Training 1- Marine soundscapes: methods for the acquisition of ocean sound from fixed and mobile platforms
Trainer: Gianni Pavan – University of Pavia
Bioacoustics and Ecoacoustics are rapidly developing disciplines to study and monitor marine ecosystems by their soundscape composition. This is a worldwide emerging research area aimed at monitoring, and possibly contrasting, the decline of biodiversity impacted by habitat reduction and degradation due to both local human activities and global environmental changes (noise pollution, climate changes and chemical pollution). The acoustic environment and the soundscape have been recognized to be an essential component of ecosystems, thus worth of being studied, monitored, protected, and even restored when altered by human activities. Ecoacoustics joins bioacoustics and ecology as an interdisciplinary science that investigates natural sounds (biological and geophysical) and anthropogenic sounds considering their interaction over a wide range of study scales, both spatial and temporal. Sounds can be both the subject and the tools of ecological research. As subject, sounds are investigated in order to understand their origin, functions, and properties. As tools, sounds are used to study and monitor ecosystems by considering anthropogenic noise and biological sounds that are the expression of animal diversity, abundance, behaviour, dynamics, and distribution. The objective of the course is to provide scholars with a basic foundation to understand bioacoustics and ecoacoustics, the equipment needed to do marine acoustic research and monitoring, the software tools, the applications in the different fields, ranging from basic research to environmental monitoring and protection.
Training 2- Requirements and methods for the production of ocean noise time series; Acoustic data FAIRness, standard formats
Requirements and methods for the production of ocean noise time series
Trainers: Enoc Martinez, Daniel Mihai Toma, Joaquin del Rio
Universitat Politècnica de Catalunya
Underwater ambient sound has been increasing in the past decades due to an increment of human activities such as shipping, seismic exploration and construction. This increment of ambient sound in the form of acoustic noise may have a severe impact on marine life. In order to achieve a good environmental status in European waters it is important to monitor and limit the impact of underwater noise, as stated in the Marine Strategy Framework Directive (MSFD). This training session will be focused on practical solutions for the production, analysis and effective sharing of underwater ambient noise time-series compliant with the MSFD. State-of-the-art technologies and tools for the acquisition and analysis of underwater noise data will be showcased and applied using real data. First, the basic acoustic concepts will be reviewed to ensure the correct handling of data. Then, a hands-on approach to signal processing applied to underwater acoustics will be provided. The session will close with an overview on the data management solutions to ensure data FAIRness (Findable, Accessible, Interoperable, Reusable).
Training 3- Methods for the detection and identification of marine mammals sounds
Trainers: Daniel E. Cline & John Ryan – Monterey Bay Aquarium Research Institute
The MBARI research activity with Blue whales is proposed as the focal content for the EMSO TSC training, because:
* It uses a combination of Machine Learning and signal processing methods, depending on the type of call.
* Blue Whales are a fascinating endangered species for which science can inform protection.
* Apply signal processing methods to discover what can be heard when Blue Whales are migrating (https://www.mbari.org/blue-whale-songs-migration/). The reason these methods (signal: noise, energy detection) are used to quantify song occurrence on key time scales is that there are times when the whales are calling so much that individual calls cannot be distinguished from each other (while the total signal from all calling whales can still be quantified relative to background).
* Demonstrate use of Machine Learning to successfully detect and classify other call types that occur less frequently and thus do not have the above “chorusing” difficulty.
MBARI are proposing to use an existing AWS (Amazon Web Services) Open Data project with both data and tutorials already up and tested in the cloud, for both Machine learning and signal processing methods applied to blue whales. To make the most effective use of a 3-hour period, MBARI suggest that we have participants simply use a web-browser and work in provided Google Colab notebooks – this is the simplest approach to quickly get up and running as it is straightforward to convert any examples in AWS to Google for a tutorial.
Interested in analyzing relevant EMSO partner Acoustic Datasets – particularly Blue Whale Signals in advance of the training to show how methods developed can be applied to other datasets;
Interested possibility of using the EMSO Virtual Research Environment as a support for conducting practical exercises during the training session.
Training 4- Introduction to ocean particle velocity and measurement techniques
Trainers: Sérgio M. Jesus and Paulo J. Santos
ISR-LarSys, University of Algarve, 8005-139 Faro, Portugal
At the macroscopic level the relevant parameter for studying ocean sound is acoustic pressure. However, at the sub-wavelength level and close to shores / acoustic surfaces, the particle velocity field shows significant departures from the acoustic pressure field. So, particle velocity and acoustic pressure are two manifestation of the same phenomena at two different spatial scales. There are two important aspects related to particle motion/velocity: one is that many marine animals, including fish and cetaceans, are extremely sensitive to it, and the other is that while acoustic pressure is scalar, particle velocity is vectorial and therefore allows for determining not only its strength but also the wavefield direction, which is of paramount importance in many underwater applications. This training session explores these two aspects with the objective of providing the necessary understanding of the role of particle velocity to ocean observation, including the impact of sound on marine life. For many years particle velocity measurements in the ocean was impossible or extremely difficult due to either too sensitive sensors picking up noise and interference, or due to bulky or delicate sensors difficult to deploy and operate in the field. Nowadays, the situation has significantly improved with vector sensor measurements being widely accessible, and applications starting to appear in many diverse fields. These will be reviewed with practical example highlights, as well as suggestions for deployment strategies and data analysis techniques.
Training 5- Methods for the detection and analyses of anthropogenic sound in the oceans
Trainers: Salvatore Viola – NATO STO-CMRE & Francesco Simeone – INGV
The goal of the course is to make participants familiar with the main techniques for the analysis of the underwater anthropogenic noise. The main sources of underwater noise will be described in terms of their acoustic features (intensity, frequency spectrum, time variability). Participants will learn to recognize, identify and characterize different classes of acoustic sources by analyzing acoustic data made available for the training.
The course will be divided into two parts: a theoretical(A) introduction, with practical examples, of the basic concepts needed during the course; followed by guided analysis of real data using intermediate level techniques.
A) Fundamentals in signal analysis of passive acoustic data:
-Terminology (e.g., source intensity, source level, transmission lost)
-Time domain representation
-Frequency domain representation
-Noise sources (impulsive, ambiental) and averaging methods (e.g., rms, percentile)
-Noise reduction techniques
-Anthropogenic noise sources (e.g., ships, airguns)
B) Real world examples from Western Ionian SN1 data
-Single source analysis (ship noise): identify the source and its characteristics
-Single source analysis (airgun): using environmental knowledge to gain source information
-Multiple sources analysis (Anthropogenic and natural): disentangle using spectral analysis and/or time series analysis
-Single source analysis (explosion) in case of reflections and reverberations
Knowledge acquired during the course:
-basic knowledge of underwater environment and sound propagation
-basic knowledge of anthropogenic sound sources
-intermediate skill on signal processing
Prerequisite: basic knowledge of signal processing