Space race to monitor whales
SCIENTISTS in Scotland are helping to develop new methods to boost the conservation of whales – by monitoring them from space.
The new “Spacewhale” project, developed by creative scientists in Germany and the UK, combines modern space technology and artificial intelligence.
The research enables whales and other large marine mammals to be surveyed at an unprecedented scale by satellite.
Scientists in Scotland helped create an algorithm that identifies the animals in high-resolution satellite images, which can cover the global ocean surface more effectively than monitoring by ship or aircraft.
The technology, supported by funding from the European Space Agency (ESA) Space Solutions program, could help answer questions about where whales go and when, and find solutions that combine their protection with human use of the seas such as determining when oil and gas or offshore wind-farm activities cause the least disturbance to wildlife.
Project manager Caroline Höschle from Bioconsult SH, said: “Earth Observation by satellites is currently developing rapidly.
“It will only take a few more years for space companies to provide highresolution images of the entire globe on a daily basis.
“That makes Spacewhale a forwardlooking tool, but it already performs fantastically well with the imagery we have today.
“Over 70 per cent of the Earth’s surface is covered by water and thus large areas are still unexplored.
“The intelligent use of satellite imagery now brings us a lot of previously inaccessible data.
“Spacewhale is a fast and efficient means of surveying whales on a global scale – at a comparable cost to traditional methods for only a small area of the oceans.”
Uk-based Hidef Aerial Surveying Ltd, including researchers in Edinburgh, developed automatic image recognition to identify whales.
The algorithm was created from aerial images of seven-to-10-metre-long minke whales taken during monthly whale monitoring flights focusing on
offshore wind farms including in the seas off the coast of Scotland.
The algorithm was subsequently able to recognise 23-metre-long fin whales and other whale species on satellite images from around the world.
Lead data scientist Dr Grant Humphries, based in the Highlands, said: “The data used to build the method included images of minke whales in Scottish waters that were used to teach the artificial intelligence how to identify species.
“We have also developed the method by which we review whales in satellite imagery.
“An artificial intelligence algorithm runs through a satellite image and if it depicts an object, our expert team of reviewers decide whether or not it is a whale.
“Really big whales such as blue whales or fin whales, when you see them even in a satellite image, are really obvious. Smaller species are more challenging, and having the expert review process in place allows us to better identify whales down to a species level.”
The application of Spacewhale can help identify critical habitats of whales, and inform marine spatial planning and impact assessments of offshore developments.
Accelerating climate change is also expected to cause profound changes, especially in Arctic waters, with associated declining ice cover and expanding human activities in whale habitat.
Höschle added: “With Spacewhale, we want to contribute to the targeted implementation of protection measures for great whales, even in the most remote areas.”
Rita Rinaldo, of ESA Space Solutions, said: “Spacewhale demonstrates that the possible applications of satellite data are endless and only becoming more so with new satellite capabilities, such as higher-resolution imagery and improved coverage, being made available at an ever increasing rate.
“This project testifies ESA engagement in protecting biodiversity and natural ecosystems, in support of the Green Deal objectives.”