Science Under the Ice

For me, the past few dives have been dreams come true–when I first thought up building Icefin, what we are doing now is what I hoped to be able to do.  The first dreams of this for me were in 2010, when I first got involved in a proposal for underwater robotics program.  By 2012, I was leading that effort.  During that program, SIMPLE, I started thinking about how to do big science–with tens of instruments over kilometers of surveys–but with low logistics.  Big vehicles and big programs take big resources, so could we do that same thing with a vehicle built with that in mind?  Thankfully, this season has proved that we can, and that exciting things lay ahead.



Icefin viewed from the McMurdo Observation Tube (credit: D. Dichek)

Up until recently, we’ve been collecting science data during primarily vehicle check out dives.  This is a critical part of a new robotics project–you have to know the vehicle well before you can use it to it’s best ability.  You spend years planning and spec’ing parts, months putting it together and testing it out just for simple functions.  Can we drive forward? Do the thrusters work? Can we see the data streams?  Then you bring the vehicle and all of the sensors and people to a totally new environment.  Harsh temperatures, testing in environments where you can’t see how it’s floating or swimming, shaking and jolting the hardware in the pisten bully, not to mention a unique kind of isolation and stress that occurs for the team when you cannot simply jump on the phone to the manufacturers, or head to your favorite restaurant to come down from a hard day.

So when you’re a scientist leading an engineering-based program, the day that the project finally makes it into science-first mode evokes grand and sweeping emotions. We’ve made it!  Below, I’ve written a summary of the last dives around McMurdo for you to enjoy.  We’re sitting in Mactown (McMurdo) waiting on the weather to clear to fly our to join the Ross Ice Shelf Programme, and try to access the never before seen subsurface at the Hot Water Drill 2 (HWD2) camp!

Dive 5, November 23–Out Under the Sea Ice

The goal for Dive 5 was to survey the ice-ocean interface.  Here, the complex oceanographic interactions throughout the Ross Ice Shelf and the McMurdo Sound conspire to create not only the sea ice itself (annual to several-year ice that melts or breaks up every year) but also a super cooled upper ocean layer that causes the ocean to freeze out forming a maze of beautiful ice crystals called “platelets” everywhere under the ice.  This is a process unique to places in Antarctica that are next to deep ice shelves.  These interactions are part of what we’re studying in order to better understand how the ice and ocean are changing here on Earth as the climate warms, but also is a great example of how processes on Europa might also work.  Europa has slowly built its thick ice shell by freezing out its ocean in a manner possibly similar to what we’re observing here in McMurdo.


The underside of the Big John crack and multi-year to single year ice transition

On our dive, our goal was to map the ocean water and the ice, understanding whether there are any changes as we got closer and farther from the ice shelf, looking for differences under the first-year and multi-year ice, and of course noticing all we can about the life under the ice.  The platelets are amazing, glittering in the bit of light coming down through the thin (2-4 m) ice and bit of light from Icefin’s forward LED. They form dune-like structures influenced by brine drainage out of the ice and the blowing currents under the ice.  But some of my favorite moments are always seeing the silvery-white “borks” hiding in and around the ice, and the tiny sea creatures darting in and out of the ice.


Platelets under the sea ice

We drove the vehicle for over four hours and several hundred meters, successfully collecting data from our conductivity/temperature (CT), dissolved oxygen (DO), and dissolved organic matter (CDOM/FDOM) sensors, along with our cameras.

Dive 6, November 24–Down to the Seafloor!

Dive 6 for Icefin was our best yet: our long-awaited dive to the seafloor. The vehicle checkout was successful, with all sensors operating nominally. We had doubled our dive weight and removed the dyneema secondary line (the Icefin tether is strong enough to hold nearly twice the vehicle’s weight, so the dyneema isn’t necessary for sea ice operations through the fish hut. We had the vehicle still oriented in bottom-facing mode, and had integrated the DVL data into the navigation solutions. We also switched the position of the DO sensor and the forward light, which drastically improved our imaging capability.


A crinoid makes a rapid pass close to the down camera, with the laser scale on

We deployed the vehicle freely spooling to the seafloor, which took about 35 minutes once in the water. The depth there ranged between 530-540m. We took successful water column data and explored the seafloor community. We also completed ~100m long back and forth surveys at 20m and 50m altitude above the seafloor to tune our sidescan sonar acquisition.

Dive 06 3D plot.png

A map of our track at the sea floor–note that the scale bar and lat/lon are not corrected in this graph.

We observed multiple crinoids, sponges, pelagic squid, jellies, sea stars, fishes, and anemones. The crinoids were a particularly interesting part of the dive, they were very active and were attracted to the vehicle’s light and movement. We were encouraged by the good match between the CTD and DO sensor temperature values, and noted that the depth matched our measurements taken in 2015 near this site with our CTD. We left the dive feeling confident and excited for one more week of sea ice operations before packing to join the RIS programme out at the drill site.

Dive 7, November 29–Erebus Glacier Tongue, Part I

Erebus Glacier Tongue is an exciting place to go with an AUV/ROV.  Here, the glacial ice flowing off the volcano flows down into the ocean and goes afloat.  Because of its relative protection from currents by the Ross Island, the tongue survives several kilometers out into the sound despite its relatively small size.  For the first EGT Dive, our mission goal was to follow along the seafloor to search for the place where the ice contacts the seafloor, called the grounding line. It was also our first deployment of the vehicle using our field deployment “Launch and Recovery System” or LARS.  As it was the first time deploying the vehicle in this way this season, the deployment process took a while longer than usual, with the team trouble shooting how to lift and position the vehicle.


Brinicles and algae growing from the ice

The dive was successful overall. We slowly surveyed the ice through the hole, with the opportunity to view filamentous bacteria and algae attached to the underside of the ice, as well as fields of brinicles. The seafloor directly under our site, about 200m from the ice tongue, was 180 m deep. We then traversed along the seafloor, surveying the rich life there. Unfortunately, during one survey, when we went to lift the vehicle off the seafloor, rocky debris got pulled through the z-thrusters and damaged their small plastic impellors. After about 2 hours of diving, we were cut short since the vehicle could not be controlled in the z-direction. We left early, but left the LARS set up for a return to the site on Friday.  However, the data recovered by the vehicle left us excited for a return.


Sponges and brittle stars near the EGT

Dive 8, December 1–Erebus Glacier Tongue, Part II

The goal for Dive 8 was the same as Dive 7: to reach the grounding line of the glacier and survey the ice-ocean conditions and life in the environment.


Sea life under the EGT

Dive 8 was a complete success on all fronts. We drove the vehicle successfully for over five hours. We started our dive with the same goal as Dive 7, to follow the seafloor and hopefully intersect grounded ice. We dropped initially to the 180m deep seafloor below our site, and then drove straight toward the glacier tongue. We successfully gathered CT, DO, CDOM, sidescan, and forward and down imaging data. We drove the vehicle 600m from the dive hole without intersecting the ice, and so we decided to lift the vehicle to confirm that we were under EGT. Along the traverse out, the seafloor gradually dropped from 180m to 290m, likely carved out by the advance of the ice. We drove the vehicle upward to 190m, but were having some trouble from the tether (it turns out that this was likely wrapped under an iceberg coming off the south side of the tongue). Since we didn’t want to get the vehicle stuck, we returned to the seafloor and drove back towards the dive hole.


Boop! A seal checks out Icefin at 250m depth

By this time, we were on fairly low battery but wanted to achieve a survey of the tongue, so we pulled the vehicle up to about 100m from the hole and then drove the vehicle to the base of the sea ice to traverse back out. We drove a straight line along a crack in the sea ice to where it intersected an iceberg that was calved off the side of the tongue. The iceberg was incredible—irregular shaped below the surface, with cracks and prominences and both broad and sharp faces. The whole surface was covered by small cusps from ablation and melt. We brought the vehicle back to the hole with less than 20% on the batteries, our best use of the power yet.


Our first view of the EGT and the Iceberg

Dive 11, December 5–Long Run Under the McMurdo Ice Shelf

Dive 9 was a short fun dive where we filmed with a BBC crew. Dive 10 was our ballasting dive where we finalized the vehicle trim. That was to be our last dive in McMurdo, since our cargo was due to be delivered on Monday in order to prepare for our scheduled Wednesday flight. However on Monday morning we learned that our flight would be delayed until at least Dec 11, and so we had part of the week to operate. This delay was unfortunate for the deep field project, but from the perspective of vehicle operations and science in McMurdo, was fortuitous so that we could complete our remaining goals in the area.


Borks schooling under the shelf and hiding in the ice

Dive 11 was focused on making a long run under the ice shelf. In a previous season, ARTEMIS made an 800m traverse under the ice shelf. We made 1 kilometer our goal for Icefin. We rotated the DVL up to enhance navigation and get a good track under the ice. We drove about 600m without intersecting the ice shelf, which helped us diagnose an issue with the IMU rotation matrix by which we navigate. We pulled the vehicle back to the hole to re-zero the navigation and head out in the correct direction. This time, we crossed under the shelf at just under 300m from the dive hut, and drove out under the ice 1.3 km, our longest dive yet and one of the longest under ice shelf dives of any vehicle in Antarctica thus far (we believe that the only longer under shelf dive was by Autosub under Pine Island Glaceir).   On the dive, we saw schools of hundreds of “Bork” fishes, imaged small crustaceans in the ice, and characterized the beautiful and ubiquitous platelet ice.


An ice cave under the McMurdo Ice Shelf

We also were excited to find a giant unexpected ice mass protruding in a deep draft mound several meters down from the otherwise smooth and shallowly sloping shelf. Within it we observed an ice cave. This is surprising to find, since it’s unclear how such a mass would form in place. Two hypotheses are that we were under a nearby rift in the shelf, and that during the rifting, some of the shelf ice became rotated and trapped, or surface melt has escaped through the rift and causes more ice accretion there. We drove the vehicle in total over 2 kilometers and over five hours.

Dive 12, December 6–Exploring a New Seafloor

Dive 12 was our chance to make it to our Ice Shelf Front site. We chose this site because of the very different oceanographic conditions there relative to our base site.  At the Icefin base site, the current is relatively benign and runs parallel to the ice shelf and largely out from under the shelf, allowing for so much ice accretion. At the Shelf Front site, the current is fast and flows into the shelf, ablating the ice and preventing accumulation of much fine-grained material at the seafloor. It is also very deep–780m.


A trio of crinoids chase the robot

This dive was also a resounding success. We had planned to do a short dive to the seafloor with the vehicle taking an oceanographic cast, then drop the drop weight to survey around briefly, and then return to the hole before completing a shelf interface run. In effect however, we had very little time to dive that day, and the seafloor turned out to be so interesting that we ended up staying at the seafloor much longer than we expected.

The current was very strong, so we started our dive in what we refer to as “pendulum mode” where we use the vehicle thrusters to drift around and make surveys, but the vehicle stays controlled by the tether. We used this mode for about 30 minutes over about 150 meters to guarantee that we got some data while at the seafloor, and were very pleased. We then dropped the drop weight and went out under vehicle control successfully. While the conditions were challenging, the control was pretty effective and we were pleased with the vehicle performance. We went out over 600m horizontally on free spooling tether.


Sea fan and brittle stars among the benthic community

We decided to pull back up through the hole to make some adjustments before heading out under the ice shelf. However, when we returned to the hole it was clogged with platelet ice. This happens occasionally and isn’t a big deal, but it took a long while to extract the vehicle. Since it had grown late by this hour, we made the decision to end the dive there.

On the way back to the hole, we got additional sensor data and had a surprising mid-ocean encounter with an elusive Antarctic Toothfish. We got great images of this large rare and threatened fish mid-swim.  If you haven’t seen the movie The Last Ocean, I highly suggest you watch it.  The punchline is, don’t eat Chilean Sea Bass.  Chilean Sea Bass is just a fancy rebranding of the Antarctic and Patagonian Sea Bass.  They are being overfished with little regard for the length of time it takes these fish to mature–17 years.  It’s easy to decimate a population like that.

As the McMurdo season draws to a close and we wait for the next adventure, I’m more excited than I’ve ever been about the project.  Onward to the Ross Ice Shelf!


Icefin Team at EGT Site, with Icefin preparing to dive


4 responses to “Science Under the Ice

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s