NOAA Conducts First Comprehensive Aerial Survey of Ice Seals in the Arctic

During spring 2025, scientists conducted the most extensive aerial survey of ice-associated seals to date. The survey took place between April 4 and June 10, 2025, off the coasts of western and northern Alaska. The goal was to determine the abundance and regional distribution of four species in U.S. waters: bearded, ringed, spotted, and ribbon seals.

This survey used multispectral camera systems enabled with artificial intelligence (AI) to detect and document seals hauled out on the spring sea ice. We will use data from these sightings to estimate species abundance and distribution-critical information used to monitor and manage wildlife populations.

Spring is the Best Season for Surveys

Bearded, ringed, spotted, and ribbon seals are known collectively as ice seals because they use seasonal sea ice as a platform to rest and raise their young. They also haul out on the ice while they undergo an annual molt cycle. Molting is an energy-intensive process of shedding the top layer of skin to reveal a new fur coat. All ages of ice seals undergo this process each spring. This provides a short window when many seals are out of the water and available to be counted from the air.

The research is part of the Alaska Ice Seal Research Plan , which outlines key research priorities for these animals. The plan is developed each year in collaboration with the Alaska Native Ice Seal Committee's Co-management Working Group. Seals are vital resources for northern coastal Alaska Native communities and are key species in Arctic marine ecosystems. Abundance estimates and distribution maps are crucial for sound decision-making about:

• Co-management of subsistence use of ice seals
• Conservation
• Permitting of activities in the Arctic that could affect these species or their habitat

Researchers included scientists from the NOAA's Alaska Fisheries Science Center and the University of Washington's Cooperative Institute for Climate, Ocean, and Ecosystem Studies .

Mapping the Arctic Seas

The research team used two specialized NOAA aircraft for the survey-a Twin Otter and a King Air. They flew out of several Alaskan communities, including Nome, Bethel, Kotzebue, Utqiagvik, and Deadhorse. Nightly, the team planned the flight paths using up-to-date satellite imagery of sea ice to ensure comprehensive sampling of the seals' changing habitat. As weather allowed, the aircraft maintained a survey altitude between 1,000 and 1,200 feet while over sea ice.

Over the course of the 68-day survey, poor weather conditions prevented flying on many days. But overall, the team completed 58 flights, covering 39,663 kilometers. That's equivalent to 24,645 miles, which is just less than one time around the planet!

High-Tech Data Collection from Above

The survey aircraft collected thermal and high-resolution color images of the sea ice below. The imagery captured a swath between 400 and 500 meters (1300-1600 ft) wide directly beneath the aircraft. One of the planes also deployed experimental ultraviolet cameras that may be useful for detecting polar bears.

Over the course of the survey, the team collected more than 1.5 million sets of images totaling more than 26 terabytes of data. To efficiently process this vast amount of data, the team used an AI-machine learning algorithm in-flight to analyze thermal imagery in real time. This system helped scientists locate animals and, in some cases, identify which images to save for further analysis.

What the Cameras See

The combination of different types of cameras provided a detailed view of the seals on the ice. In the example below, thermal imagery clearly showed the heat signature of a recently born bearded seal pup and its mother, which helped researchers spot them against the cold ice. The AI-enabled seal detection model was trained to find these heat signatures to reduce the number of images biologists need to review. This saved a tremendous amount of time. Biologists used the corresponding high-resolution color photo to identify the species present. The camera systems detected additional animals throughout the survey, including polar bears, caribou, foxes, and even some birds.

What's Next?

With the fieldwork complete, the team is deep in the meticulous process of analyzing the terabytes of data collected. Experienced ice seal biologists have been reviewing and classifying thermal detections and identifying species in the color imagery.

The team also recently developed a new thermal detection model to analyze data from the newer thermal cameras used on the King Air aircraft. Preliminary testing shows that both models detect more than 90 percent of seals in the thermal imagery.

After a series of quality control checks, the team will assemble a count of the ice seals detected. They will use statistical models to estimate the total abundance and distribution for each of the four seal species. These data allow us to understand the current status of these populations and how abundance may change over time. They will also provide perspective on how these species may adjust to a changing environment. Ultimately, all data will be made available to the public and the project's progress and findings will be shared through community meetings and reports.