[Updated 7PM] — A “Deep-SCINI” submersible with Linux-based Elphel cameras discovered surprisingly diverse life under the Antarctic ice shelf — and rapidly melting ice.
Linux has once again been spotted in the frigid seas around Antarctica, which is not so surprising considering that Tux, the penguin, enjoys a nice cold bath. In November, we learned about Woods Hole Oceanographic Institute’s expedition to measure the thickness of Antarctic sea ice using a Linux-based SeaBED AUV (autonomous underwater vehicle) from Seabed Technologies.
Deep-SCINI at the Ross Ice Shelf drillsite
(click image to enlarge)
Now, a National Science Foundation funded research team has made some remarkable discoveries under the Antarctic ice using a submersible called Deep-SCINI (Submersible Capable of under Ice Navigation and Imaging). Developed at the University of Nebraska-Lincoln, Deep-SCINI features an imaging system with three high-resolution, Elphel cameras featuring open source Linux computers (see farther below).
Deep-SCINI at the drillsite
(click images to enlarge)
Unlike the SeaBED, the tubular, 2-meter long Deep-SCINI is not an AUV, but rather a tethered, remotely operated vehicle (ROV). It took 45 minutes for the craft to cruise down through a 740-meter borehole through Antarctica’s Ross Ice Shelf, which was drilled with a jet of hot water. Deep-SCINI traveled deeper under the Antarctic ice than any previous ROV expedition.
The ROV was then set loose in a 10-meter deep layer of freezing water below the ice and above the sea floor known as the “grounding zone.” Here, the researchers were “stunned” to discover Elphel-captured images of fish and other aquatic animals living in the hostile, pitch black environment, according to a Jan. 21 story in Scientific American.
Lead driller Dennis Duling at the Ross Ice Shelf drillsite melt tank
(click image to enlarge)
A NASA ASTEP (Astrobiology Science & Technology for Exploring Planets) grant helped fund Deep-SCINI through UNL’s ANDRILL (ANtarctic geologic DRILLing) Science Management Office (SMO) under the leadership of Executive Director Frank Rack. The research was organized under the Whillans Ice Stream Subglacial Access Research Drilling (WISSARD) project, and included Northern Illinois University, Montana State University, and the University of California at Santa Cruz. UNL’s SMO was subcontracted to design and build the hot water drill system and provide Deep-SCINI for sub-ice deployments.
The drilling location is near the inward shore side of Antarctica’s France-sized Ross Ice Shelf, some 850 kilometers from the “the nearest place where the ocean is in contact with sunlight that allows tiny plankton to grow and sustain a food chain,” says Scientific American.
Ross Ice Shelf drillsite aerial view
(click image to enlarge)
Researchers had expected to find only a few microbes with sluggish metabolic rates. Yet, despite the fact that the freezing waters held little in the way of microbial food, Deep-SCINI discovered 20 to 30 species of fish and dozens of red, shrimp-like crustaceans and other unidentified marine invertebrates.
Several species have yet to be identified, and might be new discoveries. The ROV scooped up samples for further study.
It’s unclear what the fish and other creatures are eating. Possibilities include periodic waves of plankton, which were not in evidence during the dive, or microbes eating mineral grains or feeding on ammonium or methane seeps. Unlike the deep-sea ocean floor in more temperate regions, there was no evidence of mud-dwelling, epi-benthic life attached to the seafloor.
The search for life under the ice was ancillary to the main focus of the NSF-funded expedition. The chief goal was to determine how quickly the Ross Ice Shelf is melting and sliding into the ocean due to global warming. The drilling location was chosen due to its proximity to the Whillans Ice Stream, which feeds the ice shelf with new ice.
Deep-SCNI discovered a large number of pebbles on the seafloor, which were presumed to have fallen from the underside of the ice sheet as it melted. Because older, deeper sediments were found to be free of pebbles, it suggested a recent acceleration of melting due to global warming.
UNL announced the Deep-SCINI last July, and on Jan. 21, reported on its involvement via ANDRILL in the WISSARD expedition. The drilling expedition had been delayed a year, due to the U.S. government shutdown in Oct. 2013.
Bob Zook (left) and Justin Burnett working on Deep-SCINI components
(click image to enlarge)
The Deep-SCINI ROV was invented by Bob Zook, an engineer recruited by UNL’s ANDRILL office, as well as Justin Burnett, a UNL graduate student in mechanical engineering. Zook, who confirmed the presence of Linux in his craft’s cameras to LinuxGizmos in an email, told Scientific American the ROV’s nearly flawless maiden “flight” was “a minor miracle.” He added: “The rule of thumb down here is that any new technological thing does not work for the first deployment.”
The ROV has yet to be fitted with a navigation system, so it was maneuvered using “tricks,” says Scientific American. For example, the above-ice operators would “fly” the craft from one large rock to another, or reel in the tether a couple meters to tug the ROV from behind and point it away from the hole. The tether incorporated ENOP (Ethernet over Power) for communications.
The Deep-SCINI design follows an earlier Zook-designed SCINI model that had been tested in Antarctic waters starting in 2007. The original, 1.5-meter craft, which was limited to depths of 300 meters, was responsible for discovering a new species of sea anemone that lives in burrows in the underside of the Ross Ice Shelf.
The Deep-SCINI was designed to descend to 6,500 feet, or about two kilometers. The nearly two-meter long ROV has a diameter of 23 centimeters and weighs 60 to 80 pounds.
Deep-SCINI’s open source Elphel cams
The Deep-SCNI uses open source Linux-based cameras from Elphel, which are popular among academic researchers. Whereas the original SCINI craft had two older Elphel 353 cameras, the Deep-SCINI has three Elphel NC353L-369 cameras for upward, downward, and forward-looking views.
According to UNL’s Justin Burnett, each camera is fitted with 5-megapixel CMOS sensors inside a custom 3,000-psi tested pressure housing measuring 7.0 x 2.5 inches. The cameras were chosen for their openness, flexibility, large feature set, and support team. “This is a very flexible open source hardware and software IP camera,” he added.
The cameras are incorporated in Elphel’s Eyesis4Pi panoramic and stereophotogrammetric camera. A similar Elphel panoramic was originally mounted on Google Streetview cars before being replaced with an in-house design.
The Elphel NC353L-369 features an accessory board that integrates CompactFlash storage and configurable USB ports, one of which was used by ANDRILL to control a focus motor. It also provides pulse synchronization control for capturing multiple images simultaneously across a network. The NC353L-369 can record video or images to the CF card or an external SATA drive. A serial port provides access to the root console, which Rack said simplifies firmware development.
While, the Elphel camera is primarily used for video, the Deep-SCINI implementation focused on still images. “We actually stream images rather than video at variable frame rates from the device, depending on exposure and lighting conditions,” said Burnett. “The camera performs on board JPEG compression fast enough so that we are able to fly the vehicle smoothly, We are able to keep total delay under 200ms while still keeping bandwidth usage low.”
Custom viewing software lets the researchers adjust the amount of exposure, color compression, and binning to increase framerate. “It also allows us to take instantaneous full resolution images with a keystroke,” he said. “We have custom firmware written to interface with our vehicle piloting software as well, such that vehicle telemetry data is actually stored in the image exif data, greatly streamlining post processing.”
According to Scientific American, the camera lenses were reinforced with pressure-resistant sapphire crystal, as well as a streamlined body of aluminum rods and “syntactic foam comprising millions of tiny, hollow glass beads.”
In addition to the cameras, Deep-SCINI integrates five maneuvering thrusters, lights, a CT (conductivity-temperature) sensor, a syringe sampler for collecting water samples, and a gripper for picking up objects. A drop weight held in the gripper kept the vehicle oriented vertically while traveling through the borehole. An instrumented “clump weight” isolated the deployment winch and fiber optic junction box from the ROV’s tether.