Report of the Soundscapes Session

Erbe, C.1*, Southall, B.2*, Baumann-Pickering, S3., Burnham, R.4, Martin, B.5, Parcerisas, C.6, and Picciulin, M.7

1      Centre for Marine Science & Technology, Curtin University, WA, Australia
2      Southall Environmental Associates, Inc., Aptos, CA, USA
3      Scripps Institution of Oceanography, San Diego, CA, USA
4      Institute of Ocean Sciences, Fisheries and Oceans Canada, Sidney, BC, Canada
5      JASCO Applied Sciences, Halifax, NS, Canada
6      Flanders Marine Institute, Ostend, Belgium Scripps Institution of Oceanography, San Diego, CA, USA
7      CNR – National Research Council, Institute of Marine Sciences, Venice, Italy

* Session Chairs and Corresponding Authors; E-mail: C.Erbe@curtin.edu.au – brandon.southall@sea-inc.net

This report can be referenced as: Erbe, C., Southall, B., Baumann-Pickering, S., Burnham, R., Martin, B., Parcerisas, C., and Picciulin, M. (2023). Report of the Soundscapes Session, OCEANOISE2023, Vilanova i la Geltrú, Barcelona, Spain, 22-26 May. Retrieved from https://2023.oceanoise.com

__________________________

Introduction

A soundscape is the “characterization of the ambient sound in terms of its spatial, temporal, and frequency attributes, and the types of sources contributing to the sound field” (ISO 2017). Studies on ocean soundscapes and corresponding publications are rapidly increasing (Schoeman et al. 2022). Oceanoise has had a special session on soundscapes at all of its meetings, with the most recent one in 2017. The 2017 Round Table discussions highlighted several emerging needs for soundscape study.

1) Need agreement on the spatial and temporal resolution of soundscape recording and analysis. Since 2017, the Consortium on Ocean Leadership make recommendations ([COL] Consortium for Ocean Leadership 2018), as did a workshop convened by the IQOE (Ainslie, de Jong et al. 2019). These recommendations were based in part on the work of a project in the United States (ADEON, (Ainslie, Miksis-Olds et al. 2018)) and Europe (JOMOPANS, (Wang and Robinson 2020)). The fundamental recommendation of these groups was to provide spectral and decidecade sound levels with a temporal resolution of 1 minute. These recommendations are forming the basis of ISO Standard 7605 that is currently in preparation.

2) Need better ways to visualize, quantify, and compare soundscapes. Fortunately, the soundscape community has now got access to common and free tools (e.g., Merchant et al. 2015; https://sourceforge.net/projects/pamguide/).

3) Link soundscape measurements to ecology. As examples, the 2017 Round Table discussed acoustic indices, which are common in the terrestrial soundscape community (e.g., Sueur et al. 2014). We tried to have a presentation on these indices applied to underwater soundscapes at Oceanoise 2023, but failed to find a willing speaker. Several people have tried these indices under water, but voiced common concerns (e.g., long-range propagation under water leads to long listening ranges and hence soundscapes being affected by far-away rather than local source presence and ecosystem features).

4) Move soundscape study towards the “impact space”; for example, study long-term behavioural responses, communication space, and masking. We believe steady progress is being made on this front.

 

Presentations

With regards to Need 1, Bruce Martin presented on the “Essential Ocean Variable: Sound” effort of the International Quiet Ocean Experiment (IQOE) community and on the ISO Working Group aiming to standardise passive acoustic monitoring (recording, calibration, analysis, and reporting). One-minute analysis windows and both decidecade and decade band frequency resolutions are suggested by IQOE and ISO.

With regards to Need 2, Simone Baumann-Pickering presented on the U.S. Sanctuary Soundscape Monitoring Project (SanctSound; https://sanctsound.portal.axds.co/). This project has developed an online data portal and graphical user interface, whereby users can select a sanctuary, listen to recordings, view statistics of spectral levels, and browse signal detections (e.g., whales and fishes).

Also with regards to Need 2, Clea Parcerisas presented the challenges of using convolutional neural networks (CNN) to detecting and classifying signals in soundscape recordings. Recordings typically suffer from ongoing noise (i.e., the majority of time is occupied by noise rather than signals) and signals mostly occur at low signal-to-noise ratio (in particular if far away) (Stowell, 2022). Clea discussed ways to improve automated signal analysis and the best ways to test and report these models.

With regards to Needs 3 and 4, Rianna Burnham presented how ship noise reduces the active acoustic space of killer whales. She presented several soundscape metrics (e.g., the proportion of time acoustic space is reduced and the proportion of an area in which active space is reduced to certain amounts) to quantify masking. Masking (in different frequency bands, times of day, and areas) was linked to impacts on navigation, foraging, and social cohesion. The metrics may inform mitigation and management.

Further with regards to impact and management needs, Marta Picciulin presented on the Interreg Italy–Croatia Soundscape project, which has been acoustically monitoring several sites in the Adriatic Sea. Several soundscape metrics were presented, including those suggested by the European Marine Strategy Framework Directive Descriptor 11.2: one-third octave band levels at 63 Hz and 125 Hz, and the percent of area in which these levels exceed 100 dB re 1 µPa each month—as part of the definition of “good environmental status”.

Additional soundscape studies were presented in the short talk and poster sessions.

Discussions

The Round Table discussions picked up on the presentations. Progress is being made in terms of standardising soundscape measurements, so that soundscapes may be compared or synthesised. Data management concerns were raised, as the amount of data being recorded increases rapidly, and there is always a trade-off between storage and resolution. The need for big open datasets with the most common sound types for classification was pointed out (Parsons et al., 2022, Looby et al., 2022). There has been an improvement in available tools for soundscape quantification and visualization such as PAMGuide (Merchant et al., 2015). Analysis methods not requiring extensive human effort are necessary, and there seems to be a trend towards unsupervised approaches. Furthermore, calibration issues are still not completely solved, so absolute levels are not always comparable. The group then discussed whether we are really making the most of soundscape studies or whether outputs remain descriptive rather than quantitative. Additionally, the group asked if we should try to describe soundscapes from the animal perspective, focus on other species rather than only mammals, and add context to the soundscape analysis. Can we do better in terms of outcomes, usages, and impacts?

Included in this discussion was Bruce’s call for additional data to understand how well the suggested standardized methods of analyses are applied to different areas, times and data sets, and if these should be adjusted based on the area or species of focus. Also, discussions on how data collected without calibration could be calibrated post-collection using wind data and other validating methods were introduced by Simone as a means to add to the available data and make them comparable. Bruce also suggested the use of Manta as an open-source tool that would standardise measures of the soundscape.

How we can move from expressing an absolute quantification of sound level change into a more biologically relevant measure when discussing how changes in soundscape might be impactful to marine species was a topic of discussion. The use of species-centred masking metrics was discussed as one possible means to present a more easily understandable quantitative change in soundscape, while perhaps borrowing more from the definition of soundscape for humans, which is the acoustic environment as it is perceived or understood. In particular, using the changes in acoustic active space as a means to present data to management and policy makers, and allow the study of soundscapes to be better applied in conservation efforts for a species. The question of how well this approach had been received so far was asked. This discussion included making the results of soundscape studies more accessible and readily understandable to non-acousticians (rather than purely, say, discussing as changes in sound levels in decibels), in a way that might support ecological studies and assessments of impact. Standardisation of measures and considering proportional changes in active space could provide metrics that allow comparison over time and space. Also, the idea of using acoustic masking and active space to better understand the niche habitat with which an animal is actively engaged was introduced.

Much of the work of the session was focused on the commercial additions to the soundscape, but the potential impact of recreational vessels that are harder to track and quantify was also discussed. Whale-watching vessels will follow whales and expose whales to noise for extended periods (12 hours or more daily), and so the temporal scale of impact needs to be considered in impact assessments.

Bibliography

  • Ainslie, M. A., C. A. F. de Jong and J. L. Miksis-Olds (2019). IQOE Workshop Report: Guidelines for Observation of Ocean Sound. Den Haag, Netherlands.
  • Ainslie, M. A., J. L. Miksis-Olds, S. B. Martin, K. D. Heaney, C. A. F. de Jong, A. M. von Benda-Beckmann and A. P. Lyons (2018). ADEON Underwater Soundscape and Modeling Metadata Standard, Version 1.0. Technical report by JASCO Applied Sciences for ADEON Prime Contract No. M16PC00003:
  • [COL] Consortium for Ocean Leadership (2018). Report of the Workshop on Recommendations Related to Passive Ocean Acoustic Data Standards, Washington, DC.
  • International Organization for Standardization (ISO) (2017). Underwater acoustics—Terminology (ISO 18405). Geneva, Switzerland.
  • Looby A, Riera A, Vela S, Cox K, Bravo S, Rountree R, Juanes F, Reynolds LK, Martin CW. 2022. FishSounds. http://www.fishsounds.net, version 2.
  • Merchant, N. D., Fristrup, K. M., Johnson, M. P., Tyack, P. L., Witt, M. J., Blondel, P., and Parks, S. E. (2015). Measuring acoustic habitats. Methods in Ecology and Evolution, 6, 257-265, doi: 10.1111/2041-210X.12330.
  • Miksis-Olds, J.L., Heaney, K.D., Martin, B., Hawkins, A., Širović, A., Heise, K., Kaplan, M., and Mennitt, D.J. (2015). Report of the Soundscapes Session, oceanoise2017, Vilanova i la Geltrú, Barcelona, Spain, 10-15 May. (Editors Michel André and Peter Sigray). Retrieved from http://2017.oceanoise.com
  • Miksis-Olds, J. L., P. J. Dugan, S. B. Martin, H. Klinck, D. K. Mellinger, D. A. Mann, D. W. Ponirakis and O. Boebel (2021). “Ocean Sound Analysis Software for Making Ambient Noise Trends Accessible (MANTA).” Frontiers in Marine Science 8.
  • Parsons MJG, Lin T-H, Mooney TA et al. (2022) Sounding the call for a global library of underwater biological sounds. Front Ecol Evol 10:810156. https://doi.org/10.3389/fevo.2022.810156
  • Schoeman, R. P., Erbe, C., Pavan, G., Righini, R., and Thomas, J. A. (2022). Analysis of soundscapes as an ecological tool. In C. Erbe and J. A. Thomas (Eds.), Exploring Animal Behavior Through Sound: Volume 1: Methods (pp. 217-267). Cham: Springer International Publishing, doi: 10.1007/978-3-030-97540-1_7.
  • Sueur, J., Farina, A., Gasc, A., Pieretti, N., and Pavoine, S. (2014). Acoustic indices for biodiversity assessment and landscape investigation. Acta Acustica United with Acustica, 100(4), 772-781, doi: 10.3813/AAA.918757.
  • Stowell D. 2022. Computational bioacoustics with deep learning: a review and roadmap. PeerJ 10:e13152 https://doi.org/10.7717/peerj.13152
  • Wang, L. and S. Robinson (2020). JOMOPANS standard: Terminology for ambient noise monitoring. The Netherlands, Version 2.0. Report by the National Physical Laboratory for JOMOPANS. Deliverable/Task: 3.1: