Albatross research soars with help of AI conservation tools
Award-winning research is using Artificial Intelligence (AI) to support conservation efforts for the endangered Southern Royal Albatross.
Mitchell Rogers, a research fellow at Te Herenga Waka—Victoria University of Wellington, has received a 2025 Satellites for Biodiversity Award from the Connected Conservation Foundation and Airbus Foundation for his research into albatross colony health on Campbell Island, conducted in collaboration with the Department of Conservation (DOC).
Campbell Island, located more than 700 kilometres south of New Zealand, is home to 99 percent of the world’s breeding population of Southern Royal Albatross. But its remote location and unforgiving environment makes field monitoring extremely challenging and expensive.
“These birds live in one of the most isolated places on Earth,” Mitchell explains. “It’s not a place you can just drop into whenever you need new data.
“Traditionally, conservationists have relied on infrequent manual counts, often limited to small, accessible areas, which only capture a fraction of the albatross story.”
The Satellites for Biodiversity Award supports projects that promote the development and use of conservation technologies. As part of the award, Mitchell will receive an ultra-high-resolution image of Campbell Island, where each pixel represents 15 cm, an image that would be costly to obtain commercially.
This will allow him to use high-resolution satellite imagery, enhanced by AI, to better track the albatross population.
Based at the University's Te Whiri Kawe—Centre for Data Science and Artificial Intelligence, Mitchell’s research uses deep learning (AI) and high-resolution satellite imagery to detect nests, estimate population sizes, and assess breeding success—three key indicators for species health.
“The goal of the project is to develop AI models that can automate population counts for this species using satellite imagery. In collaboration with DOC, we’ll align high-resolution satellite images with GPS coordinates of nest locations, which will be collected in the summer of 2025/26.”
By aligning satellite imagery with GPS data from known nest sites, Mitchell hopes to train AI to better distinguish nests from similar looking features in the landscape.
“To my knowledge, this is the first time GPS data from nesting animals will be used to train AI models for population counts, using satellite imagery.
“The tiny white dots that mark nests can be hard to tell apart from rocks and other ground features. By training the model with nest locations, we’ll assess patterns in breeding. Some nests will be successful, some not.”
Over time, these surveys will help Mitchell track population trends and better understand the overall health of the species, which may be influenced by factors like climate change or fishing bycatch.
With 15 of the world’s 22 albatross species edging toward extinction, research and conservation efforts such as Mitchell's research is essential.
Recognised as taonga by the Kāi Tahu iwi, albatross hold a deep cultural and ecological importance. By automating population monitoring, Mitchell’s project offers a scalable, cost-effective alternative to traditional fieldwork, making long-term monitoring more accessible and achievable.
While focused on the Southern Royal Albatross and its habitat at Campbell Island, the approach has broader potential. It could be adapted to support conservation of other species around the world, contributing to global efforts to project biodiversity and preserve both natural and cultural heritage for future generations.