8:00 - 20:00

Mon - Sun

+351 918 529795

Seacolors Expeditions

Research & Information

Data input and statistics UNDER CONSTRUCTION

Sperm Whale Project

One of the primary objectives of the Sperm Whale Project is to enhance knowledge of sperm whale social groups in the waters surrounding São Miguel, thereby providing a robust scientific basis for their conservation.
The project aims to document the distribution and abundance of sperm whales around São Miguel Island and to expand and refine the photo-identification (Photo-ID) catalogue for the Azores. This catalogue will be systematically compared with Photo-ID catalogues from other islands, with a particular focus on the Macaronesia region, in order to improve understanding of large-scale distribution patterns and to identify key habitats and areas of ecological importance for sperm whales observed around the island.

Further objectives include estimating the spatial range and area use of encountered sperm whale social units, as well as documenting habitat use, activity patterns, and spatial distribution, with special emphasis on surface behaviour. Behavioural observations will be correlated with simultaneously recorded acoustic data to relate activity and behaviour to sound production.

The project also seeks to collect detailed information on the natural behaviour of sperm whales at both individual and group levels, including feeding strategies and prey composition. Morphometric data will be analysed in relation to age class and sex, and body condition assessments will be conducted to evaluate overall health status.

Additional goals include determining group size, group composition, and clan structure of sperm whale units present in the study area, as well as estimating re-sighting rates. The effects of human disturbance will be assessed by examining changes in sperm whale activity and behaviour in the presence of research vessels and drones, taking into account variables such as distance, movement patterns, altitude, speed, and angle of approach.
Finally, the project aims to contribute to a broader understanding of whale behaviour and population dynamics in the region.

Satellite Project

Project Objectives and Scientific Background


The primary objective of this satellite-based research project is the development of a robust and scalable algorithm for the automated detection of whales in the waters surrounding the Azores, including geographically remote and difficult-to-access regions. The approach is based on the analysis of high-resolution satellite imagery and aims to significantly enhance large-scale marine mammal monitoring capabilities.

Beyond mere detection, the project integrates whale occurrence data with additional environmental and atmospheric information derived from the Copernicus Sentinel satellite missions. These complementary datasets include, among others, sea surface temperature, air quality parameters, ozone concentration, solar radiation, and broader climate indicators. By combining biological observations with physical and atmospheric variables, the project seeks to establish a comprehensive framework for analyzing whale distribution and behavior in relation to environmental dynamics.



Methodological Approach


High-resolution satellite imagery represents a promising solution for monitoring whale populations over large spatial scales. However, this approach generates vast amounts of data, rendering manual image inspection infeasible. To address this challenge, advanced machine learning techniques are employed to automate the detection process. These algorithms are designed to identify image segments with a high probability of containing whales, thereby enabling efficient and objective large-scale analyses.



Machine Learning and Model Development


Within the Seacolors framework, multiple machine learning models have been evaluated for their suitability in detecting cetaceans in satellite imagery. Initial experiments indicate that traditional regression-based approaches, such as Ridge Regression, are not well suited to this classification task. In contrast, classification-oriented methods demonstrate substantially better performance.

In particular, Support Vector Classifiers (SVC) and Convolutional Neural Networks (CNN) have proven to be the most effective models for whale detection. While SVCs offer robust classification capabilities for well-structured feature spaces, CNNs are especially powerful due to their ability to automatically learn complex spatial patterns directly from image data. These models form the methodological core of the project’s detection pipeline.

Outlook and Global Applicability


Based on the results obtained in the Azores region, the developed model will be progressively expanded to cover the wider North Atlantic. In the long term, the methodology is intended to be transferable to a global scale. This will enable comprehensive analyses of marine mammal behavior under the influence of climate change and increasing human activities, such as shipping, fishing, and offshore development.

By providing a scalable, satellite-based monitoring tool, the project contributes to both marine conservation efforts and the broader scientific understanding of how environmental change impacts marine ecosystems.

AI Fin Recognation

Sea Colors Expeditions has developed an artificial intelligence–based framework for the automated photo-identification of cetaceans. The system integrates computer vision and machine-learning algorithms to analyse high-resolution photographic data of key morphological features, including flukes, dorsal fins, pigmentation patterns, and naturally occurring scars. These features are extracted and quantified to generate individual-specific visual signatures, enabling robust discrimination between conspecific individuals.

As part of long-term, systematic field research, an extensive and continuously expanding photo-identification database has been established and curated. This database provides the empirical foundation for training, validating, and refining the AI models, and supports longitudinal tracking of individual cetaceans as well as population-level monitoring across temporal and spatial scales.

The AI-assisted identification framework substantially increases the efficiency and scalability of photo-identification workflows by automating the matching of new photographic records against existing catalogues through pattern-recognition techniques. This approach significantly reduces manual processing time and observer bias, while preserving the non-invasive nature of photo-identification as a standard method in cetacean research and conservation.

APP on Sea Project

From the very first research trip, data collection was conducted using standardized paper datasheets and well-trained captains, who systematically recorded all daily sightings as well as the observed behaviour. These records were carefully compiled throughout the field season, and at the end of each season all data were manually transferred into a central database, sheet by sheet and trip by trip. Although this approach ensured a high level of data reliability, it was time-consuming and limited the immediate availability of the collected information.

With ongoing technological advancements and the increasing demand for real-time data in our research activities, the requirements for data acquisition have evolved significantly. To address these challenges, we have replaced traditional paper datasheets with a custom-developed digital application. This application enables the standardized and comprehensive recording of sightings, thereby minimizing the risk of missing relevant information and ensuring greater consistency across observers. In addition, the new system allows for the integration of supplementary environmental data, such as weather conditions and sea state, which are automatically retrieved via web services. As a result, the overall data quality, efficiency of collection, and timeliness of availability have been substantially improved, supporting more robust and responsive scientific analyses.

Whales Acoustic

This project aims to better understand cetacean distribution and habitat use in our waters, in response to the pressing need for baseline information on cetaceans in this data-poor area. This Passive Acoustic Monitoring project aims to fill important knowledge gaps and specifically investigate the temporal and spatial distribution of vocalizing cetaceans and also an opportunity to characterize temporal and spatial variability in noise levels around Whale Watching vessels activity.

Sperm whales are one of the loudest animals in the ocean and on Earth, producing sounds up to 230 decibels. This is louder than jet engines, which are about 150 decibels.

How loud are sperm whale clicks?

  • Sperm whales use clicks for communication and echolocation.
  • These clicks are so powerful that they can blow out human eardrums and vibrate a human body to death.
  • The clicks can travel for miles underwater.

  • How do sperm whales use their clicks?
  • Sperm whales use their clicks to hunt prey like squid in the dark depths.
  • They also use their clicks to keep track of each other during dives.
  • The clicks are so powerful that they can penetrate the flesh of their prey, allowing the whales to see not only where their prey is but how it looks on the inside.

  • How can we learn more about sperm whales?
  • The Sea Colors researchers are recording and analyzing the communication between sperm whales.
  • By listening to sperm whales, we can learn about bigger patterns in deep-sea ecosystems.