|Bench testing our communications system yesterday for our field experiment that starts this Saturday. Laura Palamara makes a butterfish habitat prediction as Josh Kohut looks on wondering where in the ocean all the butterfish will be. Laura's prediction map is sent via satellite phone to the glider David Aragon is wheeling out the door and onto the lawn at Rutgers IMCS. The data is then sent from the gliders satellite phone to the blue computer on the right which is partially out of view. This computer can be accessed by the other computer on the bottom right. The glider, the two computers on the right and I are going on the squid boat, the "Karen Elizabeth". Josh, Laura and David will run the operation back at the lab along with Matt Oliver from University of Delaware and Steven Gray at University of Hawaii.|
|The Slocum underwater robot glider whose communications system we plan to use to transfer data from ship to shore and back again in our field evaluation of the butterfish habitat model we made with the fisherman.|
There are a dozen reasons not to do this experiment. Actually we have already thought of a dozen that might cause us to fail and we haven’t even left the dock. But we just keep moving forward because at this point failure is not an option. Putting a regional scale habitat model into an operational nowcast mode is crazy in the first place, even if it is just experimental. But this adaptive test of our statistical hypothesis, which is what our model really is, presents a formidable communications challenge. It requires sending the model predictions Laura will compute onshore to the squid boat offshore near the edge of the continental shelf. We need to navigate the squid boat on top of these model predictions so we can accurately position our trawl tows in regions predicted to be “good” and “not so good” butterfish habitat. These predictions will change daily just like they do in the "hindcast" movie in the last post. Then, after each tow of the trawl net, we want to send the results back to shore in “real time”. Like everybody else, we take staying “connected” with cellphones and the internet for granted. So the technical challenge of getting our data 100 nautical miles offshore and back hadn’t really dawned on us until just 10 days ago. But “necessity is the mother of invention” and we have come up with a characteristically unconventional solution.
We are going use a MARACOOS underwater robot for tracking the vessel and as a satellite data transmitter and receiver during our experiment . Yes that’s right. We are going to strap a robot glider to the bridge of the squid boat and pass the data back and forth over the iridium satellite telephone nestled in its robotic tail. The glider is programed to send its position back to shore every hour so the lab can use this feature to track our squid boat in real time. MARACOOS already routinely sends commands and data back and forth between the laboratory and robot gliders at sea. So if we can set up that same kind of communication between the glider and a computer on the fishing boat our problem is solved—the glider becomes our R2D2 “operator” for our satellite telephone system. Thanks to David Aragon and Chip Haldeman of RUCOOL we now have a laptop computer with a freewave wireless antennae and communications software that can “talk” to the glider when lashed to the boat. I have another laptop wirelessly networked with this one that we will also bring with us to fish offshore. We will use this machine to write and analyze files including the model predictions Laura sends from shore through the glider. Chris Roebuck, the captain of the boat, and I can also use this shipboard wireless network to send our catch data back to shore. Based on the suggestions of Hamish Bowman, the marine GIS guru who lives in New Zealand, and who we met at last years MARACOOS annual meeting in Hoboken, New Jersey of all places (now this truly is a fortuitous and loose network!), Chris Roebuck and I will be use the Open Source Quantum GIS software to navigate the "Karen Elizibeth" on the habitat model predictions and choose locations to fish. Like every other part of this project the solution to our data transfer and visualization of model predictions depended on the generosity and advice of many other people. This is truly a collaborative, open source, research project.