March 23, 2017
The images you’re viewing on Steller Watch are from the remote cameras we placed at the westernmost reaches of the U.S., where Steller sea lions have declined dramatically. The remote cameras along with marking the animals are two research projects that are part of our mission to pinpoint why Steller sea lions in the western half of the Aleutian Islands are not rebounding like they are to the east.
In addition to Steller Watch, we just launched an effort to develop a program to automate the yearly Steller sea lion population count, our biggest and perhaps most important research project of the year. We partnered with Kaggle to accomplish this. Kaggle connects people to coders who compete to find a solution to big data challenges, saving organizations time and money. The project called NOAA Fisheries Steller Sea Lion Population Count is now live and runs through the end of June. This story was shared by the Alaska Fisheries Science Center if you’d like to read more.
Counting the Steller sea lion population is vital to monitoring the species and staying on top of abundance trends. Each summer, we head to Alaska and gear up for weeks at sea on a research vessel. Summer is best because it’s when Steller sea lions go on land, which biologists call hauling out. During the rest of the year, sea lions spend a lot of time in the open ocean.
So, how do we count? We use a handful of methods depending on the situation.
The summer months are a busy time for Steller sea lions – females give birth, begin rearing their pups and adults breed. Aggressive males stay on land and fast for months to defend their breeding territories. Even non-breeding juveniles and sub-adult males (males that are not yet sexually or socially mature) haul-out with greater consistency during the breeding season. There’s a flurry of activity at sea lion sites and scientists from the Alaska Fisheries Science Center are closely monitoring as much as possible while we count the population.
So, how do we count? We use a handful of methods depending on the situation. At smaller sites with fewer sea lions, we can count on land or by boat, either on the research vessel itself or by launching small skiffs to get a closer look. At large sites, we take aerial images from a plane or an unoccupied aircraft system, or hexacopter drone. During our 2016 survey, we collected over 20,000 aerial images that we counted 29,427 sea lions from, back at the office.
It takes two biologists up to four months to look through those aerial photos. And in those pictures, there are tens of thousands of individual sea lions, and we count and identify the age and sex of each one we count. We hope computer scientists will connect with us on Kaggle to create an innovative algorithm that can recognize sea lions in aerial photos. Automation, or partial automation, would make counting more efficient and give us biologists more time to focus on other important Steller sea lion research studies.
In our next blog, one of our biologists will share what it’s like to conduct aerial surveys in the traditional way, using a small plane.
I have been a biologist in NOAA Fisheries Alaska Fisheries Science Center studying Steller sea lion population abundance and life history for over 10 years. I am an FAA certified remote pilot and have been flying marine mammal surveys with our hexacopter since 2014.
We’d also like to know what you’d like to hear about. Head to our Talk forum to post a Question for Biologists and your question could be selected for a future blog post!