Dr. Greg Skomal is senior fisheries expert with Massachusetts Marine Fisheries, and the State’s leading shark expert. He spoke with us last year about his basking shark research. He has recently been part of a research team that is investigating the use of robot shark tracking technology, as featured on Discovery Channel’s Shark Week, which he talks about below.
Keren: Thanks for taking the time to talk to us about the exciting new shark tracker technology where robots (autonomous underwater vehicles- AUVs) follow sharks in our waters! Can you tell me how this project began?
Greg: Conceptually this goes back many years – to around 2008. We were starting to track basking sharks with various kinds of technology and learned that these sharks spend quite a bit of time diving. We were curious about what they do when they dive and thought it would be interesting to have a robot follow them and sample the water column to see what they’re eating. We worked with a plankton biologist at Woods Hole Oceanographic Institute (WHOI), and WHOI gave us the initial funding in 2008. That was the first time we were able to develop an AUV that could track marine animal.
Then, we were making a show with Discovery on white sharks in 2010 and one of the producers of the show found out what we were doing with AUVs at WHOI and he suggested that Discovery might fund that research! So in 2011 and 2012 we began fieldwork with the WHOI oceanographic systems engineering department developing and implementing the AUVs.
What’s happening currently is an engineering feat. I’m a marine biologist so I can provide a sense of how a shark might behave, and based on that engineers are able to program the software and the machinery.
Keren: Can you explain more about what the AUVs actually do?
Greg: It’s essentially an unmanned mission under water; you can program the AUV to do many different things, like map the ocean floor or do water sampling. The robots can do something as simple as ‘mow the lawn’ (go back and forth on the ocean floor taking data) or something as complex as tracking a shark.
In November 2011 we began field trials, simulating tracking something with an AUV. The first step was to have the robot try to follow boat, then a person. I dove in the water and swam around, and the AUV followed me. In 2012 the AUV did its first marine mammal track, following a basking shark.
Keren: What specific kind of information does the AUV collect?
Greg: With these kinds of AUVS you can measure pretty much anything- depth, current profiling, basic oceanographic data like temperature, dissolved oxygen, salinity. There are various modules you can clip onto the AUV, so if you want additional measurements, like amount of chlorophyll and plankton abundance, you can get that information too.
Keren: Can you explain more about how the technology works?
Greg: Sure – you can’t train a robot to recognize a fish, instead, you put a transponder on the fish. Generally, there is a communication system, which transmits a high frequency ping from the fish to the robot. In this new technology we’re using with the sharks, it’s a bit more complex- there’s actually two-way communication where the tracker asks questions that can be answered. The tracker can learn information about depth, range, and bearing from the transponder. The tracker can then navigate relative to where animal is, and follow the shark in a more precise manner. We have to catch the shark to put the transponder on it and the technology has a release mechanism so it will detach after a certain amount of time.
Keren: Similar work has been going on the West Coast, what are the differences?
Greg: The method by which the animal is being tracked is different. On the West Coast researchers attach a high frequency tag to the shark and the AUV detects the sound and positions itself based on directionality and intensity of sound source. The AUV can track for several hours doing that. This technology involves a lot of searching and listening- the communication is more passive.
The technology we have developed involves a more robust navigational system where the two devices can interact, allowing more precise tracking and positioning. The transponder can send information through underwater modem technology to the AUV, and the AUV can ‘ask questions’ back to the transponder. For instance, our transponder can detect depth and send this information back to the robot. Our system also has more video capability. However, it’s a bigger transponder. The technology on the West Coast is smaller, and therefore can attach to smaller sharks.
I fully support the development of multiple technologies because I think what will eventually emerge is the best of both worlds.
Keren: What are the main difficulties using AUVs?
Greg: You have to tag the shark, and that’s always a challenge- we’ve had great success in doing it but it can be hard. Any time you put electronics in the ocean there are many factors to consider. The engineers at WHOI do a really great job with ocean technologies.
Also, robots don’t always know what sharks are going to do next; we are always trying to figure out ways to predict the behavior of the shark. If we can predict what the shark will do we can arm the AUV with the best information for making decisions.
Keren: What are you most interested to learn?
I’m interested in understanding the fine scale behavioral data we’re getting. We’re learning about the movement of the sharks, things we had no idea about before. For instance, we’ve learned they have no problem swimming against the current at sustainable speeds of 5 or 6 knots.
We can also learn about their movements as they relate to the environment in which they’re swimming, like behavioral changes with regards to tide and time of day, do they move off shore, and if so when, in general, what are their daily behavioral patterns?
It’s a lot of cool novel data we’re gaining from these AUVS, almost everything we’re learning is new- it’s really exciting!