Invasive evergreen shrubs reshape how Welsh woodland birds feed, acoustics reveal

In the broadleaved woodlands of Wales, two non-native evergreen shrubs—Rhododendron ponticum and Prunus laurocerasus—have quietly transformed the understory, forming dense layers of foliage that shade out native vegetation and block the natural regeneration of trees. A new pilot study published in NeoBiota suggests that this structural makeover is not silencing the woodland soundscape, but it is subtly reshaping which birds are most active where the invaders take hold. Using an innovative combination of passive acoustic monitoring and citizen science, researchers detected 28 bird species from 17 families across invaded and non-invaded plots, and found that the overall volume of bird calls barely changed between the two habitat types. Yet when the team examined the birds through the lens of their functional traits—what they eat, where they sit in the food chain, and how they live—the picture changed dramatically, revealing that omnivorous birds produced more than twice as many calls in invaded plots than in comparable control areas.

The study is the first of its kind to apply bioacoustic data to the question of how invasive alien plants influence bird communities, a gap that has persisted even as invasive species are recognised as one of the five major direct drivers of biodiversity decline worldwide. Birds are widely treated as bioindicators of ecosystem health because of their sensitivity to environmental change, and invasive alien predators and reptiles have been implicated in 74 documented global bird extinctions, mainly on islands. Far less attention has been paid to invasive plants, whose impacts on birds tend to be indirect: they alter forest structure, change the availability of food resources, and sometimes provide new feed sources that certain species exploit—and even facilitate the spread of. Understanding these responses requires investigating the functional traits that mediate how bird species interact with altered habitats, the researchers argue, because specialists with narrow feeding habits or habitat preferences are typically more vulnerable to environmental alteration than generalists.

To capture those responses, the team turned to passive acoustic monitoring, a conservation technology that has transformed rapid biodiversity assessment over broad spatial and temporal scales. Sound recorders overcome many of the logistical challenges of traditional point counts and line-transect surveys, collecting large acoustic datasets in a less intrusive, faster and cheaper way. Birds are ideal subjects for such studies because each species has distinct vocalisations that machine-learning software can identify from sound spectrograms. The specific hardware was the AudioMoth, a cost-effective, easy-to-use open-source recorder that has made it increasingly feasible to fold acoustic monitoring into citizen-science projects—much as smartphone applications advanced biological recording more than a decade ago.

Five volunteer citizen scientists, recruited through events run in collaboration with the North Wales Wildlife Trust for people interested in invasive species management, suggested suitable woodland sites close to where they lived and deployed the AudioMoths themselves. Invaded plots were defined by the presence of R. ponticum and/or P. laurocerasus covering more than 50 percent of the vegetation within a minimum 3-by-3-metre area surrounding the recorder, while control plots were placed in the same woodland at least 500 metres away to avoid overlap of bird calls. The devices were configured by the volunteers using the AudioMoth Configuration App at a sample rate of 192 kHz, programmed to record in hourly intervals for six hours each day—between 05:00 and 08:00 and between 17:00 and 20:00—to capture the peak periods of bird activity. Recordings ran from late June to September 2023 across four woodlands, including Nercwys Forest, Penllergare in Swansea, the Wye Valley and AberKenfig, and after recordings affected by noise sources, nearby infrastructure or adverse weather were excluded, the survey effort amounted to 480 hours of audio.

The analysis pipeline relied on BirdNET, a machine-learning application that identifies bird species from audio recordings by analysing sound spectrographs, providing both species names and the frequency of calls per hour. Automated identifications were verified by the researchers and some of the volunteers, and unidentified calls were discarded. The detected species were then classified by trophic level—carnivore, herbivore or omnivore—by trophic niche—granivore, invertivore, omnivore or vertivore—and by primary lifestyle—generalist, insessorial or terrestrial. Trophic level describes a species’ position in the food chain and its contribution to energy transfer, while trophic niche offers a deeper view of its specific dietary resources and ecological interactions, and the distinction proved crucial to the study’s findings.

Statistically, the team aggregated call counts at the plot level and fitted generalised linear mixed models. An initial Poisson model showed strong overdispersion, so the data were refitted with a negative binomial error structure, and candidate models were ranked by automated model selection using AICc. The best-supported model included an interaction between plot type and trophic niche, with site as a random factor, and a Type II Wald chi-squared analysis of deviance confirmed the significance of that interaction. Overall species diversity was similar between control plots, with 26 species, and invaded plots, with 25, and overall call frequency did not differ between the two plot types. But the interaction term told a different story: omnivores produced an estimated 8.49 calls in invaded plots against 3.95 in control plots, a difference significant at p < 0.001, while invertivores showed a marginal decrease from 6.37 to 4.87 calls, granivores and vertivores showed no differences, and neither trophic level nor primary lifestyle produced any plot-specific effects.

The elevated acoustic activity of omnivores in invaded plots suggests these birds may benefit from the altered structure or resources that invasive evergreen shrubs provide, potentially increasing foraging opportunities in habitats where a wide variety of food resources is available. The pattern is consistent with earlier findings from other invasive systems: increased Lantana camara density has been shown to raise the number of frugivorous birds in invaded forests, and liana abundance has been linked to avian acoustic communities in tropical dry forests. The presence of R. ponticum may also indirectly affect avian foraging dynamics by altering the availability of rodents. Conversely, the slight decrease in invertivore call frequency may point to reduced insect availability beneath the dense evergreen canopy, a mechanism consistent with studies showing that invasive plants can depress invertebrate resources and the activity of insectivorous birds. Granivores and vertivores appear largely unaffected, and the absence of differences across trophic level and primary lifestyle suggests those broader traits may buffer species against the structural and resource changes the shrubs impose—echoing reports that Sosnowsky’s hogweed differently affected various foraging guilds of birds in its invaded range.

Several species were recorded in only one plot type: the Common Redstart and Eurasian Jackdaw appeared exclusively in control plots, while the Stock Dove, Eurasian Moorhen and European Goldfinch were detected only in invaded plots, though the study was not designed to evaluate species-specific impacts given its small sample size. The citizen-science dimension delivered both promise and lessons. Interest in the audio-recording technology was strong at volunteer outreach events, but the complexity of the tasks meant that sustained participation was limited—one volunteer returned no recordings at all, leaving data from four sites. The researchers suggest that future projects should budget for high levels of essential training and continued engagement with volunteers, particularly for configuring the AudioMoth, recharging equipment and transferring data, and they recommend longer project durations, widening recruitment to hobby ornithologists already aligned with the topic, and gamification and incentivisation options to keep participants committed.

The study also carries a striking economic footnote: Eryri National Park alone spent around £2.2 million between 2018 and 2024 managing R. ponticum, one of the invasive alien plants with the most severe impacts in Britain, where it shades out native vegetation and prevents woodland regeneration. The impacts of P. laurocerasus are reported to be similar but far less well documented, and the researchers believe the acoustic differences they observed are caused mainly by the dense shrub layer the two species form, altering habitat structure in ways few native shrubs would replicate. Their results add to the growing literature on R. ponticum impacts in Britain and Ireland while providing the first documented acoustic evidence of similar effects from P. laurocerasus.

As a pilot, the study acknowledges its limits—volunteer-driven site selection could introduce bias, the plots were few and small, and surveying was confined to a single season—yet the consistent patterns observed across plots provide meaningful insights, and the team calls for larger sample sizes, more extensive spatial coverage, standardised site selection and year-round surveying to capture seasonal dynamics. The core message is viral in its simplicity: the woodland may sound the same, but it is not functioning the same. Bioacoustic monitoring, the researchers conclude, is an effective tool for investigating invasive species impacts and can be integrated with citizen science to assess how invasive evergreen shrubs influence bird communities—offering a scalable, participatory way to hear the hidden consequences of biological invasion before they become visible.

Subject of Research: Using bioacoustics to assess the impact of evergreen invasive plants on bird communities in Wales, United Kingdom

Article Title: Using bioacoustics to assess the impact of evergreen invasive plants on bird communities in Wales, United Kingdom

Article References: Adekola, O. E., Jones, T. S., & Dehnen-Schmutz, K. (2026). Using bioacoustics to assess the impact of evergreen invasive plants on bird communities in Wales, United Kingdom. NeoBiota, 107, 89-103. https://doi.org/10.3897/neobiota.107.157629

Image Credits: AI Generated

DOI: 10.3897/neobiota.107.157629

Keywords: bioacoustics, assess, impact, evergreen, invasive, plants, bird, communities, Wales, United, Kingdom, scientific research