1. Introduction
Sound is a fundamental yet often overlooked component of ecosystems, shaping the behavior, distribution, and interactions of organisms that produce and perceive it. However, increasing anthropogenic noise is altering natural soundscapes worldwide, posing emerging threats to biodiversity and ecosystem function. Soundscape ecology has therefore emerged as a crucial field to investigate these dynamics, analyzing the interplay between biological, anthropogenic, and geophysical sound sources across spatial and temporal scales [
1]. In this context, the concept of soundscapes has become an essential tool to understand how human activities disrupt natural acoustic patterns and interact with biotic and abiotic components of ecosystems, with implications for species communication, habitat selection, and ecosystem resilience. This multidisciplinary approach makes it possible to monitor the health of ecosystems, identify sources of acoustic stress, and assess the impact of sound dynamics on the organism’s present [
2]. Understanding how vegetation mediates the effects of anthropogenic noise is essential for biodiversity conservation, especially in fragmented or peri-urban forest ecosystems that are increasingly exposed to human pressures. Birds in particular rely heavily on acoustic communication, and chronic exposure to noise can disrupt mating, territorial defense, and foraging behavior, ultimately leading to shifts in community composition. Identifying how landscape structure can mitigate these effects provides crucial insights for developing more effective management strategies in the face of increasing anthropization.
The soundscape of a place is shaped by a series of factors. For instance, vegetation has a crucial role in modulating the sound environment, influencing not only sound propagation but also the interactions between living organisms [
3]. Several studies have shown that vegetation structure, tree cover density, and diversity can significantly model the soundscape of an area, modifying the refraction and absorption of sound waves [
4]. The interaction between anthropogenic noise and vegetation structure is crucial in shaping the soundscape. Vegetation not only affects sound propagation but also modulates the impact of noise on biodiversity, particularly for species like birds that depend on acoustic signals for communication. In addition, vegetation should be considered as it plays a crucial role in providing nesting and feeding sites for bird communities; these roles need to be taken into account when assessing the impact of anthropophonic noise on bird communities. Studying soundscape dynamics in anthropized landscapes presents unique challenges due to the overlap of natural and anthropogenic sound sources, the spatial and temporal variability of human activities, and the heterogeneity of land use. These factors make it difficult to isolate biotic acoustic patterns and often obscure the ecological signals of interest. In addition, anthropogenic noise can mask or alter important biological sounds, making it difficult to assess ecosystem health or species abundance using passive acoustic methods. Addressing these issues requires careful study design at an appropriate scale and a robust methodological framework capable of disentangling overlapping signals. However, the relationships between soundscape and vegetation are complex and differ significantly according to the environmental contexts, including the presence of anthropogenic disturbances. Indeed, a better understanding of the anthropogenic component impacts requires local-scale studies to be adequately dimensioned to consider the landscape role. Despite increasing attention to the role of vegetation in shaping the acoustic environment, previous research has largely focused on either sound propagation mechanisms or biodiversity responses, often neglecting the integrated analysis of forest structure and anthropogenic noise. Furthermore, few studies have investigated these dynamics at the local scale, where landscape heterogeneity and human pressures interact in complex ways. This study aims to address this gap by combining ecoacoustic metrics with detailed structural vegetation analysis to investigate how forest characteristics influence soundscape composition and the spatial distribution of avian communities in anthropized forest ecosystems.
To deal with this complexity, the use of ecoacoustic indices has gained popularity in recent years. These quantitative metrics simplify the analysis of the soundscape, focusing on different sound characteristics [
5], allowing for the assessment of specific acoustic features and providing a more comprehensive picture of the environment’s complexity [
6]. In fact, the use of multiple indices allows the acquisition of more specific information regarding different aspects of sound, such as sound pressure, sound quality, and directionality; it enables the adaptation of analyses to specific application needs, such as the assessment of environmental noise or sound quality; and it reduces the likelihood to obtain a non-exhaustive or biased interpretation of the sound environment [
7].
The main objective of this study is to investigate how vegetation structure modulates soundscape composition in forested ecosystems exposed to different levels of anthropogenic disturbance, with a particular focus on the spatial distribution of avifauna. By integrating ecoacoustic indices with forest structural variables, this research aims to identify patterns and potential thresholds in noise–vegetation–fauna interactions. These findings can inform future ecoacoustic assessments by highlighting the importance of considering vegetation attributes when interpreting acoustic data, especially in contexts where human activity is present. Ultimately, the study contributes to the development of more accurate and context-sensitive tools for biodiversity monitoring and conservation planning.
The structure of the paper is organized to reflect the logical flow of the research.
Section 2 outlines the study area and the methods used for data collection and analysis. The ecoacoustic survey is described first, including instrumentation, index calculation, and statistical analyses. This is followed by the vegetation assessment, which details the estimation of structural parameters and tree biomass, as well as the analysis of vegetation variables. The final part of the Section examines the correlations between acoustic indices and vegetation metrics.
Section 3 presents the main results, while
Section 4 discusses the results in the context of the existing literature.
Section 5 concludes the paper by summarizing the main findings and implications of the study.
4. Discussion
The present study addresses the complex interactions between soundscape and vegetation structure in a protected area impacted by anthropogenic noise through a multidisciplinary methodology. Analyzing both the ecoacoustic indices and vegetation structure allowed for a detailed characterization of the spatial and temporal variability of the oxbow’s soundscape, highlighting how the proximity to the highway and the vegetation distribution significantly influence the acoustic environment. These results underscore the importance of vegetation as a key element in the attenuation and modulation of anthropogenic noise, and highlight the need for multidisciplinary approaches to enhance the understanding of soundscape patterns.
The methodology used in this study improves the discrimination of acoustic components and emphasizes the need to optimize the employed ecoacoustic indices’ parameters to ensure comparability of data across different studies [
11,
24]. The inclusion of a plurality of indices [
50,
51] allowed a more detailed view of the structure of the soundscape, enabling accurate discrimination between biophonic and anthropophonic components. This approach was guided by the evidence reported in the literature, which emphasizes the importance of using specific frequency bands to monitor distinct taxonomic groups. In particular, Metcalf et al. [
52] highlight this necessity in their recent publications, limiting their use to a narrower set of indices, such as ACI and BI. Other works, such as the one published by Bradfer-Lawrence et al. [
53], also emphasize the aforementioned necessity, albeit to a lesser extent than the other recommendations made in the study. Despite these recommendations, a significant portion of the scientific community continues to use default parameters for the calculation of acoustic indices, while others do not specify the criteria behind the selected values [
21,
22], or these values are not reported at all [
54,
55].
Although spatial autocorrelation represents a potential problem in ecological acoustics studies, our sampling scheme aimed to minimize its impact by selecting sites along environmental and anthropogenic gradients, ensuring heterogeneity in vegetation structure and exposure to noise sources. Furthermore, the same recording campaign was analyzed by Benocci et al. (2025) [
56], where we applied the Transfer Entropy Measure to the dataset to investigate directional dependencies in the flow of acoustic information. Our results confirmed the presence of spatial differentiation among sites and supported the hypothesis of non-redundancy in the composition of the soundscape. Finally, the authors conducted a direct listening analysis of the recordings [
56], which highlighted perceptible differences in acoustic content even between adjacent sites. These insights strengthen the ecological validity of our spatial arrangement, while acknowledging the lack of formal tests of spatial autocorrelation as a limitation to be addressed in future studies.
The trends observed in the ecoacoustic indices (H, ZCR, NDSI, and ACI) analysis clearly reveal a spatial gradient in relation to distance from the motorway. The H-index showed a clear reduction at the sites closest to the noise source, showing a lower sound diversity. In contrast, at more distant sites, H recorded higher values, signaling greater acoustic complexity, due to a richer presence of avian species. Similarly, the mean value of ZCR at sites closer to the highway was significantly lower, reflecting the presence of stable and unvarying acoustic signals, typical of continuous anthropogenic noise [
57]. Contrarily, at sites further away, ZCR showed greater variability, associated with the complexity of bird vocalizations. Finally, the NDSI and ACI indices revealed how the balance between biophonic and anthropophonic components varies considerably along the distance gradient. At the sites closest to the highway, the NDSI recorded extremely negative values, suggesting an almost complete dominance of the anthropophonic component [
58]. However, in some areas, such as site 9, the NDSI showed slightly higher values than the other neighboring sites, suggesting the presence of some degree of biophony, despite the anthropogenic influence. These results are corroborated by the ACI values, which confirm greater biophonic activity in the early morning hours, particularly at sites further away from traffic [
59]. Similarly to our results, the study carried out by M. G. Khanaposhtani et al. [
60] in a floodplain forest in Wisconsin (USA)—an ecosystem comparable to the Lanca del Moriano—exposed to the impact of two major highways, reported a significant inverse correlation between ADI, NDSI, and AOI (Acoustic Occupancy Index) with the distance from the road infrastructure. In addition, the inverse gradient between H with highway distance observed at the Lanca del Moriano is in line with the studies carried out at the Parco Nord of Milano, an urban Park also impacted by a highway [
61]. These trends can all be explained by the masking effect of traffic that shapes the distribution of local biophony, according to the Acoustic Habitat Hypothesis [
62] and other studies. Indeed, López-Bao et al. (2017) confirm that road noise has a significant impact on wildlife, as it masks natural sounds and interferes with the acoustic communication of animals [
63]. These effects are also exposed by other in situ studies worldwide, such as [
60], and noise effects on birds are well documented in the literature [
64].
Regarding highway noise in ecoacoustics studies, in [
60] is highlighted a potential bias in the application of NDSI in environments where noise attenuation occurs due to geometric divergence [
60]. This phenomenon has already been documented by R. Benocci et al. in a semi-natural area characterized by herbaceous layers of nemoral flora and a shrub layer, typical of Parco Nord in Milan [
61], and by R. B. Machado et al. in a Category V protected area as defined by the International Union for Conservation of Nature (IUCN) [
65]. However, the accuracy of the NDSI was improved by the methodological optimization of the input parameters for the calculation of ecoacoustic indices [
24], as proposed in the present study. Similarly, cluster analysis further supported the existence of two distinct acoustic configurations, indicating a clear dichotomy between sites dominated by a soundscape rich in biophonic signals and those predominantly influenced by anthropogenic noise. The geometric divergence bias in NDSI could also be present in other indices that present a lower frequency limit (i.e., ADI in [
60]), which has been bypassed in this study and in [
60] by adding it. On this matter, there is a need to adjust the ecoacoustic indices in the R package “soundecology” to add this parameter, which is already adjusted in the Python library “scikit-maad” [
66].
The results regarding vegetation structure and composition reveal a complex landscape, characterized by significant differences between the analyzed plots. These differences are particularly evident when comparing the forest structure and biomass of the sites inside the oxbow (P1–P6) to those near the motorway (P7–P9). This variation in vegetation likely influences the bird community, as differences in habitat complexity and resource availability can impact avian diversity and behavior [
67], ultimately shaping the observed sound dynamics [
59]. Biomass is significantly higher at sites 4, 6, and 9, indicating a denser vegetation structure and, potentially, richer biophonic activity in these areas [
68]. Despite its high biomass, site 9 is located near the highway and shows extremely negative NDSI values, indicating a predominance of the anthropophonic component over the biophonic one. These observations suggest that the masking effect of road noise reduces the ecological benefits of dense vegetation, thereby negatively affecting faunal diversity and activity. This is further supported by the reduced H-index observed at this site, highlighting the negative impact of anthropogenic noise on the complexity of the soundscape [
60].
The Shannon index showed high values at sites 4, 1, and 2, indicating greater plant diversity. Interestingly, site 7, located close to the highway, also recorded a high HS value, which can be attributed to the presence of young trees of numerous plant species. However, despite this phase of active habitat renewal, the site-specific ecoacoustic indices suggest that anthropogenic noise associated with the highway influences the soundscape, as indicated by the reduced values of H and ZCR.
The distribution graphs of the diametric classes in the areas closest to the motorway exhibit decreasing curvilinear trends, indicative of an uneven-aged forest structure. This pattern may reflect forest management practices, such as the introduction of new species, which could explain the observed variability in forest composition and structure [
69]. At the same time, the scarcity of larger individuals in the sites closest to the motorway indicates a low capacity to support high avian diversity, creating a sub-optimal environment for forest-dwelling species. This result is consistent with that reported by López-Bao et al. (2017), who discussed the negative impact of road networks on biodiversity, including birds, highlighting how road infrastructure can fragment habitats and reduce ecological resources for wildlife [
70].
Correlation analysis revealed significant relationships between vegetation structure and ecoacoustic indices. In particular, NDSI showed a positive correlation with
Quercus robur abundance (QUE, r = 0.726,
p = 0.041), which may indicate that mature trees contribute a stronger biophonic signal to the acoustic environment. This could be related to their structural complexity and ability to support a more diverse faunal community, particularly birds and insects. Previous studies have highlighted how vegetation structure can influence the balance between biophony and anthropophony in soundscapes [
13]. However, further research is needed to isolate the specific effect of tree composition from other potential confounding variables such as landscape configuration or proximity to noise sources [
19]. Similarly, BI was positively correlated with tree species richness (SR, r = 0.761,
p = 0.028), reinforcing the idea that diverse plant assemblages favor complex acoustic environments by providing diverse habitats and resources for fauna. In contrast, DSC was negatively correlated with FCover (r = −0.793,
p = 0.018), suggesting that denser vegetation attenuates high-frequency sounds, reducing sound propagation and contributing to a more stable acoustic environment.
Significant relationships were also found between the ecoacoustic indices and the spatial distribution of the sampling points. In particular, DSC showed a strong positive correlation with distance from the motorway (Dist_A7, r = 0.814,
p = 0.013), confirming that increasing distance from traffic sources leads to a reduction in high-frequency noise levels due to sound absorption by vegetation. In contrast, DSC showed a negative correlation with distance from clearing (Dist_ra, r = −0.891,
p = 0.002), suggesting that open areas characterized by lower vegetation density allow greater sound propagation and higher acoustic energy at high frequencies. Similarly, ZCR, an index of acoustic signal complexity, was negatively correlated with both distance from the bamboo patch (Dist_mb, r = −0.717,
p = 0.049) and clearing (Dist_ra, r = −0.813,
p = 0.014). This pattern may reflect higher acoustic variability in transitional or edge environments, which often host a mix of vocal species from both open and closed habitats [
19]. However, since ZCR is sensitive to a broad range of high-frequency and transient signals, further investigation is needed to clarify the ecological drivers underlying these values.
Vegetation variables also showed significant correlations with specific distances from sites. In particular, vegetation cover (FCover) was positively correlated with distance from the clearing (Dist_ra, r = 0.718, p = 0.044), confirming that moving away from open areas leads to higher forest density. In contrast, FCover was negatively correlated with both distance from the highway and forest edge, suggesting that human disturbance limits vegetation expansion. Furthermore, HS and E are negatively correlated with distance from the river, highlighting the influence of water availability on vegetation structure.
While this study focused on local-scale distances from specific landscape features (e.g., highway, clearing, river), it is important to recognize that broader landscape structure [
71], including habitat connectivity and fragmentation across the Ticino Valley Regional Park, likely plays an additional role in shaping avian assemblages. Fragmented landscapes can limit dispersal, reduce nesting site availability, and isolate populations, while well-connected habitats facilitate movement and support higher biodiversity [
72,
73]. Future research integrating landscape metrics (e.g., patch size, edge density, connectivity indices) derived from remote sensing or GIS analysis could further clarify how large-scale spatial configuration modulates local soundscape and biodiversity patterns.
The positive correlation between NDSI and biomass suggests that vegetation structure plays a fundamental role in shaping acoustic diversity. Forested environments with higher biomass, characterized by well-developed vegetation layers, appear to promote a richer and more complex biophonic component, likely due to their ability to support a greater abundance and diversity of soniferous species [
59,
60]. This finding is consistent with ecological theories linking habitat complexity to acoustic variability, as structurally rich ecosystems provide enhanced niches and resources for bioacoustic activity [
22,
26,
74].
Furthermore, the observed positive correlation between biomass and bird species richness reinforces the idea that productivity and structural characteristics of vegetation influence the diversity and abundance of bird communities. High biomass levels typically correspond to increased availability of resources, including food sources such as insects and seeds, and nesting sites, which are critical for the persistence of specialized bird species. In particular, mature forest stands dominated by large trees, such as
Quercus robur, contribute significantly to avian biodiversity by providing microhabitats that support a diverse assemblage of species [
67]. The structural complexity of these trees, including their extensive canopies and the presence of cavities, provides essential ecological resources for cavity-nesting birds, a group that is highly dependent on the availability of mature forest elements [
75,
76].
In addition to their role as physical habitat providers, mature trees may also influence the spatial organization of acoustic signals within forest ecosystems. The hypothesis of acoustic habitat filtering suggests that certain structural features of the habitat may favor species whose vocalizations are best transmitted or least masked in that particular acoustic environment [
77,
78]. In this context, large individuals of
Quercus robur, with their complex tree architecture—such as irregular branch angles, cavities, and layered foliage—may generate multiple sound reflections, absorptions, and diffusions. Such structural complexity may lead to the formation of distinct spectral and spatial acoustic niches, facilitating acoustic niche partitioning and enabling the coexistence of bird species that use different frequency bands and calling strategies [
79,
80]. Our results support this framework, suggesting that the ecological value of mature oak extends to shaping the acoustic landscape in a way that supports diverse acoustic communities.
Conversely,
Carpinus betulus, although a common tree species at some sampling sites, did not show a significant correlation with avian species richness. This result suggests that tree maturity, rather than mere presence, is a key determinant of avian diversity. Unlike
Quercus robur,
Carpinus betulus lacks the structural characteristics necessary to support specialized forest birds, particularly those that rely on cavities for nesting or require stable, resource-rich environments [
81]. Its lower ecological contribution may be attributed to its comparatively limited ability to provide trophic and structural resources, reinforcing the idea that forest stands dominated by older, larger trees are critical for maintaining complex and ecologically functional bird communities.
5. Conclusions
In this study, we applied a multidisciplinary approach to investigate the interaction between soundscape characteristics and vegetation structure within a protected area affected by a nearby highway. Using eight ecoacoustic indices, we quantified the diversity and complexity of soundscapes at eight sites, revealing a clear distinction between areas exposed to anthropogenic noise and those immersed in denser vegetation. Sites with higher biomass and structurally complex forests, such as sites 4 and 9, showed high acoustic diversity indices, suggesting that mature forests not only provide critical habitat for diverse faunal assemblages but also buffer the propagation of anthropogenic noise. In contrast, areas characterized by simpler vegetation structures and lower biomass, such as sites 2 and 8, were associated with reduced biophony, highlighting the role of habitat quality in shaping acoustic biodiversity.
These results highlight the ability of vegetation to modulate the soundscape by influencing both noise propagation and ecological niche availability. Key structural attributes, such as basal area, tree species richness, and biomass, emerged as primary drivers of ecoacoustic variability, reinforcing the idea that the size of individuals of different habitats plays a fundamental role in supporting biodiversity. In particular, the presence of large tree species, such as Quercus robur, was positively associated with avian species richness, supporting the hypothesis that structural complexity and individual size are critical determinants of both habitat quality and acoustic complexity.
From a broader perspective, this study highlights the value of ecoacoustics as a non-invasive and scalable tool to assess biodiversity patterns and the ecological impact of human activities. Integrating soundscape analysis with vegetation structure assessments provides a powerful framework for understanding ecosystem dynamics, offering new insights into the interactions between biotic and abiotic factors. This approach has significant potential to guide conservation planning, inform sustainable forest management, and mitigate the effects of environmental degradation. Future research should extend these findings by incorporating long-term monitoring strategies, assessing seasonal variations in ecoacoustics models, and exploring the applicability of these methods in different ecological contexts.
By advancing our understanding of how vegetation modulates the soundscape, this study contributes to a growing body of research that supports the conservation of mature forests as a key strategy for maintaining acoustic biodiversity. In an era of rapid environmental change, promoting interdisciplinary approaches that connect ecoacoustics, forest ecology, and conservation science will be essential to safeguard the integrity and resilience of natural ecosystems.