Fish sounds and boat noise are prominent soundscape contributors in an urban European estuary

Systematic review of machine learning methods applied to ecoacoustics and soundscape monitoring

Soundscape ecology is a promising area that studies landscape patterns based on their acoustic composition. It focuses on the distribution of biotic and abiotic sounds at different frequencies of the landscape acoustic attribute and the relationship of said sounds with ecosystem health metrics and indicators (e.g., species richness, acoustic biodiversity, vectors of structural change, gradients of vegetation cover, landscape connectivity, and temporal and spatial characteristics). To conduct such studies, researchers analyze recordings from Acoustic Recording Units (ARUs). The increasing use of ARUs and their capacity to record hours of audio for months at a time have created a need for automatic processing methods to reduce time consumption, correlate variables implicit in the recordings, extract features, and characterize sound patterns related to landscape attributes. Consequently, traditional machine learning methods have been commonly used to process data on different characteristics of soundscapes, mainly the presence-absence of species. In addition, it has been employed for call segmentation, species identification, and sound source clustering. However, some authors highlight the importance of the new approaches that use unsupervised deep learning methods to improve the results and diversify the assessed attributes. In this paper, we present a systematic review of machine learning methods used in the field of ecoacoustics for data processing. It includes recent trends, such as semi-supervised and unsupervised deep learning methods. Moreover, it maintains the format found in the reviewed papers. First, we describe the ARUs employed in the papers analyzed, their configuration, and the study sites where the datasets were collected. Then, we provide an ecological justification that relates acoustic monitoring to landscape features. Subsequently, we explain the machine learning methods followed to assess various landscape attributes. The results show a trend towards label-free methods that can process the large volumes of data gathered in recent years. Finally, we discuss the need to adopt methods with a machine learning approach in other biological dimensions of landscapes.

Detection of invasive fish species with passive acoustics: Discriminating between native and non-indigenous sciaenids

Invasive alien species have been rising exponentially in the last decades impacting biodiversity and ecosystem functioning. The soniferous weakfish, Cynoscion regalis, is a recent invasive sciaenid species in the Iberian Peninsula and was first reported in the Tagus estuary in 2015. There is concern about its possible impacts on native species, namely the confamiliar meagre, Argyrosomus regius, as there is overlap in their feeding regime, habitat use, and breeding behaviour. Here, we characterised the sciaenid-like sounds recently recorded in the Tagus estuary and showed that they are made by weakfish as they have similar numbers of pulses and pulse periods to the sounds made by captive breeding weakfish. We further demonstrate that breeding grunts from weakfish and the native sciaenid, recorded either in captivity or Tagus estuary, differ markedly in sound duration, number of pulses and pulse period in the two species, but overlap in their spectral features. Importantly, these differences are easily detected through visual and aural inspections of the recordings, making acoustic recognition easy even for the non-trained person. We propose that passive acoustic monitoring can be a cost-effective tool for in situ mapping of weakfish outside its natural distribution and an invaluable tool for early detection and to monitor its expansion.

Impact of small boat sound on the listening space of Pempheris adspersa, Forsterygion lapillum, Alpheus richardsoni and Ovalipes catharus

Anthropogenic stressors, such as plastics and fishing, are putting coastal habitats under immense pressure. However, sound pollution from small boats has received little attention given the importance of sound in the various life history strategies of many marine animals. By combining passive acoustic monitoring, propagation modelling, and hearing threshold data, the impact of small-boat sound on the listening spaces of four coastal species was determined. Listening space reductions (LSR) were greater for fishes compared to crustaceans, for which LSR varied by day and night, due to their greater hearing abilities. Listening space also varied by sound modality for the two fish species, highlighting the importance of considering both sound pressure and particle motion. The theoretical results demonstrate that boat sound hinders the ability of fishes to perceive acoustic cues, advocating for future field-based research on acoustic cues, and highlighting the need for effective mitigation and management of small-boat sound within coastal areas worldwide.

Boat noise impacts Lusitanian toadfish breeding males and reproductive outcome

Anthropogenic noise is a growing threat to marine organisms, including fish. Yet very few studies have addressed the impact of anthropogenic noise on fish reproduction, especially in situ. In this study, we investigated the impacts of boat noise exposure in the reproductive success of wild Lusitanian toadfish (Halobatrachus didactylus), a species that relies on advertisement calls for mate attraction, using behavioural, physiological and reproductive endpoints. Two sets of artificial nests were deployed in the Tagus estuary and exposed to either ambient sound or boat noise during their breeding season. Toadfish males spontaneously used these nests to breed. We inspected nests for occupation and the presence of eggs in six spring low tides (in two years) and assessed male vocal activity and stress responses. Boat noise did not affect nest occupation by males but impacted reproductive success by decreasing the likelihood of receiving eggs, decreasing the number of live eggs and increasing the number of dead eggs, compared to control males. Treatment males also showed depressed vocal activity and slightly higher cortisol levels. The assessment of oxidative stress and energy metabolism-related biomarkers revealed no oxidative damage in noise exposed males despite having lower antioxidant responses and pointed towards a decrease in the activity levels of energy metabolism-related biomarkers. These results suggest that males exposed to boat noise depressed their metabolism and their activity (such as parental care and mate attraction) to cope with an acoustic stressor, consistent with a freezing defensive response/behaviour. Together, our study demonstrates that boat noise has severe impacts on reproductive fitness in Lusitanian toadfish. We argue that, at least fishes that cannot easily avoid noise sources due to their dependence on specific spawning sites, may incur in significant direct fitness costs due to chronic noise exposure.

Boat-induced pressure does not influence breeding site selection of a vulnerable fish species in a highly anthropized coastal area

The brown meagre (Sciaena umbra) is a vulnerable vocal fish species that may be affected by boat noise. The breeding site distribution along the anthropized Venice sea inlets was investigated, by using the species' chorusing activity as a proxy of spawning. Passive acoustic campaigns were repeated at 40 listening points distributed within the three inlets during three-time windows in both summer 2019 and 2020. The role of temporal, morphological, and hydrodynamic variables explaining the observed distribution patterns was evaluated using a GLM approach, considering also human-induced pressures among the candidate predictors. The GLM analysis indicates a higher probability of recording S. umbra chorus after sunset in deeper areas of the inlets, characterized by low water current, while the underwater noise overlapping the species' hearing range and boat abundance did not play any role. This suggests that the species' breeding site choice in the inlets was not influences by boat-induced pressure.

Boat noise impacts early life stages in the Lusitanian toadfish: A field experiment

Marine traffic is the most common and chronic source of ocean noise pollution. Despite the evidence of detrimental effects of noise exposure on fish, knowledge about the effects on the critical early life stages - embryos and larvae - is still scarce. Here, we take a natural habitat-based approach to examine potential impacts of boat noise exposure in early life stages in a wild fish population of the Lusitanian toadfish (Halobatrachus didactylus). In-situ experiments were carried out in the Tagus estuary, an estuary with significant commercial and recreational boat traffic. Nests with eggs were exposed to either ambient (control) or boat noise (treatment), for 1 fortnight. Eggs were photographed before being assigned to each treatment, and after exposure, to count number of eggs and/or larvae to assess survival, and sampled to study development and oxidative stress and energy metabolism-related biomarkers. Data concerns 4 sampling periods (fortnights) from 2 years. Results indicate that offspring survival did not differ between treatments, but boat noise induced a detrimental effect on embryos and larvae stress response, and on larvae development. Embryos showed reduced levels of electron transport system (ETS), an energy metabolism-related biomarker, while larvae showed higher overall stress responses, with increased levels of superoxide dismutase (SOD) and DNA damage (oxidative stress related responses), ETS, and reduced growth. With this study, we provided the first evidence of detrimental effects of boat noise exposure on fish development in the field and on stress biomarker responses. If these critical early stages are not able to compensate and/or acclimate to the noise stress later in the ontogeny, then anthropogenic noise has the potential to severely affect this and likely other marine fishes, with further consequences for populations resilience and dynamics.

A Fish and Dolphin Biophony in the Boat Noise-Dominated Soundscape of the Cres-Lošinj Archipelago (Croatia)

Spatio-temporal variability of marine soundscapes reflects environmental dynamics and local habitat health. This study characterizes the coastal soundscape of the Cres-Lošinj Natura 2000 Site of Community Importance, encompassing the non-tourist (11–15 March 2020) and the tourist (26–30 July 2020) season. A total of 240 h of continuous recordings were manually analyzed and the abundance of animal vocalizations and boat noise was obtained; sound pressure levels were calculated for the low (63–2000 Hz) and high (2000–20,000 Hz) frequency range. Two fish sound types were drivers of both seasonal and diel variability of the low-frequency soundscape. The first is emitted by the cryptic Roche’s snake blenny (Ophidion rochei), while the second, whose emitter remains unknown, was previously only described in canyons and coralligenous habitats of the Western Mediterranean Sea. The high-frequency bands were characterized by bottlenose dolphin (Tursiops truncatus) vocalizations, indicating dolphins’ use of area for various purposes. Boat noise, however, dominated the local soundscape along the whole considered periods and higher sound pressure levels were found during the Tourist season. Human-generated noise pollution, which has been previously found 10 years ago, is still present in the area and this urges management actions.