For the commute to and from work or university, the speed is on the lower side of this range due to higher density of traffic, but there is an increase in noise caused by braking, turning and acceleration at junctions (Kinkaid et al., 2003), along with the combination of noise from the movement of several cars. Similarly, wind noise over the ears is often loud at higher cycling speeds and can block external noise. The effects are both physiological and psychological as evidenced by sleep disturbance studies (Lin et al., 2011), annoyance levels (Bluhm et al., 2004; Rylander, 2004) and cardiovascular effects (Babisch, 2008; Babisch et al., 2005; Barregard et al., 2009; Bluhm et al., 2007; Jarup et al., 2008; Kaltenbach et al., 2008). A review (Babisch, 2008) of two cross-sectional and five analytical studies find an increase in cardiovascular risk over 60 dB(A), showing a dose–response relationship. External costs of transport noise have been estimated to be 7.45–23.81 Swiss francs for per night of noise-free sleep (Riethmuller et al., 2008), while the willingness to pay can be as high as €10 per 1 dB(A) improvement per person, per year above 43 dB(A) leading to annual costs of traffic noise in Germany to be 7.8–9.6 billion Euro (University of Texas School of Public Health). A review article of 64 studies on the valuation of noise conducted by the United Kingdom Department of Transportation in 1999 estimated this value to be (University of Texas School of Public Health) £15–£30 per decibel (dB) per household per year, 0.02–2.27% GDP and 0.08–2.30% change in property value per dB. In the Republic of Ireland (HSA, 2007) the daily noise exposure level is considered to be the average exposure level over an 8h day and is expressed as LEX, 8 h dB(A), the peak sound pressure is the maximum value of the noise pressure and is expressed as ppeak dB(C) and it is recommended that if it is difficult to hear a normal conversation at a distance of 2m from the person speaking, it is likely that the noise level in the area is above 80 dB(A). Limits to these values, beyond which remedial action should be taken relate to lower exposure action values of LEX,8h = 80 dB(A) and ppeak = 135 dB(C) while the upper exposure action values are LEX,8h = 85 dB(A) and ppeak = 137 dB(C). The maximum limit of daily noise exposure or peak sound pressure corresponds to LEX,8 h = 87 dB(A), and ppeak = 140 dB(C). These values form the context and background for this paper.

2.3 Whole-body vibration

WBV have detrimental effects on the human body, including discomfort, hand/arm numbness, white fingers, and trapped nerves and lower back pain (Cardinale and Pope, 2003). The ISO 2631-1 (ISO, 1997) standard for evaluation of human exposure to WBV looks at WBV in terms of health, comfort, perception and motion sickness. It categorises three zones based on the weighted acceleration (aw) of the vibration sustained and the duration of exposure (T) to the vibration related to running root mean square (rms) and fourth power vibration dosage, respectively. Zone 1 is an area where resulting health problems have not been documented, zone 2 relates to a risk of a resulting health problem and zone 3 is the area which signifies a region for which documented health problems have occurred for this level of WBV (ISO 2631-1 (ISO, 1997) Annex B). Ranges of weighted accelerations in zone 1 also represent levels of comfort from ‘not uncomfortable’ to ‘extremely uncomfortable’ (ISO 2631-1 (ISO, 1997) Annex C).

3.1 Choice of transects

The exposure of cyclists to air pollution, vibration from varying road conditions, and noise were measured in Cork for transects often frequented by university students (Figure 1). Transects 1 and 2 are typical for several students in terms of their daily commute and correspond to transects considered in a previous study where physiological and psychological responses were linked for cyclists (Doorley et al., 2015, 2019b). This work concluded that there is a correlation between the physiological responses (interbeat intervals of heart rate) and psychological ratings of perceived risk, and such correlation can also lead to errors around estimated calorie count of fitness apps. A more recent study in this direction (Mac Hale et al., 2019) indicated that while passing events of vehicles lead to a physiological change in a cyclist in terms of heartbeat, this does not necessarily relate to a higher perceived risk rating by them. The zone of the study in this paper covers the major roads surrounding the university and link up with major university accommodations. The transects were chosen for comparison with this previous study and to investigate the idea of a movable sensor-based map for the future of the city, as has been investigated by the authors with mobile sensors for looped transects in Dublin (Doorley et al., 2016). The transects are showed on a map in Figure 1.

figure parent remove

Figure 1. University College Cork and Cork city transects for experiment

The Environmental Protection Agency (EPA) Ireland air pollution monitoring station is located at the city hall on the opposite side of the city as compared to the university and so proximity of this station could be used to check significant anomalies, especially when simple sensors are used as the one in this study. Transects 4 and 5, in conjunction with the previous ones, link the university to Cork bus station and would be travelled regularly by student cyclists. Table 1 provides a summary of the transects. The sampling location exposures are important here since the relative variation of exposure conditions are mostly independent of the cyclist. This is an advantage since significant data can be gathered without the need of a control population, which is not possible for other studies like risk perception. Traffic data for this localised choice of transects and granularity in time is not available for Cork, and this must be considered as a limitation in this study.


Table 1. Attributes and summary of chosen transects for the experiment

Table 1. Attributes and summary of chosen transects for the experiment

Transect attribute Transect 1 Transect 2 Transect 3 Transect 4 Transect 5
Length: km 1.48 2.17 1.77 2.95 1.93 km
Proportion of transect with full-cycle lane 40% 27% 9%
Proportion of transect with bus lane 13% 40% 11%
Number of bus stops 3 7 4 6 6
Number of traffic lights without advance stop lines for cyclists 2 4 4 14 10
Number of traffic lights with advance stop lines for cyclists 4 4 3 4 2
3.2 Data collection

A modern racing hybrid was used to conduct this experiment. This bicycle has narrow tyres with little suspension, allowing us to record the worst-case scenario of WBV experienced by the cyclist as a result of the road surface quality. The tyre pressure of the cycles, while pumped to qualitative adequacy, was not measured and must be considered as a limitation in this study. The accelerometer was mounted to the frame in order to ensure that the measured data represents true data (which is experienced by the cyclist) as opposed on the cyclist (where the vibration is filtered through the body). Additionally, the accelerometer on the frame allows the device to remain level in relation to the bicycle and cyclist for the duration of the testing. Video footage was recorded from a bicycle-mounted camera to match events or markers from other instruments with actual events on the road. Any significant variations can also be further investigated using the camera footage. Similarly, riders do not feel the shocks from rough roads directly but are mitigated by a range of factors such as choice of saddle and handgrips, but more notably, by standing on the pedals and relaxing the wrists, as experienced riders do over holes and roughness.

Noise exposure data were collected at 8000 Hz as LAeq and peak dB using the ‘Noise Meter’ app on a Google Nexus 4 smartphone. The accuracy of such smartphone apps may vary, and a recent study indicates that despite such variation, some of the apps can be useful (Kardous and Shaw, 2014). The maxima, minima and an average value, along with the data, was emailed as a text file to an external email directly from the device. The smartphone was mounted in a compatible armband on the cyclist's right arm with a microphone exposed to the cyclist's environment, keeping the device on the traffic side of the cyclist.

A Libellium Waspmote accelerometer, measuring data at 50 Hz in three perpendicular directions, was mounted to the bicycle frame on a basket over the rear wheel. Accelerations were measured at 50 Hz and saved to a micro SD card. The recorded accelerations were weighted by getting a root-mean-squared (RMS) value for each second of the data.

Levels of oxygen (O2), carbon dioxide (CO2) and nitrogen dioxide (NO2) were measured along various transects using a portable, compact Libellium Waspmote Plug and Sense waterproof device and were stored on an internal micro SD card. The sampling rate for air pollution data is significantly slower compared to that of the acceleration data. Additionally, there is an over-heating tendency of the gas probes observed during the experimental study, which requires the sensor to be shut down for a cooling period between each reading. Due to this limitation, air samples were taken at fixed locations at approximately 100 m intervals along the cycle transects to obtain a better portrayal of the air qualities measured at various locations to assess exposure conditions of cyclists. The levels of each of the three gases analysed were taken at a total of 106 locations across the five designated transects (Figure 2). Data were collected on weekdays during peak traffic conditions of 17 : 00 and 19 : 00. The height of the pollution monitor was kept constant at 0.85 m (securely fastened to the carrier), the distance from the kerb for cycle lanes was generally kept at 0.25–0.5 m, while for mixed-mode traffic there was an attempt to keep it at 0.5 m where possible, except for turning. Figure 3 presents the test bicycle with a related data collection set-up.

figure parent remove

Figure 2. Gas sampling locations in Cork city on each of the five chosen transects (source: from GoogleMaps)

figure parent remove

Figure 3. Racing hybrid bicycle used for testing along with the bicycle-mounted camera, phone, gas sensors and accelerometer locations indicated

3.3 Calibration aspects
The oxygen probe has a linear relationship expressed as
is the voltage recorded by the device and
is the oxygen level as a percentage of air. The nitrogen dioxide probe calibration relates voltage outputs for 0.053 and 0.002 ppm of nitrogen dioxide as 1.267 and 0.101 V, respectively, while for carbon dioxide, the voltage outputs for 350 and 0.002 ppm were 1.158 and 3.3 V, respectively. Values in-between are linearly interpolated.
4.1 Air pollution
4.1.1 Carbon dioxide

The results of relative carbon dioxide exposure for the cyclists on different transects (Figure 4) are presented with an accepted average level of carbon dioxide as 411.3 ppm, which is the average level of carbon dioxide in the atmosphere in 2019 as per UK Met Office (2019). Transect 1 along the perimeter of the main campus of the university shows relatively low levels of carbon dioxide (338 ppm), and the maximum value of 353 ppm corresponds to a busy traffic light junction. Transect 2 near student accommodations has lower levels (lowest 328 ppm) of carbon dioxide and the transect is further away from the city with overall lower traffic levels, although the maximum value is the same (at a busy traffic light) as that observed in transect 1. A comparative count of traffic could have been helpful for comparison but was not available and is highlighted as a limitation. Transect 3 linking the university to the city centre through congested roads registered a minimum value of 335 ppm and a maximum of 354 ppm, while transect 4 with significantly heavier traffic does not record carbon dioxide levels below 350 ppm at any point. The overall air quality in Ireland is good, but this recorded value of carbon dioxide is significantly lower than the global levels and should also be interpreted based on the lower fidelity of measurement devices. The relative variation of such concentration levels remains relevant with such limited measurement options. The lowest level recorded on transect 4 was 359 ppm, with regular increase of 377 ppm. Transect 5 results are similar to transect 4 areas as it is in a congested area as well, and the carbon dioxide concentration recorded was 373 ppm. The concentration of 411.3 ppm was not exceeded anywhere.

figure parent remove

Figure 4. Carbon dioxide exposure levels for cyclists on different transects in Cork

4.1.2 Nitrogen dioxide

The average safe level of nitrogen dioxide in the air is taken here as 0.053 ppm. To obtain reasonable and stable results, a 99 s interval was used for measurement, which allowed sufficient time for the gas board to be shut off and for the nitrogen dioxide probe to cool. This increased the accuracy of the data, although the probes occasionally overheated and thus had to be monitored. Therefore, instead of representing the average safe value by a horizontal line at 0.053 ppm, the error of the sensor probe was incorporated by plotting the curve of the values recorded by the nitrogen dioxide sensor for the same duration as the field tests in a controlled environment. Under such circumstances, the relative changes between transects are still possible to be compared (Figure 5). Transect 1 consistently records low concentrations of nitrogen dioxide. This is consistent with the conclusions drawn from the analysis of the carbon dioxide data for this transect. Transect 2 is similar to transect 1, with levels recorded between 0.01 and 0.015 ppm below levels recorded in the control test. After approximately 1400 s, the highest quality of air is recorded on this transect, at a location near the student accommodation where there is no traffic junction. Transect 3 recorded levels consistently at approximately 0.02 ppm below the control test. The three spikes in the data correspond to traffic lights and congested traffic. Transect 4 recorded values approximately 0.01 ppm below the control test, while the peak corresponds to a traffic junction. Transect 5 recorded the maximum values at approximately 0.01 ppm greater than the control test at one point, corresponding to two extremely busy bus stops with heavy traffic. There are several buses in that location often for an extended length of time, with their engines running to align with their timetables.

figure parent remove

Figure 5. Nitrogen dioxide exposure levels for cyclists on different transects in Cork

4.1.3 Oxygen

The accepted level of oxygen is taken as 20.95% in this study and is lower in built-up urban areas due to higher concentrations of carbon dioxide and NOx from motor vehicles and lack of vegetation. Levels of oxygen recorded in Cork (Figure 6) are lower than expected. This can be partially attributed to the quality of the oxygen probe, leading to some underestimation but also indicative of a lower level of oxygen. Additionally, the probe was observed to systematically stabilise only after a certain amount of distance was traversed. However, the variation of measured oxygen levels of different transects provides insight into the changing air quality, and thus the variations are explored in this study. Transects 1 and 2 have similar oxygen levels, which increase away from the city due to the better quality of air. Towards the end of the transect, lower levels of oxygen are recorded due to constant westbound traffic that occurs on this part during the evening rush hour. Transect 3 values are similar to that of transect 2 but with lower levels of oxygen due to higher traffic density and less vegetation in the city centre. The exact cause of this can be debated, but it is observed that there is a narrowing of this stretch of road from two lanes to one to accommodate the contraflow bicycle lane, leading to an increase in traffic density levels. Transect 4 has the lowest concentration of oxygen due to its location in the city centre. The lowest values correspond to a busy junction with traffic lights and a large car park with a significant rush during the evening peak. Transect 5 also has low concentrations of oxygen due to its location in the city centre. The slightly better quality of air here as compared to transect 4 is due to the fact that the vast majority of this transect is located adjacent to the River Lee, and this helps increase the concentration of oxygen. The lowest values recorded here correspond to a very busy bus stop serving 13 different bus services and for busy junctions with traffic lights.

figure parent remove

Figure 6. Relative variation of underestimated oxygen levels for all transects recorded at 100 m intervals

These results demonstrate that significant information can be obtained for variation of exposure conditions, even without the support of very accurately calibrated devices. Further studies on air pollutants, including PM2.5, PM10, benzene, NOx and other markers, can be particularly helpful in assessing the condition in a city.

4.2 Road quality and vibrations

The WBV data on the cycled transects are presented (Figure 7), and the weighted accelerations are compared and categorised under three major zones identified in ISO 2631-1 (ISO, 1997) annex B. Weighted accelerations in zone 1 are assessed under the ranges of comfort given in ISO 2631-1 (ISO, 1997) annex C. The ranges are for instantaneous recordings of the weighted acceleration, and they will decrease as the duration of exposure increases. However, for this exposure study, the focus will be on the instantaneous level as a weighted acceleration value for each second of the sampled transects.

figure parent remove

Figure 7. Instantaneous vibration level estimates for each transect cycled

Transect 1 shows a mix of road surface conditions. Initially, the road quality is significantly worse than the latter part. The first 100 s show multiple peaks between 4.6 and 6.8 m/s2, and the surface can be identified as poor quality by visual inspection. The stationary periods correspond to traffic lights, as observed in previous sections, and are related to higher pollution levels, although this is probably slightly reduced as cyclists prefer to filter themselves to the front of such a vehicular queue. This is applicable to all subsequent transects as well. In this study, the cyclists did not filter themselves to the front, and consequently, the measurement reflects the effects of being at the rear of car exhausts. Directly after the traffic junction, the 8 m/s2 is related to a temporary obstruction for road works in operation. The remaining 3 min of transect 1 shows weighted accelerations mostly between 1 and 3 m/s2 corresponding to the good road surface. For transect 2, high accelerations can be observed after 150 s at a roundabout. Between 260 and 320 s, 11.3 m/s2 accelerations correspond to road surface visually in need of improvement, following which the surface is of an acceptable level. Transect 3 has overall good road conditions with the brief exceptions of localised road conditions relating to 6.5 and 7.5 m/s2 accelerations recorded. Transect 4 is the longest, covering the city centre but the acceleration response is good as these roads are better maintained. The 11.1 m/s2 recorded is due to a localised bump. On the contrary, transect 5 is poorer in terms of road quality and vibrations. Beyond the initial 2 min on a new two-way bicycle lane, there is an 80 s period of the significantly poor road surface, which eventually improves later to a certain extent. Analysing overall acceleration data indicated that zones 1, 2 and 3 for WBV corresponded with 80, 18.5 and 1.5% of the transects, respectively. Since most of the roads fall in zone 1, this is further sub-divided (Figure 8), and for the most part, the level of comfort experienced by a cyclist was ‘very uncomfortable’. This stated experience was related to the response to a questionnaire filled up by all cyclists, where the comfort levels, along with other detailed questions around culpability (Byrne et al., 2017), were recorded. With a sharp increase in integration of smartphone-based accelerometers in recent times, the results also indicate a need to carry out a comparative study of the performance of various phones and mobile applications (Cahill et al., 2019), which could be indicative of whether such mobile data can also be used for assessing the ride quality and map extensive urban areas for road surface maintenance.

figure parent remove

Figure 8. Ranges of comfort experienced by cyclists on roads of zone 1

4.3 Noise pollution

In this study, 80 dB(A) is kept as the maximum acceptable level of noise. Figure 9 presents the noise levels measured on all transects with transect 2 experiencing maximum exposure (74.9 dB) related to high levels of commuter traffic during rush hours, with the noise level for this entire transect remaining between 72.9 and 74.9 dB. Noise pollution is thus less of an issue in Cork. When these noise levels were compared to the relationship between annoyance and noise level, it was found that for the duration of the transect, the cyclist was subjected to an annoyance level of 30%. Noise pollution in Cork, as expected, is greater in the city centre than it is further out.

figure parent remove

Figure 9. Variation of noise exposure levels on different transects cycled

This paper presents Cork as a case study for estimating environmental exposure variations for cyclists using simple sensors with limited capabilities. It is observed that despite limitations in measurement, problems around calibration, or design of experiments, useful information can be obtained easily, which can subsequently inform urban planning and design for cycling.

For Cork, higher levels of air pollutions are observed at or just before busy traffic junctions related to motor vehicle emissions in their static condition and in denser parts of the city. Analysing the vibration data in conjunction with the gas levels, the duration of exposure of the cyclist to higher levels of pollutants could be identified. An advance start for cyclists at junctions may help this situation to some extent, along with safety benefits. The transects tested as part of this study contain a total of 47 sets of traffic lights, 17 of which contain an advance start, which allows a cyclist to remain at the top of the queue for the duration of a traffic light cycle. A cycle lane spanning the length of a typical tailback for a junction can also help in terms of planning. There are several junctions in Cork where an advance start is provided, but due to the narrow nature of the road leading up to the junction, the participating cyclists were sometimes concerned about the possibility of safely passing the queue of vehicles on a bicycle to gain access to the advance start during peak traffic conditions.

In terms of vibration responses, while a high percentage of the roads were of good quality, more than 20% of the road surfaces tested were not adequate. Only around 1.5% of the roads tested led to significant vibration responses. The selection of improvement of road surfaces can be chosen based on these results since they are more scattered around all transects. Road surfaces were generally of higher quality in the city centre than towards the outskirts. Lowest vibrations were recorded on the newly laid cycle lanes.

Cyclists in Cork are not exposed to significant levels of noise pollution. The highest level of noise recorded was of the magnitude of 74.9 dB, related to a 30% annoyance level. The highest level of noise pollution was related to high volumes of commuter traffic in the morning rush hour with speeds faster than that typically encountered in the city centre. The lowest levels of noise pollution were related to low vehicular traffic.

Overall, the combined exposure of cyclist to various levels of pollution are not very high for a range of selected transects qualitatively representing a range of traffic, cycling infrastructure and pollution conditions. Such an observation should encourage the authorities to provide and promote cycling further and to extend this type of database over an extended length of time and region.

The study is also relevant for cities currently experiencing or expecting an increase in cycling. It can encourage the owners of road and cycling infrastructure to make better-informed decisions regarding the impact of cycling on their networks using simple campaigns similar to the one presented, even if there are limitations around such measurements.


weighted acceleration


voltage recorded by the device


oxygen level as a percentage of air


duration of exposure


Cited By

Related content

Related search

By Keyword
By Author

No search history

Recently Viewed