Passive acoustic monitoring as a tool for offshore wind farm biodiversity assessments: a pilot study

2. Introduction

2.1 The importance of monitoring

Marine biodiversity is rapidly changing due to the direct and indirect impacts of climate change and ocean sprawl, which is the proliferation of human-made structures in the ocean and coastal areas (Duarte 2014).The Intergovernmental Science-Policy Platform on Biodiversity and Ecosystem Service (IPBES) declared a global biodiversity emergency (IPBES, 2019), and it is therefore essential that effective long-term monitoring programmes are established to ensure sustainable development (Borja and Elliott 2021). This is particularly important in an offshore marine environment, which can be difficult and expensive to access and where impacts are not always clearly visible (Nathanael et al. 2024). Due to the dynamic nature of marine ecosystems, it can be challenging to obtain a clear signal to connect human activities (and the associated pressures) with an ecosystem response. On the other hand, some interventions, if done correctly, can potentially enhance biodiversity and improve ecosystem outcomes (Strain et al. 2018). Therefore, monitoring methods should be able to detect when biodiversity increases as well as potential decreases as a route to demonstrating environmental gains, which is especially important in light of recent discussions around Marine Net Gain (Hooper et al. 2021).

Traditional monitoring methods for monitoring the biodiversity associated with benthic habitats include grab sampling, box cores, benthic trawls and visual surveys (remotely operated vehicles, autonomous underwater vehicles and time-lapse images/videos) (Scottish Government report, 2023). Except for time-lapse cameras attached to benthic observation stations and landers (Kutti et al. 2014; Osterloff et al. 2016; De Clippele et al. 2023; Clark et al. 2024), these approaches only provide a snapshot of information at one point in time. Some of these tools can introduce sampling biases and cause damage to sensitive features (Beisiegel et al. 2017). There is a need, therefore, to improve monitoring regimes and investigate the effectiveness of non-invasive cost-effective tools, such as Passive Acoustic Monitoring (PAM).

2.2 Passive Acoustic Monitoring

Passive Acoustic Monitoring is emerging as a tool to assess and monitor changes in aquatic benthic biological communities (Minello et al. 2021; Havlik et al. 2022). Traditionally, in marine ecosystems, PAM has been used to monitor the presence and abundance of marine mammals. However, in recent years, PAM has been proposed and used for the monitoring of benthic habitats at a high temporal resolution and for long periods (months to years) with global efforts rapidly increasing (Mooney et al. 2020; Van Hoeck et al. 2021; Di Iorio et al. 2021; Havlik et al. 2022; Looby et al. 2023; Darras et al. 2024). Sounds produced by vertebrates and invertebrates such as fish, and crustaceans, worms and bivalves contribute to the biophonical component of the acoustic landscape or soundscape, while currents and wind contributes to the geophonical component, and boat noise and other man-made noise contribute to the anthroponical component (Duarte et al. 2021). Characterising the biological component of underwater soundscapes and monitoring changes in acoustic communities composition, diversity and abundance can provide information on animal presence and behaviour (i.e. mating, aggression), as well as on species diversity (Ladich 1997; Fudge and Rose 2009; Desiderà et al. 2019a; Mooney et al. 2020; Di Iorio et al. 2021).

2.3 Benthic soundscapes

Both fish and crustaceans can produce either active or passive sounds (Rountree et al. 2018; Caiger et al. 2020; Raick et al. 2021). Active sounds are here defined as sounds which animals intentionally produce for communication purposes, i.e. all those sounds emitted by a sender which influence the behaviour of a receiver in a fashion that provides fitness-related advantages at least to the sender. For example, cod can vibrate their sonic muscles around their swim bladder, which produces “grunt” sounds used to attract mates during spawning events (Brawn 1961) and spiny lobsters can make “rasp” stridulation sounds to fend off predators (Buscaino et al. 2011). Other sounds, which are here referred to as passive sounds, are non-intentional and can be a by-product of certain behaviours, such as air movement, jaw movement, swimming or food consumption (Rountree et al. 2018; Parsons et al. 2022; Looby et al. 2023). These sounds typically have more variable acoustic characteristics (Rountree et al. 2018). For example scallops (Pecten maximus) produce a “cough” sound while moving their valves (Di Iorio et al. 2012).

Some of these sounds may be more abundant during certain times of the day and year. The circadian cycles (i.e. the daily cycle) of acoustic activity are regulated by changes in light levels generated by the sun or the moon cycles. Times when acoustic activity of benthic species may be increased depend on the species. For examples. The European spiny lobster (Palinurus elephas), is most active at night (Buscaino et al. 2011), while the tropical domino damsel fish (Dascylus albisella) produces more sound during the day (Lammers et al. 2008). Increased acoustic activity can sometimes be observed at dusk and dawn, as well as during new moon periods (e.g., Lammers et al., 2008; Radford et al., 2008). Seasonal patterns are usually associated with fish’s spawning (Lobel 1992, Luczkovich et al 1999, Bolgan et al. 2017, Hawkins and Amorim 2000), and breeding periods (Mann et al 1997). Therefore, analysing the sounds present in offshore marine environments could be a valuable tool for monitoring specific biological processes and identifying both natural and shifting patterns of biological activity across various locations.

2.3.1 Acoustic Community and Phonic Richness

The acoustic community can be defined as the aggregation of soniferous species, and the phonic richness is a measure of the total number of types of sounds made by soniferous species. These acoustic metrics have been used as a proxy for biodiversity and habitat condition (Mooney et al. 2020), with alterations to their spatial or temporal patterns potentially being indicative of changes in community composition or ecological state (Rabin et al. 2022, Sueur et al. 2019, Lamont et al. 2022). For example, differences in acoustic community composition have been found related to variations in water depth and the percentage cover of structure-forming benthic habitats, such as coralligenous reefs, in the north-western Mediterranean (Di Iorio et al., 2021). Habitats that are regarded to be healthy or in better condition tend to have a higher phonic richness, although significant differences are not always found (Lamont et al., 2020, Di Iorio et al., 2021). Visually and aurally analysing the presence and abundance of the different types of sounds is, therefore, a useful approach to providing an understanding of patterns in biodiversity, which is otherwise difficult to monitor.

2.5 Acoustic Indices

Since manually annotating the different types of sounds present in a recording is very time-consuming, the application of computational acoustic indices, initially developed for terrestrial habitats, is increasingly being applied in marine habitats. Rather than investigating patterns of species-specific calls, acoustic indices target diversity assessment at the community level or use spatio-temporal changes of sounds to estimate alterations in ecosystem functionality (Sueur et al., 2018, Pieretti and Danovaro 2020) (Table 1).

Soundscape indices can be used separately or in combination with each other to quantify the complexity of the soundscape in one or more “simple” values (Lillis et al. 2014, Williams et al. 2022). The list of indices used in this study, along with a summary description and the R packages used is provided in Table 1. The acoustic complexity index (ACI) is the most commonly used index and appears to be the most reliable metric for assessing fish diversity, habitat quality and certain key ecosystem functions (Harris et al. 2015, Elise et al. 2019, Minello et al. 2025). The total-, spectral-, and temporal entropy indices are also commonly used in marine habitats (Sueur et al. 2008, Villanueva-Rivera et al. 2011, Depraetere et al. 2012, Minello et al. 2025), with the total entropy appearing to work well for temperate reefs (Harris et al. 2015). The bioacoustics index has been shown to detect fish chorusing to some extent and was found to be important when using a multi-metric approach (Siddagangaiah et al. 2019, Williams et al., 2022). Changes in the amplitude of recordings over time, have been used to indicate the presence of fish choruses (Siddagangaiah et al. 2021) or discriminate between different habitats or habitat states (Elise et al. 2019, Lamont et al. 2021).

Although multiple reviews highlighted the potential of using acoustic indices in monitoring, they also pointed out the need for further studies due to sometimes a lack of a clear relationship between more traditional biodiversity indices, including phonic richness and values generated by acoustic indices (Mooney et al. 2020; Pieretti and Danovaro 2020; Minello et al. 2021). In addition, noise pollution, equipment self-noise fish chorusing, and even geophysical sounds can disrupt the reliability and interpretability of the indices (Parks et al. 2014, Rice et al. 2017, Minello et al. 2021, Nguyen Hong Duc et al., 2021). For example, equipment self-noise can artificially increase ACI values, while fish chorusing can reduce it (Nguyen Hong Duc et al., 2021).

Standardisation of the data collection, while accounting for geophysical drivers and removing or reducing anthropogenic noises have been suggested to improve the reliability of the indices (Minello et al. 2021). Increasing the spatial and temporal sample sizes, along with habitat-specific testing can improve their effectiveness (Minello et al. 2021). Furthermore, it is essential to recognise that acoustic indices do not enable the discrimination of fish community properties, which are crucial for management, such as abundance and diversity (Bolgan et al., 2018; Raick et al., 2023a).

2.6 Offshore wind

Coastal habitats with structure-forming benthic animals and plants, such as coral reefs, kelp and seagrass, have been the primary focus of marine benthic soundscape studies (Havlik et al. 2022). Sedimentary habitats that are characterised by sandy, muddy and coarse sediments are mostly only being used as a control rather than being the focus of previous studies (Ceraulo et al. 2018). Offshore developments typically affect the latter. Using PAM as a monitoring tool could be particularly relevant for the planned global offshore wind farm developments, as these have the potential to change benthic biodiversity at a local and regional scale through the introduction of hard substrates and changes in sediment deposition, food supply and local hydrodynamics (Mangi 2013; Clark et al. 2014; Dannheim et al. 2020). The application of PAM could provide long-term, remote and scalable information on aquatic biodiversity which are not achievable by other monitoring techniques (Mooney et al. 2020; Ross et al. 2023).

The Scottish Government has set a range of targets to cut greenhouse gas emissions and to generate more energy from renewable sources. The Climate Change (Scotland) Act 2019 commits the Scottish Government to reach net zero emissions of all greenhouse gases by 2045. Offshore wind will play a vital part in meeting these targets and is set to expand substantially in Scotland over the next decade and beyond, as evident by the recent ScotWind seabed leasing round by the Crown Estate Scotland. In January 2021 and through its clearing process in August 2022, Crown Estate Scotland offered Option Agreements for 20 offshore wind projects around Scotland, totalling 27.6 Giga Watts (GW) of potential new offshore wind developments. The related offshore wind Innovation and Targeted Oil and Gas (INTOG) decarbonisation planning and leasing process outlined a further 5.4 GW of projects that have been offered exclusivity agreements.

This pilot study set out to demonstrate the usefulness of PAM recordings as a tool to detect spatial changes in offshore benthic biodiversity associated with sedimentary European Nature Information System (EUNIS) classified sandy to muddy and coarse habitat types. It will do this by first assessing which soniferous species could be present in the study sites by extracting data from open source platforms such as the OneBenthic and National Biodiversity Network (NBN), then by establishing if there are differences in benthic acoustic communities associated with the different habitat types, and finally by establishing if passive acoustic biodiversity metrics such as phonic richness and acoustic indices can be used as a proxy for benthic biodiversity using both raw data and data cleaned from noise pollution. This pilot study is a first step towards determining the value of this tool in monitoring programmes and survey requirements for marine renewable energy devices (MREDs) in Scottish waters. The Marine Directorate has been using PAM to monitor cetacean species in Scottish waters over the past 10 years, with click detectors and broadband acoustic recorders. There is therefore the potential to ‘unlock’ data from the comprehensive PAM network to add value to Scottish marine monitoring programmes.