Big Fish, Big Data

In the mid-1980s, the technology for electronically tracking ocean wildlife was just being developed. Early electronic tags relied on acoustic pings to communicate location and depth, and required captains and crews using directional hydrophones to actively follow marine animals through the ocean – for hours or days at a time.

The first challenge in using these technologies, of course was getting close enough to the animals to attach the tags to them. Fortunately, tournament gamefishers had been successfully catching and releasing billfish for decades. Especially off Hawaii’s Kona Coast, where the Hawaiian International Billfish Tournament (HIBT) had been run annually since 1959.

Since the 1980s, Stanford University’s Dr. Barbara Block and her lab have been working with the organizers, anglers, captains and crews of the HIBT. Initially, they deployed acoustic tags, which provided information about the depths and speeds at which the incredible Pacific blue marlin traveled. Eventually they started tagging them with the first prototype pop-up satellite tags, which recorded information about depth, temperature and light over weeks or months – revealing for the first time the epic migrations made by these magnificent predators once they left the waters of Hawaii.

In 2009, Peter Fithian, founder and director of the HIBT, wanted to create a special tagging event for the HIBT’s 50th Anniversary, to commemorate the HIBT’s special role in developing animal tagging technology. A group of us, including (at the time) graduate student George Shillinger, tournament coordinator Bob Kurz, and me, came up with the idea of having tournament anglers sponsor tags as part of a competition-within-a-competition, where the tag that went the furthest from where it was deployed to where it released itself from the marlin and “popped up,” would be declared the winner. Thus the Great Marlin Race was born. The Great Marlin Race concept proved popular with the anglers, and has returned to the HIBT every year since.

In 2011, the International Game Fish Association (IGFA) officially partnered with the Great Marlin Race, with a goal of running similar events at billfish tournaments around the world at locations including Australia, Puerto Rico, Kenya, the Galapagos, the Bahamas, the Seychelles, Papua New Guinea, Mexico, Japan, Costa Rica and Nicaragua. A total of 314 billfish have been tagged in 54 different Great Marlin Race events, generating more than 165,000 nautical miles of tracks.

So what does that have to do with big data? Each one of these electronic tags consists of sensors which measure light, pressure, and temperature, and record these data every 60 seconds. After a pre-set period of time (currently 240 days for IGFA Great Marline Race events), the tag releases itself from the fish and floats to the surface. Using the ARGOS satellite system, the tag transmits a summary of its data back to the laboratory over the course of about 10 days before its battery fails. These data summaries contain a staggering amount of data – and if the tags can be physically recovered (which happens occasionally), a complete, minute-by-minute record of the fish’s behavior can be downloaded.

The data from a single tag can provide unprecedented insights into animal behavior. For instance, a blue marlin tagged off Puerto Rico popped up off the southern tip of Africa, documenting connectivity between these remote locations. Black marlin have made their way from Lizard Island, off Australia’s east coast, almost all the way across the South Pacific towards South America, and a blue marlin tagged in Richard’s Bay, South Africa (in the Indian Ocean) made it past Point Agulhas into the Atlantic, documenting movement from one ocean basin to another.

Although these individual findings are fascinating, deeper insights into the “why” questions – e.g., Why do these animals go where they go? What factors control their migrations? – can be answered only after large numbers of tracks have been recorded and the data combined, allowing researchers to look for patterns of behavior that occur across an entire sample population.

Using the multi-year dataset in the Pacific, most of which was collected on blue marlin tagged at the HIBT, researchers from the Block Lab, led by Dr. Aaron Carlisle, were able to carry out the first such meta-analysis of Pacific blue marlin behavior. This study showed that both oxygen and temperature affect how deep marlin dive and where they travel; and in places where both oxygen and temperature are low – as is the case in a large portion of the Pacific Ocean between Hawaii and North America – their available habitat becomes vertically compressed, confining them to the uppermost part of the water column. It also demonstrated that during El Nino conditions, the tongue of cold surface water which extends westward from South America along the Equator creates a barrier to the migration of blue marlin, effectively confining them to the northern hemisphere. Although this is probably inconsequential now, with El Nino conditions occurring only every few years; some climate scientists have suggested that as the ocean warms due to climate change, El Nino conditions could become “the norm” in the Pacific – raising the possibility that Pacific blue marlin populations in the northern and southern hemispheres could become isolated from one another.

These insights would not have been possible without the very large dataset gathered through years of electronic tagging, and the sophisticated computing tools required to carry out the analyses used in this study. It is also worth pointing out that the collection of this “big data” set would never have been possible without the incredible collaboration among biologists and computer scientists, billfish tournament organizers and anglers, captains and crews, generous sponsors, and the members and staff of the International Game Fish Association. The diversity of expertise, and the dedication to working together to answer big questions, is not unique to the IGFA Great Marlin Race, and may be an important, if often overlooked, aspect of the big data world.

At the Oceans of Data Institute, we are developing ways to share data – including tracking data from tagged marine animals, in our Ocean Tracks program – for students to explore. Working with the same datasets scientists are using, students using the Ocean Tracks interface can study the ways that marine animals migrate across the Pacific Ocean, and how those movements are shaped by the environment around them. We welcome you to join us in this journey of discovery at http://oceantracks.org.