New publication: Elevating photo-identification: Aerial-identification improves re-sight rates and supports long-term monitoring of humpback whales

Written by MMRP PhD student Lewis Evans

We are excited to share our newly published manuscript in Marine Mammal Science.

Titled: Elevating photo-identification: Aerial-identification improves re-sight rates and supports long-term monitoring of humpback whales

Authors:

Lewis I. Evans, Martin van Aswegen, Sonja Feinberg, Jens J. Currie, Stephanie H. Stack, Andrew Szabo, Lars Bejder

Why did we conduct this study?

Not all whales are the same. They differ in size, body condition, behavior, and where they feed or breed — differences that shape survival, reproduction, and how they respond to disturbance. While some studies have managed to explore this individual variability, they usually come with small sample sizes (It turns out that gathering the data for these studies is really tough!) As a result, most research ends up relying on population-level snapshots, averaging across individuals. These studies are enormously valuable for establishing baselines and broad trends, but the trade-off is that averages can blur out the individual variation that really matters — the unique traits and behaviors that can tip the balance for survival, reproduction, or response to change. To better understand how whales may respond to future human or environmental change, we need to move beyond population averages and capture individuals through time. The challenge? Telling “who is who” in the first place…

Traditionally, humpback whales are identified by the unique patterns on the underside of their tail flukes. By using these markings, we can identify individuals across different sightings to know which whales we are sampling. When the same whale is identified on more than one occasion, for example, across days, months, or years, we call this a ‘re-sight’. Re-sights let us follow the same whale across time and space, so that we can track things like changes in blubber mass, as well as monitor growth in calves, or see how whales shift habitats and social groups. All of these individual stories add up to a bigger picture of how populations function.

Figure 1: Tail fluke photograph of a humpback whale in Maui, Hawai‘i. NMFS Permit No: 21476. 

But on the breeding grounds, mothers rarely lift their tails, so historically, we could only capture these fluke identifier photographs (let's call them “fluke-IDs”) in 37% of encounters (and 60% for other whales). We likely encountered the same individuals multiple times, but without IDs, we couldn’t be sure — and every missed ID was a lost chance at a re-sight. To address this, we tested whether drones could help improve re-sight rates by capturing unique features from above, such as the bumps (tubercles) and circular scars left by cookiecutter sharks on their heads.

Figure 2: The image depicts tubercles (orange circle) and cookiecutter shark scars (White circle) on top of the whale's head. Photo credits: Jens Currie

Our objectives

The objectives of this study were to (1) determine whether aerial-ID can reliably distinguish individual humpback whales, (2) assess whether aerial-ID increases both individual and total re-sight rates within a season compared to fluke-ID, and (3) test whether tubercle patterns can be used to support long-term identification.

How did we do it?

During the 2022 whale season in Maui, we flew drones over nearly 1,500 whales and captured top-down screenshots of their heads. From each encounter, we selected clear images that showed two distinctive features: the raised bumps (tubercles) on the rostrum and the circular scars left by cookiecutter shark bites.

We then used a free software program called I3S Classic, which works a bit like facial recognition for whales. Analysts marked a few stable reference points on each image (Figure 2), like the rostrum tip and eyes, and then placed digital markers on the visible bumps or scars so the software could recognize their positions. The program turned these annotations into a digital fingerprint, aligned them with those from other whales in the catalog, and compared how well the features lined up to generate a ranked list of likely matches — from most to least similar (Figure 2). Human reviewers then checked the suggestions to confirm whether each whale was new or a re-sight.

Figure 3: A top-down drone screenshot of a humpback whale for aerial-identification annotation. The reference points for tubercle (TB) annotations are in orange and cookiecutter shark scars (CCS) are in white: (1) medial rostrum, (2) left pectoral insertion point, (3) right pectoral insertion point, (4) left eye, (5) right eye. The dashed lines indicate the annotation regions for TB (orange) and CCS (white). 

To test whether this approach could also work beyond a single season, we expanded the method to 54 whales that had already been matched by fluke-ID across different years and even between Hawai‘i and Southeast Alaska. This allowed us to see how well tubercle patterns hold up for long-term identification.

It worked! Significantly improving our ability to re-sight whales

It worked, and far better than we imagined! During our 86-day study, aerial-ID captured ID images for over 95% of whales, compared to just 37% of mothers and 60% of others with fluke-ID. The number of whales with at least one re-sight rose sharply, up 202% for mothers and 96% for others. And when we added up the total number of re-sights (every repeat encounter across all whales) the increase was 227% for mothers and 40% for others.

These benefits also extended beyond a single season. Across the 2018–2025 period, we re-sighted those 54 humpback whales that had already been verified with fluke-ID, including 28 that moved between Hawai‘i and Southeast Alaska. Using tubercles, aerial-ID successfully re-identified all 54, showing performance comparable to our within-season re-sights. Re-sight intervals ranged from just 7 months to more than 6 years, with a median of 3.5 years, suggesting that tubercles provide consistent markers over long timescales and may remain useful for even longer.

What does this mean? 

Aerial identification gives us a new way to identify whales that were previously slipping through the cracks — especially mothers on the breeding grounds. By boosting re-sight rates and extending reliable matches across years and regions, this method holds potential to strengthen studies that can lend insights into population trajectories and fill critical gaps left by fluke-ID alone. Working as a complement to fluke-ID, aerial-ID broadens our toolkit for monitoring individuals. Essentially, more re-sights mean more chances to understand how individual whales grow, migrate, and respond to a changing ocean. This, in turn, helps us build stronger science to support targeted conservation and management of these animals across their range.

Want to learn more?

This post highlights just a portion of our work. See our abstract below for a quick overview, or dive into the full scope of the study, including limitations, future directions, and a deeper discussion of aerial-ID’s pros and cons, in the complete manuscript in Marine Mammal Science here: 

Evans, L. I., M. van Aswegen, S. Feinberg, J.J. Currie., S. H. Stacket., A. Szabo, and L. Bejder. 2025. “Elevating Photo-Identification: Aerial-Identification Improves Re-Sight Rates and Supports Long-Term Monitoring of Humpback Whales.” Marine Mammal Science e70078. http://doi.org/10.1111/mms.70078

If you’d like to connect or have questions about this publication, feel free to reach out to me at levans4@hawaii.edu

Abstract:

‘Photo-identification (photo-ID) is a widely used, non-invasive method for monitoring individual animals, including humpback whales (Megaptera novaeangliae; HBWs), and has provided valuable insights into their population dynamics, movement patterns, and social structures. Traditional identification relies on the trailing edge and ventral pigment patterns of the tail fluke (fluke-ID); however, not all whales present their flukes, limiting identification and re-sighting opportunities. We developed a novel aerial-identification (aerial-ID) approach using drone imagery to identify individual HBWs based on the arrangement of two features, tubercles (TB) and cookiecutter shark scars (CCS). Between January and March 2022, we sampled 1,498 HBWs, including repeated individuals, capturing fluke-ID images for 772 and aerial-ID images for 1,437. Fluke-ID yielded 164 re-sightings (76 lactating females, 88 others), while aerial-ID yielded 372 (249 and 123, respectively), representing a 227% increase for lactating females and 40% for others. We extended this approach to a multi-year, cross-regional dataset (2018–2025) of 54 individuals verified with fluke-ID. All were matched using aerial-ID, with the longest re-sight spanning 2,737 days (6.5 years), representing the maximum interval within our study period.  Aerial-ID thus offers a powerful complement to fluke-ID, expanding demographic coverage, increasing re-sighting rates, and enabling long-term, cross-regional monitoring.’

In our commitment to making our science readily accessible to the Native Hawaiian community, we provide an abstract of this paper in Hawaiian. By sharing this work in ‘Ōlelo Hawai‘i, we hope to strengthen the connection between our research and the community it serves, reflecting our commitment to both the conservation of these species and the protection of the land and waters they inhabit. We welcome discussion and further engagement from all who wish to connect.

‘He ʻano hana laha a hoʻoluhi ʻole ka hōʻoia kiʻi no ke kilo ʻana i nā holoholona, e like me ke koholā (Megaptera novaeangliae), a hōʻike i ko lākou ʻano lehulehu, kā lakou holo ʻana, me nā pilina ʻohana. Kaukaʻi ka hōʻoia laha i ke ʻano o ke kaʻe o ka hiʻu a me ka waihoʻoluʻu i lalo o ka hiʻu (hōʻoia hiʻu); akā naʻe, ʻaʻole hōʻike ʻia ka hiʻu o nā koholā a pau, e hoʻēmi ana i nā hōʻoia a me nā ʻike hou. Ua kūkulu ʻia kekahi ʻano hana hōʻoia lani e mākou i mea e hōʻoia ai i nā koholā ma o ke ʻano o ʻelua mea, nā puʻu a me nā ʻālina manō cookiecutter. Ma waena o ʻIanuali me Malaki 2022, ʻike ʻia 1,498 mau koholā, a me nā ʻike hou, a paʻi kiʻi ʻia 772 kiʻi hōʻoia hiʻu me 1,437 kiʻi hōʻoia lani. Ua loaʻa 164 ʻike hou (76 wāhine e hānai ana, 88 koholoā ʻē aʻe) mai ka hōʻoia hiʻu, a loaʻa 372 (249 a me 123) mai ka hōʻoia lani, he hoʻonui 227% ma nā wāhine e hānai ana a 40% ma nā koholā ʻē aʻe. Ho’ohana nō hoʻi mākou i kēia ʻano hana ma kekahi papa ʻike o 54 mau koholā i hōʻoia ʻia e ka hōʻoia hiʻu mai 2018-2025 i nā wahi he nui. Hoʻohālikelike ʻia nā koholā a pau e ka hōʻoia lani, a 2,737 lā (6.5 makahiki) ka wā ʻike hou lōʻihi loa, ka wā nui loa ma kā mākou noiʻi. He ukali maikaʻi nō ka hōʻoia lani i ka hōʻoia hiʻu, e hoʻonui ana i ka ʻike ʻana o nā ʻano koholā ʻokoʻa a me ka ʻike hou ʻana o kekahi koholā. No laila, ʻoi aku ke kilo ʻana no ka wā lōʻihi aʻe a me nā wahi ʻē.’

Acknowledgments:

This work would not have been possible without the support and collaboration of so many partners.

Fieldwork in Hawai‘i was supported by the University of Hawai‘i at Mānoa, the U.S. Department of Defense’s Defense University Research Instrumentation Program, the Office of Naval Research, ‘Our Oceans’ Netflix, Wildspace Productions & Freeborne Media, the Omidyar Ohana Foundation, PacWhale Eco-Adventures, and the generous members and donors of Pacific Whale Foundation. Research in Southeast Alaska was funded by the National Geographic Society (NGS), Lindblad Expeditions-National Geographic (LEX-NG) Fund, and the North Pacific Research Board.

We are especially grateful to our key collaborators at the Pacific Whale Foundation, Griffith University, and the Alaska Whale Foundation, whose partnership, knowledge, and dedication made this work possible.

A heartfelt thank-you goes to the many research assistants who braved long days in the field and countless hours of data processing, particularly Abigail Machernis, Grace Olson, Florence Sullivan, Elizabeth Beato, Jessie Hoffman, and Dana Bloch, for their tireless commitment.

We also thank PacWhale Eco-Adventures, Christina Lovitt & Emma Nelson (Maui Ocean Adventures), Lee James (Ultimate Whale Watch), and the Guth family for their invaluable logistical support on the water.

A special thanks to Cameron Nemeth for translating the abstract into Hawaiian and Ekolu Kelley for verification and editing, an important step in making our research accessible to the Native Hawaiian community.

All data were collected under NOAA/NMFS research permits 21476 and 20311 and with approval from the University of Hawai‘i Institutional Animal Care and Use Committee (18-2971). All drone operations were conducted by qualified FAA Part 107 pilots.