Childhood sound disturbance and sleep problems in Alpine valleys with high levels of traffic exposures and greenspace

Environmental Research 242 (2024) 117642 Available online 22 November 2023 0013-9351/© 2023 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). Childhood sound disturbance and sleep problems in Alpine valleys with high levels of traffic exposures and greenspace Angel M. Dzhambov a , b , c , d , * , 1 , Peter Lercher b , 1 , Dick Botteldooren e a Environmental Health Division, Research Institute at Medical University of Plovdiv, Bulgaria b Institute of Highway Engineering and Transport Planning, Graz University of Technology, Austria c Research Group “Health and Quality of Life in a Green and Sustainable Environment”, Strategic Research and Innovation Program for the Development of MU - Plovdiv, Medical University of Plovdiv, Bulgaria d Department of Hygiene, Faculty of Public Health, Medical University of Plovdiv, Bulgaria e Department of Information Technology, Research Group WAVES, Ghent University, Belgium ARTICLE INFO Keywords: Air pollution Exposome Garden Greenness Restorative quality Traffic noise ABSTRACT Sound disturbance and sleep problems are regarded as the most common adverse effects of environmental noise but evidence of the role of air pollution and greenspace is scant. This is especially true for children who find themselves in a sensitive developmental period and experience their environment differently than adults. This study examined the joint effects of traffic exposures and residential greenspace on child sound disturbance and sleep problems via perceptions of neighborhood quality. We used cross-sectional data for 1251 schoolchildren (8–12 years) in the Tyrol region of Austria/Italy. Questionnaires provided information on sociodemographic and housing factors, perceived neighborhood quality, sound disturbance in different situations, and sleep problems. Modelled acoustic indicators included day-evening-night sound levels and the highest percentile level, and nighttime sound level and a bespoke sleep disturbance index. Nitrogen dioxide served as a proxy for traffic-related air pollution. The normalized difference vegetation index was calculated as a measure of residential greenspace, and presence of a domestic garden was self-reported. Results showed that higher level of traffic-related exposures was positively associated with sound disturbance and sleep problems, while living in a greener area, especially in a house with a garden, was associated with lower sound disturbance and less sleep problems even in the presence of traffic. Traffic exposures contributed to more unfavorable, and greenspace to more positive perceptions in terms of traffic-related stressors, opportunities for outdoor recreation, and general satisfaction with the neighborhood. This indirect path seemed more important for greenspace than for traffic exposures. In conclusion, it seems advantageous to combine traffic-related mitigation with improving access to greenspace in interventions for supporting the acoustic comfort of children during day and nighttime. Even highly nature-dominated environments could still benefit from proximal green infrastructure, especially from domestic gardens. 1. Introduction Annoyance and sleep disturbance are regarded as the most common adverse effects of environmental noise exposure (WHO Regional Office for Europe, 2018). Though soft outcomes themselves, they act as pathways to poor mental and physical health (WHO Regional Office for Europe, 2018). In Europe alone, 22 million people suffer from chronic high annoyance and 6.5 million from chronic high sleep disturbance due to noise (European Environment Agency, 2020). Yet, surprisingly little is known about these impacts on children. Children find themselves in a sensitive developmental period when various stressors can adversely shift the trajectory of their later health (Smith and Pollak, 2020; Person Waye et al., 2023). They have a number of characteristic vulnerabilities to environmental harms such as spending more time outdoors than adults, meaning they are exposed for longer periods of time (Grelat et al., 2016). Children also have a more limited coping repertoire, especially towards chronic cumulative stressors, which can undermine self-regulatory processes and consequently impair adaptive capacities (Bruni et al., 2011; Evans and Kim, 2013; Bilotta et al., 2018). That is, children may not recognize environmental threats to their health, nor * Corresponding author. Department of Hygiene, Faculty of Public Health, Medical University of Plovdiv, 15A Vassil Aprilov Blvd., 4000, Plovdiv, Bulgaria. E-mail addresses: [email protected], [email protected], [email protected] (A.M. Dzhambov). 1 These two authors contributed equally. Contents lists available at ScienceDirect Environmental Research journal homepage: www.elsevier.com/locate/envres https://doi.org/10.1016/j.envres.2023.117642 Received 8 October 2023; Accepted 9 November 2023 Environmental Research 242 (2024) 117642 2 enjoy the same behavioral flexibility as adults in meeting adaptive challenges and threats from the environment (e.g., through window-opening habits, avoidance of a particular setting or situation). Furthermore, from a methodological stance, single-item noise annoyance measurements adopted from adult studies may not sufficiently characterize the effects of noise on children, as few studies have shown (Lercher et al., 2000; Haines et al., 2003; Persson Waye et al., 2011). Such a narrow approach disregards other potentially positive appraisals and coping opportunities provided by the overall physical environment. Beyond annoyance, exposure to traffic noise at night is a prominent stressor undermining normal sleep, according to multiple experimental and observational studies in adults (World Health Organization, 2009; Basner and McGuire, 2018) and fewer studies in children (Bruni et al., 2011; Quehl et al., 2021). Quality sleep is essential for the recovery of children from psychological wear and tear experienced in their daily life, as well as from environmental and social demands and pressures in the residential and school settings. Although earlier studies have shown lower awakening rates in children (Eberhardt, 1988; ¨ Ohrstr¨ om et al., 2006), stress-related cardiovascular reactions are more pronounced in them (Muzet, 2007). A recent review underlines that disruption of normal sleep of children can contribute to a broad spectrum of adverse health effects, such as non-specific symptoms, poor academic performance, neurodevelopmental problems, and somatic disorders like high blood pressure and adiposity (Matricciani et al., 2019). However, most studies on noise-related sleep problems have relied on acoustic indicators based on the average sound energy over a period of time, even though alternative indicators reflecting disturbing and distracting events in a soundscape context may be better predictors of sleep outcomes (De Coensel et al., 2009; Botteldooren et al., 2015; Fredianelli et al., 2022; Lercher and Dzhambov, 2023). Hitherto, these alternative acoustic appraisals have not been applied in studies related to children’s annoyance and sleep. Another critical gap in the noise and child sleep literature is that noise was studied primarily as a stand-alone exposure (e.g., Tiesler et al., 2013; Weyde et al., 2017; Rudolph et al., 2019; Quehl et al., 2021). However, there is a broad spectrum of relevant factors influencing sleep quality (Philippens et al., 2022). Among those, the evidence of an association between air pollution and sleep disorders in children is growing (e.g., Lawrence et al., 2018; S´ anchez et al., 2019), but most of these studies did not account for co-exposure to noise (Liu et al., 2020). While noise acts as a stressor activating subcortical structures in the brain and leads to arousal of the vegetative nervous system and release of stress hormones (Basner et al., 2018), air pollution is believed to affect sleep via neurochemical changes (e.g., serotonin neurochemistry), impaired neurodevelopment of brain structures involved in circadian rhythm regulation, and sleep disordered breathing (Liu et al., 2020). In addition, air pollution has been found to contribute to noise annoyance since perceptions of traffic noise and air pollution can be intertwined (Frei et al., 2014; Oiamo et al., 2015; Lercher, 2019). As a counterpart to these stressors, residential greenspace might be protective against sleep problems, offsetting some of the mechanisms leading to poor sleep. Greenspace, which encompasses home gardens, street trees, parks, and vegetation dominated landscapes in general, can mitigate noise (Van Renterghem et al., 2015) and air pollution (Nowak et al., 2006). Moreover, lower noise annoyance and more favorable appraisals of the acoustic environment have been observed in the presence of green infrastructure like green sound barriers or through green window views (Van Renterghem, 2019; Sch¨ affer et al., 2020). Greenspace also provides opportunities for relaxation, regaining depleted neurocognitive adaptive resources, and supporting health-enhancing behaviors in children (Markevych et al., 2017). A greener environment can be conducive to greater levels of outdoor physical activity and thereof better sleep (Dzhambov et al., 2023). The evidence in children is mixed but in general, physical activity during daytime can improve sleep quality by supporting normal body weight and respiratory function, counteracting the effects of neuroinflammation and stress hormones, and dampening psychological stress (Antczak et al., 2020). A handful of studies in adults have already indicated that people living in greener neighborhoods, especially dominated by trees, have longer and better sleep (Astell-Burt et al., 2013; Johnson et al., 2018; Astell-Burt and Feng, 2019; Li et al., 2022; Stenfors et al., 2023). Some of these studies accounted for co-exposure to air pollution (Li et al., 2022) or noise (Johnson et al., 2018). However, as with the other causal links discussed so far, the evidence in children is scarce. One study did not support such a protective association despite the above-mentioned findings in adult populations (Feng et al., 2020), while another study found a borderline significant association between greenspace and longer sleep duration, which was mediated by perceived stress (Zhong et al., 2023). In light of the paucity of evidence on these associations in children, in the current study, we endeavored to examine the joint effects of traffic exposures and residential greenspace on child sound disturbance and sleep problems. We hypothesized that higher noise and air pollution levels would relate to unfavorable perceptions of neighborhood quality, sound disturbance, and sleep-related problems, while greenspace would work in the opposite direction. Moreover, we anticipated that higher levels of greenspace would weaken the effects of environmental stressors. To test these hypotheses, we employed a suite of acoustic, air pollution, and greenspace indicators, as well as measures of perceived neighborhood characteristics relevant for sleep. 2. Methods 2.1. Study design and setting This study used data previously collected for the Brenner Base Tunnel Study in 2004–2005. Detailed description of the design and sampling has been reported elsewhere (Dzhambov et al., 2019). Briefly, 1251 8–12 years old children were recruited from 49 schools in the Tyrol region of Austria and Italy. The mothers of these children filled out a questionnaire about sociodemographic and housing factors, while children were asked about perceptions of their residential environment, sound disturbance, and sleep problems. Ethical approval was obtained from the Ethics committee of the Medical University Innsbruck (Ethics commission number 2105/2004). Children were sampled from the Lower Inn, Wipp, and side valleys. The area is distinctive in that it brings together well-developed traffic infrastructure and rich natural features. The Lower Inn and Wipp valleys represent major transportation routes through the Alps, with small residential villages located along them. Traffic lines stretch along the valley floor creating unfavorable acoustic conditions with sound wave propagating over great distances to the slopes, meeting few obstacles. While the meteorology and topography of the Lower Inn valley favor the build-up of air pollution, towards the south the valleys are better ventilated and air pollution levels tend to be lower. In contrast, the Alpine landscape in the area offers scenic nature views with lush vegetation, providing multiple opportunities for outdoor recreation. 2.2. Sleep problems and sound disturbance Children answered questions on different aspects of their night-time sleep. Separate items asked about recent time problems falling asleep, uneasy sleep, and feeling tired in the morning. Answers were provided on a 5-point scale (1 =never, 2 =seldom, 3 =sometimes, 4 =often, or 5 =very often). The sum of the three items represented a sleep problems scale where higher values indicated more sleep problems. Nine items were used to measure disturbance by different sounds and in different situations. The agreement with each statement on perceived disturbance was rated on a scale from 1 to 4 (1 =not right at all, 2 = rather wrong, 3 =rather correct, or 4 =exactly right). Specifically, children reported their disturbance by (1) car, (2) truck, and (3) railway noise, (4) general noisiness of the neighborhood, as well as disturbance A.M. Dzhambov et al. Environmental Research 242 (2024) 117642 3 by traffic noise during (5) doing homework, (6) relaxation, (7) watching TV, (8) being outdoors, and (9) trying to fall asleep. We summed responses to these nine items into a sound disturbance scale, with higher disturbance indicated by higher values. 2.3. Perceived neighborhood quality We constructed a summary perceived neighborhood quality scale, which was treated as a mediator. Perceived neighborhood quality was measured with seven items covering different neighborhood characteristics that could be influenced by traffic exposures and greenspace. Children were asked, on a scale from 1 to 4 (the same as for perceived disturbance in Section 2.2), if their neighborhood (1) had a lot of space to play, (2) meadows and trees, and (3) clean air, if it was (4) quiet, if (5) children were allowed to run around, if (6) people were helpful to children, and (7) if children generally enjoyed living there. 2.4. Noise and air pollution by traffic The main sources of traffic noise in the area under investigation include road traffic on the highway in the central valley, other road traffic, and rail traffic including an important contribution of freight on the railroad that belongs to the Trans European Network. We calculated day-evening-night sound levels (L den ) at each address (see Dzhambov et al., 2019 for detailed calculation notes). In addition to the most common noise exposure indicators, L den and night-time sound level (L night ), more explorative indicators that might be relevant for sleep disturbance were added. To account for the loudest events, several indicators have been proposed (Can et al., 2015; Can et al., 2016). Here, the highest percentile level (L 01 ) over the day for all sources combined was also used. This indicator reflected the level during the 864 loudest seconds of the day. It was modelled by augmenting L den at the most exposed façade of the dwelling with an expected 1-s time series. To calculate this augmentation, the distance to the highway, main roads, and the railway were considered while assuming free field propagation, neglecting buildings and terrain. An overall background level with mean value of 37 dB A was added. The source distribution for generating the time series was based on the known traffic intensity and a Poisson distribution of traffic. Further analysis of the L 01 showed that values above 80 dB A were obtained only near railway tracks while the contribution of the main road was more evenly spread over all locations yet remains lower than about 80 dB A. In an attempt to take as much of the existing knowledge on sleep problems in children into account, a new indicator that was modelled ab initio is introduced, the sleep disturbance index (SDI). This indicator was based on detectable indoor noise events expressed in A-weighted level. Dwelling sound insulation (seen as a form of coping) was assumed to depend on average outdoor levels (Locher et al., 2018). For each detected event, the probability of an effect on sleep pattern was estimated based on the vast amount of literature reporting sleep changes in lab during noise events (Griefahn et al., 2008; Basner et al., 2011). In addition to the level itself, also rise time was accounted for. Overall, the probability of sleep changes due to noise is given by PSD =f(N,dN), where N is the loudness of the event and dN is the initial increase in loudness. The function f was modelled as a linear function truncated by zero and 1, and its effect threshold was adapted to the age of the child to account for a lower sensitivity at young age (Lopez-Poveda, 2014; ¨ Ohrstr¨ om et al., 2006). The single event response was then further aggregated over 10-min sleep epochs assuming independence of disturbances: PSD,Tk =∑ Tk i=1 PSD,i∏ i−1 j=1(1−PSD,j) where T k is the duration of the sleep epoch. Finally, a weighted average, stressing morning hours, of the probability of disturbed sleep epochs over the night was taken. This average accounted for the longer sleeping time of young children. To calculate SDI, the temporal evolution of the sound level should be known. Typical road traffic noise models do not account for this. Detailed traffic models can produce this temporal information, yet their deployment in large areas is time consuming mainly because of the setup-time of the micro-traffic simulation. Therefore, the same model that was used for calculating L 01 was also used here. A background level of 35 dB A was assumed. Given sound sources in the Alpine area under study, a total SDI score including highway, major road, and railway noise was calculated. The same model was also used to estimate L night . Mean annual concentrations of nitrogen dioxide (NO 2 ) were used as a proxy for long-term exposure to traffic-related air pollution. NO 2 was derived by the meteorological model Graz Mesoscale Model at a horizontal resolution of 10 ×10 m and a vertical resolution of 2 m (see Dzhambov et al., 2019). Modelled NO 2 levels were assigned to children’s home address. Since noise and air pollution indicators can lead to disturbance and sleep problems via different mechanisms, but are also highly correlated, they were analyzed both individually and as a combined traffic exposures factor. 2.5. Greenspace The normalized difference vegetation index (NDVI) was used as a measure of residential greenspace (Tucker, 1979). It was calculated at a 30 ×30 m resolution using satellite data (cloud free images from July–August 2003) from the Landsat 4–5 Thematic Mapper. The average NDVI value in 100-m, 500-m, and 1000-m circular buffers around child’s home represented general vegetation level, where values closer to +1 indicate high greenness and values closer to −1 indicate build-up land cover and water. NDVI 100-m was used as the main greenspace indicator, since immediate greenspace in the vicinity of the home is theoretically more relevant for child health outcomes and has been found to correlate with sleep difficulties in adults stronger than more distant greenspace (Stenfors et al., 2023). Still, we report single-exposure sensitivity analyses with NDVI in the 500-m and 1000-m buffers, which supported the choice of the 100-m buffer. Mothers also reported whether their house had a garden. By garden, we broadly meant private greenspace, green yards, orchards, floral gardens, or vegetable gardens. 2.6. Confounders and effect modifiers Additional potentially influential variables were available at both the individual and area-levels. Child’s age, sex, and maternal education were reported by the mother. Maternal education served as a proxy for the family’s socioeconomic status (basic for ≤9 years of school, skilled labour, vocational, or higher/A-level). Indoor crowding, which is also related to a family’s socioeconomic status and may affect disturbance, was calculated as people/rooms ratio. Dwelling type could reflect non-acoustic characteristics of the dwelling relevant for noise exposure and disturbance. Dwelling types were single family detached house, row house, or multiple dwelling. Given structural differences in topography, meteorology, and traffic exposures across the valleys in the study area, where children lived was also of interest. The valleys in the study were Lower Inn valley, side Wipp valley north, side Wipp valley south, main Wipp valley north, and main Wipp valley south. 2.7. Statistical analysis The dataset was examined for missing values and variable distributions. Due to missing data, some statistical tests used a smaller analysis sample (i.e., 1017 observations in the main analysis), but the missing A.M. Dzhambov et al. Environmental Research 242 (2024) 117642 4 data on any given variable did not exceed 10%. Following descriptive analyses, bivariate associations between the core variables in the study were tested using Spearman correlations. Due to its distribution (Figure S1), the SDI variable was dichotomized as equal to zero or greater. Separate linear regression models were fitted to examine the associations between L den , L night , L 01 , SDI, NO 2 , NDVI 100-m , and garden, on the one hand, and sound disturbance and sleep problems, on the other. The models were adjusted for child’s sex and age, and maternal education, which were identified as the minimum adjustment set to avoid overfitting. Tolerance (>0.2) and Variance Inflation Factor (<5.0) statistics suggested that these models did not suffer from multicollinearity. We then tested a structural equation model (SEM) including the joint associations between exposures (L den , L 01 , SDI, L night , NO 2 , NDVI 100-m , and garden) and outcomes (sound disturbance and sleep problems), as mediated by perceived neighborhood quality (Fig. 1). As control variables, we also included child’s sex and age, maternal education, and crowding. L den , L 01 , SDI, L night NO 2 were assumed to load onto the same latent construct that we called traffic exposures. The decision to group all traffic-related indicators was based on model fit comparison and the variance explained in them by the latent construct. A priori, we assumed covariances between the exposure variables as shown in Fig. 1. Besides direct and indirect effects of the exposures, we were interested in potential effect modification (moderation) by child sex, dwelling type, and presence of a home garden. We investigated whether these categorical variables modified the direct and indirect paths of interest with the multigroup SEM approach. First, we estimated a multigroup model allowing the paths to be estimated separately and freely across the categories of the putative modifier (e.g., separately for boys and girls, when the modifier under study was sex). Next, we estimated a reduced model, which imposed equality constraints on the paths of interest (i.e., constraining them to be equal across subgroups). We used a chi-square difference test to compare the unconstrained and constrained models. A significant difference in chi-square of these nested models was taken as an indication that the model with more freely estimated parameters fitted the data better, i.e., that the constrained paths differed significantly across different subgroups. To estimate these models, we employed the diagonally weighted least squares (DWLS) estimation method with robust standard errors. All variables were standardized for easier interpretation of the path estimates. Indirect effects were computed as the product of the regression weights associated with their constituent paths, and standard errors for these defined parameters were computed using bootstrapping (1000 draws), except for multi-group models, where the Delta method was used for efficiency instead. The term “effect” is used for consistency with the accepted language in the SEM literature but bears no claims of causality. As a sensitivity analysis, we also rendered the main model with a full-information maximum likelihood (FIML) estimator and bootstrapping to impute missing values. We evaluated goodness-of-fit using indices of acceptable model fit suggested by Hu and Bentler (1999): non-significant χ 2 (p >0.05); CFI ≥0.95; RMSEA ≤0.06 with a 90% CI ≤0.06; and SRMSR ≤0.08. The parsimony normed fit index (PNFI) was also assessed, as it takes into account model complexity. PNFI was expected to be >0.50 (Iacobucci, 2010). Structural equation modeling was conducted with the package lavaan v. 0.6–10 (Rosseel, 2012) in R v. 4.1.2. (R Core Team (2021). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria.). All other analyses were conducted with Stata MP v. 18 (StataCorp. 2023. Stata Statistical Software: Release 18. College Station, TX: StataCorp LLC.). A p-value of <0.05 was considered statistically significant. 3. Results 3.1. Descriptive analyses and bivariate associations The study included 1251 children with a mean age of about 9 years and slightly higher prevalence of girls. Characteristics of the study sample are provided in Table 1. On average, children reported moderate sound disturbance and sleep problems. Most children were exposed to relatively low noise and air pollution levels, while greenspace levels were generally high (Figure S1), and most children lived in a home with a garden. Bivariate associations between the central variables in the study are given in Table S1. A moderately strong positive correlation existed between sound disturbance and sleep problems and these variables were negatively correlated with perceived neighborhood quality. Higher noise and air pollution levels were associated with higher sound disturbance and more sleep problems though the correlations with sleep problems were weak. Conversely, higher NDVI 100 m and home garden were inversely associated with sound disturbance and more sleep problems, again with the correlation with sound disturbance being stronger. The exposures were related to each other in line with theory, with noise and air pollution indicators being positively correlated between them and negatively with NDVI 100 m and home garden. Fig. 1. Conceptual model of the paths linking traffic exposures and greenspace to sound disturbance and sleep problems. Notes: Green lines represent positive associations, while orange lines represent negative (inverse) associations. Abbreviations: L den – day-evening-night sound levels; L 01 – sound level exceeded for 1% of the time of the measurement duration; L night – night-time sound level; NDVI – normalized difference vegetation index; NO 2 – nitrogen dioxide; SDI – sleep disturbance index. A.M. Dzhambov et al. Environmental Research 242 (2024) 117642 5 From Table 2, sound disturbance was higher with higher noise and air pollution levels, while NDVI and home garden were associated with lower disturbance. Similarly, the acoustic indicators were positively associated with sleep problems, while greenspace was associated with less sleep problems. As expected, NDVI in the 100-m buffer was a stronger predictor than NDVI in the larger buffers. 3.2. Main structural equation model The SEM converged normally in 43 iterations and was reasonably consistent with the data: χ 2 (56) =238.19, p <0.001; CFI =0.97, PNFI = 0.59; RMSEA =0.06 (90% CI: 0.05, 0.06); SRMR =0.05. Fig. 2 shows the estimated significant pathways in the model (See Figure S2 for the full path diagram). Sound disturbance and sleep problems were positively correlated. The model explained 29% of the variance in sound disturbance, as opposed to only 5% in sleep problems. Higher traffic exposures and lower neighborhood quality were the most influential predictors of both outcomes. In addition, lower traffic exposures, having a garden, higher NDVI 100-m , and higher maternal education were associated with higher perceived neighborhood quality, which was explained at 14%. All hypothesized indirect effects were supported in the estimated model (Table 3). That is, higher traffic exposures were associated with lower perceived neighborhood quality, and then with higher sound disturbance and more sleep problems. On the other hand, higher NDVI 100-m and garden related to higher perceived neighborhood quality, and in turn, to lower sound distance and less sleep problems. Most total effects also went in the expected direction except for the null effect of NDVI 100 m on sleep problems. When missing values were considered in the estimation to make use of the full dataset using FIML, the model fit was slightly worse. The only materially important difference in the overall associations was that the total effect of garden on sound disturbance was no longer significant, though in the same direction, and that of NDVI 100-m on sleep problems became significant (Figure S3). 3.3. Multi-group structural equation modeling Table S2 shows the results of multigroup SEMs investigating differences in the strengths of the paths from environmental exposures to perceived neighborhood quality and then to the two outcomes. Significant chi-square differences between the unconstrained and corresponding constrained models provided statistical support for effect modification by sex, dwelling type, and garden. Indirect and total effects from these multigroup models are shown in Tables S3–S5. In boys, the total effects of traffic exposures, NDVI 100 m , and garden were more pronounced. In girls, by contrast, the only total effect observed was from traffic exposures to sound disturbance. Stratification by presence of a home garden revealed that the total effect of traffic exposures on sound disturbance was stronger in homes with a garden, and the effect on sleep problems was only present in those homes. NDVI 100-m was associated with lower sound disturbance more strongly in those homes that did not have a garden. Depending on dwelling type, traffic exposures were associated with sound disturbance more strongly in row houses and single-family homes as opposed to multiple dwellings. In addition, in row houses, NDVI 100-m was almost associated with less sleep problems, while for gardens, an inverse association with sleep problems was observed only in single family homes. 4. Discussion 4.1. General findings This study suggested that traffic exposures and greenspace could work simultaneously but in opposite directions as determinants of sound Table 1 Study population characteristics (N =1251). Characteristics Descriptives Missing Socio-demographics Age [years] (Mean ±SD) 9.36 ±0.65 0% Boy (N, %) 623 (49.80) 0% Maternal education (N, %) 3.36% Basic 279 (22.30) Skilled labor 396 (31.65) Vocational 287 (22.94) A-level 247 (19.74) Sound disturbance (Median, 25th – 75th) 15.00 (12.00–19.00) 3.12% Sleep problems (Median, 25th – 75th) 4.00 (1.00–6.00) 0.72% Perceived neighborhood quality (Median, 25th – 75th) 24.00 (22.00–26.00) 2.88% Environmental exposures L den [dBA] (Median, 25th – 75th) 49.89 (42.54–59.54) 0.00% L night [dBA] (Median, 25th – 75th) 41.95 (35.68–52.62) 0.32% L 01 [dBA] (Median, 25th – 75th) 52.10 (41.96–62.84) 0.00% SDI >0 (N, %) 410 (32.77) 0.00% NO 2 [ μ g/m 3 ] (Median, 25th – 75th) 12.68 (9.73–18.53) 9.11% NDVI 100 m (Median, 25th – 75th) 0.43 (0.33–0.53) 0.00% Home garden (N, %) 927 (74.10) 0.96% Other characteristics Crowding [people/rooms] (Median, 25th – 75th) 1.00 (0.80–1.25) 1.60% Valley (N, %) 0.00% Inn valley 251 (20.06) Side Wipp valley north 326 (26.06) Side Wipp valley south 133 (10.63) Main Wipp valley north 326 (26.06) Main Wipp valley south 215 (17.19) Dwelling type (N, %) 0.00% Single family attached 780 (62.35) Row house 176 (14.07) Multiple dwelling 295 (23.58) Notes: Depending on their distribution, we report mean and standard deviation (SD) for normally-distributed variables, median and percentiles for nonnormality distributed variables, and number of cases and percentage across categories for categorical and ordinal variables. Abbreviations: L den – dayevening-night sound levels; L 01 – sound level exceeded for 1% of the time of the measurement duration; L night – night-time sound level; NDVI – normalized difference vegetation index; NO 2 – nitrogen dioxide; SDI – sleep disturbance index. Table 2 Single-exposure models testing the associations of traffic exposures and greenspace with sound disturbance and sleep problems. Exposure indicator Sound disturbance Sleep problems β p β P L den 0.29 <0.001 0.12 <0.001 L night 0.28 <0.001 0.11 <0.001 L 01 0.27 <0.001 0.12 <0.001 SDI 0.23 <0.001 0.09 0.002 NO 2 0.24 <0.001 0.06 0.029 NDVI 100-m −0.22 <0.001 −0.09 0.001 NDVI 500-m −0.18 <0.001 −0.07 0.015 NDVI 1000-m −0.14 <0.001 −0.04 0.227 Garden −0.13 <0.001 −0.09 0.001 Coefficients shown are standardized linear regression coefficients (β) with corresponding significance values (p-value). Exposure indicators are tested one-ata-time in separate models. All models are adjusted for child age, sex, and maternal education. Abbreviations: L 01 – traffic sound level exceeded for 1% of the day; L den – dayevening-night sound levels; L night – night-time road traffic sound level; NDVI – normalized difference vegetation index; NO 2 – nitrogen dioxide; SDI – sleep disturbance index. A.M. Dzhambov et al. Environmental Research 242 (2024) 117642 6 disturbance and sleep quality in schoolchildren. While higher levels of traffic noise and air pollution were positively associated with sound disturbance and sleep problems, living in a greener area, especially in a house with a garden, was associated with lower sound disturbance and less sleep problems even in the presence of traffic. The underlining mechanism for which there was statistical support in the current study was children’s perception of their neighborhood’s characteristics. That is, traffic exposures contributed to more unfavorable, and greenspace to more positive perceptions in terms of traffic-related stressors, opportunities for outdoor recreation, and general satisfaction with the neighborhood. This indirect path seemed more important for greenspace because its overall effect was mostly realized through perceived neighborhood quality. Conversely, traffic exposures had convincing direct effects. We set out to characterize the interweaving of positive and negative, as well as objectively-measured and perceived, aspects of the neighborhood environment. In addition to sleep, it was necessary to consider sound disturbance as a separate outcome rather than its precursor, since in our cross-sectional study, it was not straightforward to ascertain if annoyance conceptually preceded sleep problems or the other way around. Moreover, children may spend their playtime in the same room where they sleep, so the self-reports of these two constructs could overlap (cf. Grelat et al., 2016). We decided to go beyond conventional noise annoyance operationalization employed in most studies among adults (cf. Guski et al., 2017; Nguyen and Yano, 2023). Though it is understood that traffic noise is a stressor for children causing annoyance and a range of adverse emotional responses (Haines et al., 2003) and that childhood is a critical period of developing coping skills and cognitive schemas needed for adaptation to social and environmental adversities (cf. Bruni et al., 2011; Evans and Kim, 2013; Bilotta et al., 2018), empirical research still largely relies on single items to measure noise annoyance in children (Lercher et al., 2000; Haines et al., 2003; Persson Waye et al., 2011). In earlier studies on adults a broader noise reaction approach was supported (Job et al., 2001), and more recent attempts point in the same direction (Kroesen and Schreckenberg, 2011). However, the body of evidence in children is limited and standard scales used in adults may not capture the variety of situations and reactions related to children’s noise disturbance (cf. Lercher et al., 2000; Haines et al., 2003; Persson Waye et al., 2011). In our study, we constructed a multi-item sound disturbance scale, motivated by the fact that children engage in a range of activities, both in and out of doors, typical for their age group but not covered in classic noise annoyance scales developed for adults. Multi-item scales also generally outperform single items in terms of predictive validity of a construct (Diamantopoulos et al., 2012). Although not shown here, association between traffic-related exposures and a single traffic-disturbance item were weaker than with our multi-item disturbance scale. Further, as intermittent and even single sound events are sufficient to cause arousal and sleep fragmentation (Basner and McGuire, 2018), we speculated that using only standard acoustic indicators, which represent the average sound energy during day or nighttime, would not be sensitive enough. To detect the influence on sleep of both continuous sound levels and intermittent acoustic events or emergences above background, we employed L den , L night , L 01 , and SDI as a set, which we considered conceptually relevant for both disturbance and sleep Fig. 2. Structural equation model showing estimated paths linking traffic exposures and greenspace to sound disturbance and sleep problems (N =1020). Notes: Solid lines are associated with statistically significant standardized regression estimates at p <0.05, while dashed lines correspond to non-significant estimates. Observed variables have rectangular outline and latent variables have oval outline. Percentages indicate the variance explained in endogenous variables. Control variables (child’s sex, age, maternal education, and crowding), covariances, and errors terms are not displayed to enhance readability. Abbreviations: L 01 – traffic sound level exceeded for 1% of the day; L den – day-evening-night sound levels; L night – night-time road traffic sound level; NDVI – normalized difference vegetation index; NO 2 – nitrogen dioxide; SDI – sleep disturbance index. Table 3 Estimated paths from traffic exposures and greenspace to sound disturbance and sleep problems in the structural equation model (N =1020). Paths Standardized estimate (95% CI) p-value Indirect effects Traffic exposures → Neighborhood quality → Sound disturbance 0.07 (0.04, 0.11) <0.001 Traffic exposures → Neighborhood quality → Sleep problems 0.03 (0.01, 0.04) 0.004 NDVI 100 m → Neighborhood quality → Sound disturbance −0.10 (−0.13, −0.07) <0.001 NDVI 100 m → Neighborhood quality → Sleep problems −0.04 (−0.06, −0.02) <0.001 Garden → Neighborhood quality → Sound disturbance −0.05 (−0.08, −0.02) 0.002 Garden → Neighborhood quality → Sleep problems −0.02 (−0.03, −0.01) 0.005 Total effects Traffic exposures → Sound disturbance 0.27 (0.19, 0.35) <0.001 Traffic exposures → Sleep problems 0.11 (0.04, 0.18) 0.002 NDVI 100 m → Sound disturbance −0.13 (−0.20, −0.13) <0.001 NDVI 100 m → Sleep problems −0.03 (−0.10, 0.04) 0.397 Garden → Sound disturbance −0.09 (−0.16, −0.03) 0.007 Garden → Sleep problems −0.09 (−0.16, −0.02) 0.010 Abbreviations: NDVI – normalized difference vegetation index. A.M. Dzhambov et al. Environmental Research 242 (2024) 117642 7 outcomes (cf. Fredianelli et al., 2022; Sanok et al., 2022). With their help, we found associations in line with our hypothesis. These acoustic indicators differed not only in their definition but also in the way they were modelled. L 01, took into account all traffic noise sources and starts by adding expected temporal fluctuations to the overall L den . It did so by considering the distance to the three main traffic noise sources in the area: the highway, the main road, and the railway. The calculation of L night and SDI was based on the same simplified traffic model. In contrast to the commonly used L night , SDI represented the probability of sleep disturbance, where sleep disturbance was defined as any change in sleep pattern. In its calculation, we considered that children have increased behavioral tolerance for noise during sleep (Aeschbach, 2021; Arregi et al., 2022) and longer sleep duration. It should be noted though that SDI should be more relevant for autonomous response and long-term health effects than for self-reported sleep disturbance. That could explain why we observed weaker associations with SDI than with L den . Thus, together with L den , which reflected the acoustic situation during the entire day and thus was relevant for sound disturbance and overall neighborhood quality, these indicators were associated with sound disturbance both directly and indirectly via perceived neighborhood quality. Future refinement of the SDI should carefully select sleep outcome measures for which it would be more informative. Much less evidence is available that living in a greener environment can mitigate sound disturbance (Lugten et al., 2018) and even less supported is the contribution of greenspace to better sleep. Studies in adults have observed that having views of green landscapes or simply living in a green neighborhood and closer to greenspace relates to lower noise annoyance, even after accounting for objective sound levels (Van Renterghem and Botteldooren, 2016; Dzhambov, 2017; Van Renterghem, 2019). However, we are unaware of any previous study that has demonstrated or even investigated the associations between greenspace and sound disturbance reactions in children. There is also ambiguous evidence of association between green space and sleep quality in children (Feng et al., 2020). One recent study in adolescents showed that lower stress mediated the association between greenspace and sleep duration (Zhong et al., 2023). Our study offers support for one potential mechanism involving general appreciation of the living environment mediating the association between greenspace and sound disturbance. Upon closer inspection of the items that we used to construct our perceived neighborhood quality scale, one can see that they largely tapped dimensions consistent with scales on perceived restorative quality (Bagot, 2004; Han, 2018), such as relative absence of stressors/nuisances, opportunities for outdoor recreation, and hence restorative experiences. In that, our study seems to add to a growing body of evidence that perceived restorative quality is a central link between natural environments and better psychological well-being (Collado et al., 2017; Dzhambov et al., 2017, 2023). Regarding NO 2 , there were grounds for including it in our set of traffic exposure indicators, as air pollution has been shown to contribute to ratings of noise annoyance beyond noise levels, due to shared perceptual attributes (Frei et al., 2014; Oiamo et al., 2015; Lercher, 2019). It is also biologically plausible that air pollution could lead to sleep problems through respiratory symptoms and by affecting brain chemistry over time (Liu et al., 2020). And yet, our findings for NO 2 should be regarded cautiously, since the correlation between NO 2 and traffic noise in our data could explain the associations found in single exposure models with sound disturbance and sleep problems, i.e., NO 2 could have acted as a proxy for traffic noise. According to a systematic review, most studies considering the role of air pollution in children investigated as outcomes primarily breathing disorders during sleep (Liu et al., 2020). Our sleep problems scale was generic and related to insomnia only and therefore did not tap respiratory symptoms related to air quality. Further, we used NO 2 as a proxy for traffic-related air pollution, but particulate matter has been found to be more strongly associated with sleep disorder symptoms (Lawrence et al., 2018). By showing that traffic exposures were inversely related to perceived neighborhood quality, we replicated a previously observed mechanism in adults (von Lindern et al., 2016) and adolescents (Dzhambov et al., 2017), where traffic noise and air pollution seemed to constrain the perceived restorative quality of the living environment via annoyance and thereby led to worse mental health. Of note, a bidirectional relationship between satisfaction with the living environment and noise annoyance is plausible. For instance, a study in French children showed the role of residential satisfaction in lower child noise annoyance from different sound sources (Grelat et al., 2016). Longitudinal datasets can be used to test which is the stronger predictor over repeated measurements or if in fact a bidirectional relationship is revealed over time. 4.2. Effect modification To understand whether greenspace not only worked in the opposite direction to traffic exposures but could buffer their effects, we compared the pathways of interest in the SEM between children living in a home with and without a garden. We preferred to stratify the model by garden as it was not so strongly spatially correlated with traffic sources like the NDVI. Moreover, gardens represent readily accessible greenspace where children can play and relax. To our surprise, higher traffic exposures were associated with stronger sound disturbance when the child’s home had a garden, and there was no negative effect of traffic on sleep in children living in a home without a garden. We do not mean to suggest that gardens enhance the effects of noise, rather they may reflect unmeasured aspects of local context. For instance, single-family homes in these areas with a garden might be less protected against noise compared with larger apartment homes. This remains a conjecture though. Likewise, the differences in effects observed between children living in different dwelling types could be explained by the different configuration of those. The path from NDVI 100 m also differed depending on the presence of a home garden. It was not that unexpected that higher NDVI was associated with lower sound disturbance only in the absence of a garden, since children without access to a garden were arguably relying more on available greenspace in their neighborhood, while a garden would supplant the need to use more distant greenspace. Most evidence on positive greenspace effects on health is derived from highly urbanized countries/regions (Ye et al., 2022) where home gardens are not so common features as in our study area. In one of the few studies in children examining both urban and rural areas, the amount of greenness, park, and water showed significant relationships with quality of life only in urban/suburban populations but not in rural areas (Tillmann et al., 2018). But that study, as many others, did not include noise and air pollution. Another study that did measure air pollution showed positive effects of greenspace on quality of life regardless of asthma-rhinitis status (Boudier et al., 2022). Our findings contribute to the small evidence base that greenspace, especially gardens, may have positive effects also in rural areas, in spite of the typically abundant background vegetation and the presence of noise and air pollution. Regarding sex, most associations we found were stronger in boys and we speculate that this may be attributed to boys being more physically active outdoors, which would result in greater importance of the outdoor environment. We are unaware of existing literature on the subject of child gender differences for the outcomes of interest in our study. Previous studies in adults have generally pointed to stronger associations between greenspace and physical health outcomes in women (Sillman et al., 2022). 4.3. Implications Childhood mental disorders (Barican et al., 2022) are often seen in schoolchildren who face various taxing demands in their social and academic environment. Social challenges in the family and friend spheres, information overload at school, high parental expectations, and excessive media consumption are common stressors for children (Basu A.M. Dzhambov et al. Environmental Research 242 (2024) 117642 8 and Banerjee, 2020). In this company, physical environment exposures may not be key determinants of childhood mental health on the individual level, but modifications to the living environment have the potential to benefit the child population as a whole (Ye et al., 2022). However, along with mitigation of stressors like noise, interventions should consider improving access to greenspace at the same time. Most greenspace and health research has taken place in urban and suburban areas; therefore, we need to know more about the benefits of greenspace in rural context. Even though our study was undertaken in an area with relatively high levels of greenspace, associations between greenspace and more favorable outcomes were still found. This suggests that even highly nature-dominated residential environments could still benefit from development of proximal green infrastructure, especially from domestic gardens and yards. Our findings are specifically relevant for residents of the study area, as high levels of road and railway traffic were a major source of environmental nuisance at the time of data collection for this survey, and even nowadays the areas are exposed to the same amount of traffic noise. Efforts to address this acoustic situation depend on the faster implementation of the WHO-noise guidelines (WHO Regional Office for Europe, 2018). 4.4. Limitations As with other cross-sectional studies, causality cannot be established in this study. We acknowledge that alternative directions of some relationships observed in our model are plausible. To address the uncertain causality between sound disturbance and sleep problems, we assumed a correlational rather than causal relationship between them. Nevertheless, in other instances, we made judgement calls, as with the positioning of perceived neighborhood quality in conceptual space before sound disturbance. One can appreciate how, in reverse, negative appraisal of the acoustic situation can erode general satisfaction with other aspects of an area (cf. Lercher and Dzhambov, 2023). Our model included a parsimonious set of psychosocial and behavioral determinants of sound disturbance and sleep problems. For instance, we did not use information on media use and screen time, outdoor physical activity, or health status even though these variables could be modifiers or mediators for the exposures under study (Whiting et al., 2021). If they consider these, future studies may be able to explain sleep problems to a greater degree than we could. Symptoms of trouble falling asleep, staying asleep, or feeling restored in the morning are commonly used to operationalize insomnia in children and adolescents (Owens, 2005). However, the agreement between self-reported and objectively measured sleep quality is not ideal (Matthews et al., 2018). Critics may also argue against the conceptualization of the latent “traffic exposures” variable as it combines annual average indicators with event-based and daytime-specific indicators. Our decision on these indicators was based on model performance evaluation, which indicated that, for example, having SDI as a stand-alone predictor of the outcomes in the SEM did not result in better model fit. On the other hand, it was explained to a reasonable degree by the latent variable. Further, exposure misclassification is another possibility since we did not account for children’s time-activity patterns and actual exposure to traffic or greenspace. Tracking participants’ movements across microenvironments can provide better estimates of accumulated exposure (Almanza et al., 2012), but was technically not feasible at the time of data collection. We did not use land use/land cover data to extract indicators of greenspace access, as the study area is rural and public parks are not a relevant type of greenspace exposure. Regarding the acoustic indicators used, one might notice that L night and SDI performed similarly or worse than L den as predictors of sound disturbance and sleep problems in the linear models. This may lead one to question the utility of those indicators if they did not provide superior prediction. This implies that nonmonotonic relationships may exist with said outcomes. However, all these indicators contributed to the latent variable in our model. At this point, further development and testing of the SDI is indicated. 5. Conclusions Higher levels of traffic exposures were positively associated with sound disturbance and sleep problems, and living in a greener area was associated with lower sound disturbance and less sleep problems even in the presence of traffic. Traffic exposures contributed to more unfavorable, and greenspace to more positive perceptions in terms of trafficrelated stressors, opportunities for outdoor recreation, and general satisfaction with the neighborhood. This indirect path seemed more important for greenspace than for traffic. Our findings suggest that mitigation of traffic-related exposures should be coupled with improving access to greenspace, in interventions for supporting the acoustic comfort of children during day and nighttime. Even highly nature-dominated environments could still benefit from proximal green infrastructure, especially from domestic gardens. Credit author statement Angel M. Dzhambov: Conceptualization, Methodology, Software, Formal analysis, Investigation, Writing – original draft, Writing – review & editing. Peter Lercher: Conceptualization, Methodology, Investigation, Writing – review & editing, Project administration. Dick Botteldooren: Methodology, Software, Investigation, Writing – review & editing. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Data availability The authors do not have permission to share data. Acknowledgements We want to thank the inhabitants of the Lower Inn and Wipp valleys. Our thanks also go to the Austrian Ministry of Science and Transportation for funding the framework of the Environmental Health Impact Assessment (EHIA), the government of the Tyrol region for providing GIS data and informational support from the BEG (Brenner Eisenbahn Gesellschaft). The BBT survey got support from the BBT company within a legally required EHIA through EU-support. The noise mapping was done by Ghent University, and the air pollution assessment by an Italian-Austrian consortium. We also thank the large EHIA-teams in both studies who did the fieldwork. Finally, we thank Iana Markevych and Matthew Browning for calculating the normalized difference vegetation index. This project has received funding from the European Union’s Horizon 2020 research and innovation program under grant agreement No 874724. Equal-Life is member of the European Human Exposome Network which brings together nine research projects, receiving over € 100 million from Horizon 2020, the EU’s framework program for research and innovation. Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi. org/10.1016/j.envres.2023.117642. A.M. Dzhambov et al. Environmental Research 242 (2024) 117642 9 References Aeschbach, D., 2021. Effects of nocturnal aircraft noise and non-acoustical factors on short-term annoyance in primary school children. Int. J. Environ. Res. Publ. Health 18, 6959. https://doi.org/10.3390/ijerph18136959. Almanza, E., Jerrett, M., Dunton, G., Seto, E., Pentz, M.A., 2012. A study of community design, greenness, and physical activity in children using satellite, GPS and accelerometer data. Health Place 18 (1), 46–54. https://doi.org/10.1016/j. healthplace.2011.09.003. Antczak, D., Lonsdale, C., Lee, J., Hilland, T., Duncan, M.J., , Del Pozo 867 Cruz B, Hulteen, R.M., Parker, P.D., Sanders, T., 2020. Physical activity and sleep are inconsistently related in healthy children: a systematic review and meta-analysis. Sleep Med. Rev. 51, 101278 https://doi.org/10.1016/j.smrv.2020.101278. Arregi, A., Lertxundi, A., Vegas, O., García-Baquero, G., Ibarluzea, J., Anabitarte, A., Barroeta, Z., Jimeno-Romero, A., Subiza-P´ erez, M., Lertxundi, N., 2022. Environmental noise exposure and sleep habits among children in a cohort from northern Spain. Int. J. Environ. Res. Publ. Health 19 (23), 16321. https://doi.org/ 10.3390/ijerph192316321. Quehl, J.; Bartels, S.; Fimmers, R. Astell-Burt, T., Feng, X., Kolt, G.S., 2013. Does access to neighbourhood green space promote a healthy duration of sleep? Novel findings from a cross-sectional study of 259 319 Australians. BMJ Open 3 (8), e003094. https://doi.org/10.1136/bmjopen2013-003094. Astell-Burt, T., Feng, X., 2019. Does sleep grow on trees? A longitudinal study to investigate potential prevention of insufficient sleep with different types of urban green space. SSM Popul Health 10, 100497. https://doi.org/10.1016/j. ssmph.2019.100497. Erratum in: SSM Popul Health. 2020 Dec 10;12:100715. PMID: 32025565; PMCID: PMC6996010. Bagot, K.L., 2004. Perceived restorative components: a scale for children. Child. Youth Environ. 14 (1), 107–129. Barican, J.L., Yung, D., Schwartz, C., Zheng, Y., Georgiades, K., Waddell, C., 2022. Prevalence of childhood mental disorders in high-income countries: a systematic review and meta-analysis to inform policymaking. Evid. Base Ment. Health 25 (1), 36–44. https://doi.org/10.1136/ebmental-2021-300277. Basner, M., McGuire, S., 2018. WHO environmental noise guidelines for the European region: a systematic review on environmental noise and effects on sleep. Int. J. Environ. Res. Publ. Health 15 (3), 519. https://doi.org/10.3390/ijerph15030519. Basner, M., Müller, U., Elmenhorst, E.M., 2011. Single and combined effects of air, road, and rail traffic noise on sleep and recuperation. Sleep 34. Basu, S., Banerjee, B., 2020. Impact of environmental factors on mental health of children and adolescents: a systematic review. Child. Youth Serv. Rev. 119 https:// doi.org/10.1016/j.childyouth.2020.105515. Article 105515. Bilotta, E., Vaid, U., Evans, G.W., 2018. Environmental stress. Environmental psychology: An introduction 36–44. Botteldooren, D., Andringa, T., Aspuru, I., Brown, A.L., Dubois, D., Guastavino, C., et al., 2015. From sonic environment to soundscape. Soundscape and the built environment 36, 17–42. Boudier, A., Markevych, I., Jacquemin, B., Abramson, M.J., Accordini, S., Forsberg, B., et al., 2022. Long-term air pollution exposure, greenspace and health-related quality of life in the ECRHS study. Sci. Total Environ. 849, 157693. Bruni, L., Novelli, R., Ferri, R., 2011. Sleep disturbance in children by noise. In: Encyclopedia of Environmental Health. Elsevier Science and Technology, Amsterdam, The Netherlands, pp. 88–94. Can, A, Guillaume, G, Gauvreau, B, 2015. Noise indicators to diagnose urban sound environments at multiple spatial scales. Acta Acust unit Acust 101, 964. Can, A., Aumond, P., Michel, S., de Coensel, B., Ribeiro, C., Botteldooren, et al., 2016. Comparison of noise indicators in an urban context. In: Inter-Noise (Ed.), 45th International Congress and Exposition of Noise Control Engineering. HAMBOURG, Germany, p. 9 hal-01373857v3. Collado, S., Staats, H., Corraliza, J.A., Hartig, T., 2017. Restorative environments and health. Handbook of environmental psychology and quality of life research 127–148. De Coensel, B., Botteldooren, D., De Muer, T., Berglund, B., Nilsson, M.E., Lercher, P., 2009. A model for the perception of environmental sound based on notice-events. J. Acoust. Soc. Am. 126 (2), 656–665. Diamantopoulos, A., Sarstedt, M., Fuchs, C., et al., 2012. Guidelines for choosing between multi-item and single-item scales for construct measurement: a predictive validity perspective. J. Acad. Market. Sci. 40, 434–449. https://doi.org/10.1007/ s11747-011-0300-3. Dzhambov, A.M., 2017. More residential greenspace is associated with lower noise annoyance: results from a quantitative synthesis of the literature. In: Łucjan, C., G´ erard, D. (Eds.), Traffic Noise: Exposure, Health Effects and Mitigation. Nova Science Publishers, New York, pp. 77–104. Dzhambov, A.M., Lercher, P., Vincens, N., Persson Waye, K., Klatte, M., Leist, L., Lachmann, T., Schreckenberg, D., Belke, C., Ristovska, G., Kanninen, K.M., Botteldooren, D., Van Renterghem, T., Jeram, S., Selander, J., Arat, A., White, K., Julvez, J., Clark, C., Foraster, M., van Kamp, I., 2023. Protective effect of restorative possibilities on cognitive function and mental health in children and adolescents: a scoping review including the role of physical activity. Environ. Res. 233, 116452 https://doi.org/10.1016/j.envres.2023.116452. Dzhambov, A.M., Markevych, I., Lercher, P., 2019. Associations of residential greenness, traffic noise, and air pollution with birth outcomes across Alpine areas. Sci. Total Environ. 678, 399–408. https://doi.org/10.1016/j.scitotenv.2019.05.019. Dzhambov, A., Tilov, B., Markevych, I., Dimitrova, D., 2017. Residential road traffic noise and general mental health in youth: the role of noise annoyance, neighborhood restorative quality, physical activity, and social cohesion as potential mediators. Environ. Int. 109, 1–9. Eberhardt, J.L., 1988. The influence of road traffic noise on sleep. J. Sound Vib. 127 (3), 449–455. European Environment Agency, 2020. Environmental Noise in Europe — 2020. https:// doi.org/10.2800/686249. Evans, G.W., Kim, P., 2013. Childhood poverty, chronic stress, self-regulation, and coping. Child development perspectives 7 (1), 43–48. Feng, X., Flexeder, C., Markevych, I., Standl, M., Heinrich, J., Schikowski, T., Koletzko, S., Herberth, G., Bauer, C.P., von Berg, A., Berdel, D., Astell-Burt, T., 2020. Impact of residential green space on sleep quality and sufficiency in children and adolescents residing in Australia and Germany. Int. J. Environ. Res. Publ. Health 17 (13), 4894. https://doi.org/10.3390/ijerph17134894. Fredianelli, L., Lercher, P., Licitra, G., 2022. New indicators for the assessment and prevention of noise nuisance. Int. J. Environ. Res. Publ. Health 19 (19), 12724. Frei, P., Mohler, E., R¨ o¨ osli, M., 2014. Effect of nocturnal road traffic noise exposure and annoyance on objective and subjective sleep quality. Int. J. Hyg Environ. Health 217 (2–3), 188–195. https://doi.org/10.1016/j.ijheh.2013.04.003. Grelat, N., Houot, H., Pujol, S., Levain, J.P., Defrance, J., Mariet, A.S., Mauny, F., 2016. Noise annoyance in urban children: a cross-sectional population-based study. Int. J. Environ. Res. Publ. Health 13 (11), 1056. https://doi.org/10.3390/ijerph13111056. Griefahn, B., Br¨ ode, P., Marks, A., Basner, M., 2008. Autonomic arousals related to traffic noise during sleep. Sleep 31 (4), 569–577. https://doi.org/10.1093/sleep/31.4.569. Guski, R., Schreckenberg, D., Schuemer, R., 2017. WHO environmental noise guidelines for the European region: a systematic review on environmental noise and annoyance. Int. J. Environ. Res. Publ. Health 14 (12), 1539. https://doi.org/10.3390/ ijerph14121539. Haines, M.M., Brentnall, S.L., Stansfeld, S.A., Klineberg, E., 2003. Qualitative responses of children to environmental noise. Noise Health 5 (19), 19–30. Han, K.T., 2018. A review of self-report scales on restoration and/or restorativeness in the natural environment. J. Leisure Res. 49 (3–5), 151–176. https://doi.org/ 10.1080/00222216.2018.1505159. Hu, L.T., Bentler, P.M., 1999. Cutoff criteria for fit indexes in covariance structure analysis: conventional criteria versus new alternatives. Struct. Equ. Model. 6, 1–55. Iacobucci, D., 2010. Structural equations modeling: fit Indices, sample size, and advanced topics. J. Consum. Psychol. 20, 90–98. https://doi.org/10.1016/j. jcps.2009.09.003. Job, R.F.S., Hatfield, J., Carter, N.L., Peploe, P., Taylor, R., Morrell, S., 2001. General scales of community reaction to noise (dissatisfaction and perceived affectedness) are more reliable than scales of annoyance. J. Acoust. Soc. Am. 110 (2), 939–946. Johnson, B.S., Malecki, K.M., Peppard, P.E., Beyer, K.M.M., 2018. Exposure to neighborhood green space and sleep: evidence from the Survey of the Health of Wisconsin. Sleep Health 4 (5), 413–419. https://doi.org/10.1016/j. sleh.2018.08.001. Kroesen, M., Schreckenberg, D., 2011. A measurement model for general noise reaction in response to aircraft noise. J. Acoust. Soc. Am. 129 (1), 200–210. Lawrence, W.R., Yang, M., Zhang, C., Liu, R.Q., Lin, S., Wang, S.Q., Liu, Y., Ma, H., Chen, D.H., Zeng, X.W., Yang, B.Y., Hu, L.W., Yim, S.H.L., Dong, G.H., 2018. Association between long-term exposure to air pollution and sleep disorder in Chinese children: the Seven Northeastern Cities study. Sleep (9), 41. https://doi.org/ 10.1093/sleep/zsy122. Lercher, P., 2019. Combined transportation noise exposure in residential areas. In: Encyclopedia of Environmental Health. Elsevier, pp. 695–712. Lercher, P., Brauchle, G., Kofler, W., Widman, U., Meis, 2000. The assessment of noise annoyance in schoolchildren and their mothers. In: Proceedings of the 29th International Congress and Exhibition on Noise Control Engineering. Soci´ et´ e Française d’Acoustique, Nice, pp. 2318–2322. Lercher, P., Dzhambov, A.M., 2023. Soundscape and health. In: Schulte-Fortkamp, B., Fiebig, A., Sisneros, J.A., Popper, A.N., Fay, R.R. (Eds.), Soundscapes: Humans and Their Acoustic Environment, Springer Handbook of Auditory Research, vol. 76. Springer, Cham. https://doi.org/10.1007/978-3-031-22779-0_9. Li, G., Zhu, Z., Hu, M., He, J., Yang, W., Zhu, J., Zhao, H., Zhang, H., Huang, F., 2022. Effects of carbon dioxide and green space on sleep quality of the elderly in rural areas of Anhui Province, China. Environ. Sci. Pollut. Res. Int. 29 (14), 21107–21118. https://doi.org/10.1007/s11356-021-17296-7. Liu, J., Wu, T., Liu, Q., Wu, S., Chen, J.C., 2020. Air pollution exposure and adverse sleep health across the life course: a systematic review. Environ. Pollut. 262, 114263 https://doi.org/10.1016/j.envpol.2020.114263. Locher, B., Piquerez, A., Habermacher, M., Ragettli, M., R¨ o¨ osli, M., Brink, M., Cajochen, C., Vienneau, D., Foraster, M., Müller, U., Wunderli, J.M., 2018. Differences between outdoor and indoor sound levels for open, tilted, and closed windows. Int. J. Environ. Res. Publ. Health 15 (1), 149. https://doi.org/10.3390/ ijerph15010149. Lopez-Poveda, E.A., 2014. Development of fundamental aspects of human auditory perception. In: Development of Auditory and Vestibular Systems, fourth ed. Lugten, M., Karacaoglu, M., White, K., Kang, J., Steemers, K., 2018. Improving the soundscape quality of urban areas exposed to aircraft noise by adding moving water and vegetation. J. Acoust. Soc. Am. 144 (5), 2906–2917. Markevych, I., Schoierer, J., Hartig, T., Chudnovsky, A., Hystad, P., Dzhambov, A.M., de Vries, S., Triguero-Mas, M., Brauer, M., Nieuwenhuijsen, M.J., Lupp, G., Richardson, E.A., Astell-Burt, T., Dimitrova, D., Feng, X., Sadeh, M., Standl, M., Heinrich, J., Fuertes, E., 2017. Exploring pathways linking greenspace to health: theoretical and methodological guidance. Environ. Res. 158, 301–317. https://doi. org/10.1016/j.envres.2017.06.028. Matricciani, L., Paquet, C., Galland, B., Short, M., Olds, T., 2019. Children’s sleep and health: a meta-review. Sleep Med. Rev. 46, 136–150. Matthews, K.A., Patel, S.R., Pantesco, E.J., Buysse, D.J., Kamarck, T.W., Lee, L., Hall, M. H., 2018. Similarities and differences in estimates of sleep duration by A.M. Dzhambov et al.