Jean Marc Wunderli

Intermittency ratio: A metric reflecting short-term temporal variations of transportation noise exposure

Most environmental epidemiology studies model health effects of noise by regressing on acoustic exposure metrics that are based on the concept of average energetic dose over longer time periods (i.e. the Leq and related measures). Regarding noise effects on health and wellbeing, average measures often cannot satisfactorily predict annoyance and somatic health effects of noise, particularly sleep disturbances. It has been hypothesized that effects of noise can be better explained when also considering the variation of the level over time and the frequency distribution of event-related acoustic measures, such as for example, the maximum sound pressure level. However, it is unclear how this is best parametrized in a metric that is not correlated with the Leq, but takes into account the frequency distribution of events and their emergence from background. In this paper, a calculation method is presented that produces a metric which reflects the intermittency of road, rail and aircraft noise exposure situations. The metric termed intermittency ratio (IR) expresses the proportion of the acoustical energy contribution in the total energetic dose that is created by individual noise events above a certain threshold. To calculate the metric, it is shown how to estimate the distribution of maximum pass-by levels from information on geometry (distance and angle), traffic flow (number and speed) and single-event pass-by levels per vehicle category. On the basis of noise maps that simultaneously visualize Leq, as well as IR, the differences of both metrics are discussed.

More than clean air and tranquillity: Residential green is independently associated with decreasing mortality

Green space may improve health by enabling physical activity and recovery from stress or by decreased pollution levels. We investigated the association between residential green (greenness or green space) and mortality in adults using the Swiss National Cohort (SNC) by mutually considering air pollution and transportation noise exposure. To reflect residential green at the address level, two different metrics were derived: normalised difference vegetation index (NDVI) for greenness, and high resolution land use classification data to identify green spaces (LU-green). We used stratified Cox proportional hazard models (stratified by sex) to study the association between exposure and all natural cause mortality, respiratory and cardiovascular disease (CVD), including ischemic heart disease, stroke and hypertension related mortality. Models were adjusted for civil status, job position, education, neighbourhood socio-economic position (SEP), geographic region, area type, altitude, air pollution (PM10), and transportation noise. From the nation-wide SNC, 4.2 million adults were included providing 7.8years of follow-up and respectively 363,553, 85,314 and 232,322 natural cause, respiratory and CVD deaths. Hazard ratios (and 95%-confidence intervals) for NDVI [and LU-green] per interquartile range within 500m of residence were highly comparable: 0.94 (0.93-0.95) [0.94 (0.93-0.95)] for natural causes; 0.92 (0.91-0.94) [0.92 (0.90-0.95)] for respiratory; and 0.95 (0.94-0.96) [0.96 (0.95-0.98)] for CVD mortality. Protective effects were stronger in younger individuals and in women and, for most outcomes, in urban (vs. rural) and in the highest (vs. lowest) SEP quartile. Estimates remained virtually unchanged after incremental adjustment for air pollution and transportation noise, and mediation by these environmental factors was found to be small. We found consistent evidence that residential green reduced the risk of mortality independently from other environmental exposures. This suggests the protective effect goes beyond the absence of pollution sources. Environmental public health measures should not only aim at reducing pollutant exposure, but additionally maintain existing and increase residential green in areas where lacking.

A Model to Predict Sound Reflections from Cliffs

In the context of the Swiss railway noise calculation model sonRAlL a model to predict sound exposure level spectra for reflections from cliffs was developed and validated with measurements. It is an engineering model with low computational cost that is applicable to any type of sound sources. The scattered energy reflected from the cliffs is modelled by the superposition of secondary sources placed on the cliff surface. The only free parameter of the model, the scattering coefficient, was determined by measurements with detonations of 200 g TNT e q in the alpine part of Switzerland. Measurements yielded good agreement not only in situations with reflecting cliffs but also in the presence of large boulder fields.

Differences between Outdoor and Indoor Sound Levels for Open, Tilted, and Closed Windows

Noise exposure prediction models for health effect studies normally estimate free field exposure levels outside. However, to assess the noise exposure inside dwellings, an estimate of indoor sound levels is necessary. To date, little field data is available about the difference between indoor and outdoor noise levels and factors affecting the damping of outside noise. This is a major cause of uncertainty in indoor noise exposure prediction and may lead to exposure misclassification in health assessments. This study aims to determine sound level differences between the indoors and the outdoors for different window positions and how this sound damping is related to building characteristics. For this purpose, measurements were carried out at home in a sample of 102 Swiss residents exposed to road traffic noise. Sound pressure level recordings were performed outdoors and indoors, in the living room and in the bedroom. Three scenarios—of open, tilted, and closed windows—were recorded for three minutes each. For each situation, data on additional parameters such as the orientation towards the source, floor, and room, as well as sound insulation characteristics were collected. On that basis, linear regression models were established. The median outdoor–indoor sound level differences were of 10 dB(A) for open, 16 dB(A) for tilted, and 28 dB(A) for closed windows. For open and tilted windows, the most relevant parameters affecting the outdoor–indoor differences were the position of the window, the type and volume of the room, and the age of the building. For closed windows, the relevant parameters were the sound level outside, the material of the window frame, the existence of window gaskets, and the number of windows.

sonRAIL – The New Swiss Calculation Model for Railway Noise

The new Swiss calculation model for railway noise called sonRAIL is presented. The sound propagation model as well as the emission model is designed to simulate fundamental physical processes. Therefore, they allow a reproduction of current situations and prediction of the effects of different mitigation measures. The model precisely calculates emission levels, taking into account all relevant variables. The emission levels are determined as sound power spectra for five source heights. The rolling noise is calculated with wheel and rail roughness spectra. The influence of different superstructure types is taken into account. The propagation model considers the direct sound propagation, meteorological effects, and reflections at buildings, walls, and other rigid surfaces, and diffuse reflections at forest edges and cliffs are described. As a first broad application of the model, the results of the sound exposure along the Gotthard railway line in Switzerland are shown.

Diurnal variability of transportation noise exposure and cardiovascular mortality : a nationwide cohort study from Switzerland

BACKGROUND Most epidemiological noise studies consider 24 h average noise exposure levels. Our aim was to exploratively analyze the impact of noise exposure at different time windows during day and night on cardiovascular mortality. METHODS We generated Switzerland-wide exposure models for road traffic, railway and aircraft noise for different time windows for the year 2001. Combined noise source equivalent continuous sound levels (Leq) for different time windows at the most exposed façade were assigned to each of the 4.41 million Swiss National Cohort adult participants. Follow-up period was from 2000 to 2008. Hazard ratios (HR) of noise effects on various cardiovascular primary causes of death were computed by Cox regression models adjusted for potential confounders and NO2 levels. RESULTS For most cardiovascular causes of death we obtained indications for a diurnal pattern. For ischemic heart disease the highest HR was observed for the core night hours from 01 h to 05 h (HR per standard deviation of Leq: 1.025, 95% CI: 1.016-1.034) and lower HR for the daytime 07 h to 19 h (1.018 [1.009-1.028]). Heart failure and daytime Leq yielded the highest HR (1.047 [1.027-1.068]). CONCLUSION For acute cardiovascular diseases, nocturnal intermittent noise exposure tended to be more relevant than daytime exposure, whereas it was the opposite for chronic conditions such as heart failure most strongly associated with continuous daytime noise. This suggests that for acute diseases sleep is an important mediator for health consequences of transportation noise.

Sleep spindle characteristics and arousability from nighttime transportation noise exposure in healthy young and older individuals

Study Objectives Nighttime transportation noise elicits awakenings, sleep-stage changes, and electroencephalographic (EEG) arousals. Here, we investigated the potential sleep-protective role of sleep spindles on noise-induced sleep alterations. Methods Twenty-six young (19-33 years, 12 women) and 18 older (52-70 years, 9 women) healthy volunteers underwent a repeated measures polysomnographic 6-day laboratory study. Participants spent one noise-free baseline night, followed by four transportation noise-exposure nights (road traffic or railway noise; continuous or intermittent: average sound levels of 45 dB, maximum sound levels of 50-62 dB), and one noise-free recovery night. Sleep stages were scored manually and fast sleep spindle characteristics were quantified automatically using an individual band-pass filtering approach. Results Nighttime exposure to transportation noise significantly increased sleep EEG arousal indices. Sleep structure and continuity were not differentially affected by noise exposure in individuals with a low versus a high spindle rate. Spindle rates showed an age-related decline along with more noise-induced sleep alterations. All-night spindle rates did not predict EEG arousal or awakening probability from single railway noise events. Spindle characteristics were affected in noise-exposure nights compared to noise-free nights: we observed a reduction of the spindle amplitude in both age groups and of the spindle rate in the older group. Conclusions We have evidence that spindle rate is more likely to represent a trait phenomenon, which does not seem to play a sleep-protective role in nighttime transportation noise-induced sleep disruptions. However, the marked reduction in spindle amplitude is most likely a sensitive index for noise-induced sleep alterations.

OP XI – 5 The proportion of high noise sensitivity and annoyance in the hermes cohort study of swiss adolescents

Background/aim The World Health Organization considers children to be of particular risk of suffering negative cognitive and health consequences through exposure to transportation noise. However, little is known about how adolescents experience exposure to noise in their everyday life. The purpose of this study is to examine level of both noise annoyance and sensitivity in a cohort of adolescents. Methods Noise sensitivity and annoyance was measured in a cohort of 892 participants aged 12 to 17 with one year follow-up. Noise annoyance for five different noise sources, including transportation vehicles, industry and people, was measured using a verbal 4 point Likert scale, where the annoyance was classified as follows: none (0), little (1), clearly (2), severely (3). Participants were classified as highly annoyed if they scored 2 or higher. Noise sensitivity was measured with the Zimmer/Ellermeyer (1998) scale ranging from 0 to 27. Individuals were classified as sensitive if they had a score over 13.5. Proportion of noise annoyed and noise sensitive participants are presented. Change between baseline and follow-up proportions were compared with the McNewmar-test. Results 847 students participated at the baseline measurement of the annoyance of road noise. Follow-up participation was at 806 (participation rate: 95.1%). The proportion of highly annoyed students was 4.40% (n=39) at baseline and 2.74% (n=23) at follow-up. The change in proportion after a year was quite significant (p=0.028). The noise sensitivity questionnaire was filled out by 887 students at baseline and follow-up was at 840 (participation rate: 94.7%). The proportion of highly noise sensitive students was 24.32% (n=206) at baseline and 24.07% (n=194) at follow-up, with no indications for a difference between baseline and follow-up (p=0.834). Within all comparisons gender was not substantial, except for baseline noise annoyance of two noise sources. Conclusion Our study suggests that prevalence of high noise annoyance is low in this age group and tends to decrease during puberty. Noise sensitivity seemed to be prevalent in a quarter of adolescents and stayed constant after a year. Future analyses should explore the relationship between annoyance, sensitivity, actual noise exposure and health outcomes.

A systematic analysis of mutual effects of transportation noise and air pollution exposure on myocardial infarction mortality: a nationwide cohort study in Switzerland

AIMS The present study aimed to disentangle the risk of the three major transportation noise sources-road, railway, and aircraft traffic-and the air pollutants NO2 and PM2.5 on myocardial infarction (MI) mortality in Switzerland based on high quality/fine resolution exposure modelling. METHODS AND RESULTS We modelled long-term exposure to outdoor road traffic, railway, and aircraft noise levels, as well as NO2 and PM2.5 concentration for each address of the 4.40 million adults (>30 years) in the Swiss National Cohort (SNC). We investigated the association between transportation noise/air pollution exposure and death due to MI during the follow-up period 2000-08, by adjusting noise [Lden(Road), Lden(Railway), and Lden(Air)] estimates for NO2 and/or PM2.5 and vice versa by multipollutant Cox regression models considering potential confounders. Adjusting noise risk estimates of MI for NO2 and/or PM2.5 did not change the hazard ratios (HRs) per 10 dB increase in road traffic (without air pollution: 1.032, 95% CI: 1.014-1.051, adjusted for NO2 and PM2.5: 1.034, 95% CI: 1.014-1.055), railway traffic (1.020, 95% CI: 1.007-1.033 vs. 1.020, 95% CI: 1.007-1.033), and aircraft traffic noise (1.025, 95% CI: 1.006-1.045 vs. 1.025, 95% CI: 1.005-1.046). Conversely, noise adjusted HRs for air pollutants were lower than corresponding estimates without noise adjustment. Hazard ratio per 10 μg/m³ increase with and without noise adjustment were 1.024 (1.005-1.043) vs. 0.990 (0.965-1.016) for NO2 and 1.054 (1.013-1.093) vs. 1.019 (0.971-1.071) for PM2.5. CONCLUSION Our study suggests that transportation noise is associated with MI mortality, independent from air pollution. Air pollution studies not adequately adjusting for transportation noise exposure may overestimate the cardiovascular disease burden of air pollution.

Adverse impact of nocturnal transportation noise on glucose regulation in healthy young adults: Effect of different noise scenarios

BACKGROUND Epidemiological evidence indicates an association between transportation noise exposure and a higher risk of developing type 2 diabetes. Sleep disturbances are thought to be one of the mechanisms as it is well established that a few nights of short or poor sleep impair glucose tolerance and insulin sensitivity in healthy good sleepers. OBJECTIVES The present study aimed to determine the extent to which exposure to nocturnal transportation noise affects glucose metabolism, and whether it is related to noise-induced sleep alterations. METHODS Twenty-one young healthy volunteers (nine women) participated in a six-day laboratory study starting with a noise-free baseline night, then four nights sleeping with randomly-presented transportation noise scenarios (three road and one railway noise scenario) with identical average sound level of 45dB but differing in eventfulness and ending with a noise-free recovery night. Sleep was measured by polysomnography. Glucose tolerance and insulin sensitivity were measured after the baseline, the last noise night and the recovery nights with an oral glucose tolerance test using Matsuda and Stumvoll insulin sensitivity indexes. Eleven participants were assigned a less eventful noise scenario during the last noise night (LE-group), while the other ten had a more eventful noise scenario (ME-group). Baseline metabolic and sleep variables between the two intervention groups were compared using a non-parametric Mann-Whitney U test while mixed models were used for repeated measure analysis. RESULTS All participants had increased glucoseAUC (mean±SE, 14±2%, p<0.0001) and insulinAUC (55±10%, p<0.0001) after the last noise night compared to the baseline night. 2h-glucose level tended to increase only in the ME-group between baseline (5.1±0.22mmol·L-1) and the last noise night (6.1±0.39mmol·L-1, condition: p=0.001, interaction: p=0.08). Insulin sensitivity assessed with Matsuda and Stumvoll indexes respectively decreased by 7±8% (p=0.001) and 9±2% (p<0.0001) after four nights with transportation noise. Only participants in the LE-group showed beneficial effects of the noise-free recovery night on glucose regulation (relative change to baseline: glucoseAUC: 1±2%, p=1.0 for LE-group and 18±4%, p<0.0001 for ME-group; Stumvoll index: 3.2±2.6%, p=1.0 for LE-group and 11±2.5%, p=0.002 for ME-group). Sleep was mildly impaired with increased sleep latency of 8±2min (<0.0001) and more cortical arousals per hour of sleep (1.8±0.6arousals/h, p=0.01) during the last noise night compared to baseline. No significant associations between sleep measures and glucose tolerance and insulin sensitivity were found. CONCLUSION In line with epidemiological findings, sleeping four nights with transportation noise impaired glucose tolerance and insulin sensitivity. Based on the presented sound exposure, the eventfulness of the noise scenarios seems to play an important role for noise-induced alterations in glucose regulation. However, we could not confirm our hypothesis that transportation noise impairs glucose regulation via deterioration in sleep quality and quantity. Therefore, other factors, such as stress-related pathways, may need to be considered as potential triggers for noise-evoked glucose intolerance in future research.