Monday :: Oct 29, 2007

southward ho

by Christina Hulbe

A couple of weeks from now, if everything goes according to plan, I will be living in a tent west Antarctica (simple map). I'll be back in late December. This is the second of two field seasons in a project looking for signatures of recent (the last few hundred years) change in a region of Antarctica that may be capable of rapid change. I thought before I head out of the hemisphere, I should write a bit about what we're doing and what the work is like.

One of the outstanding issues in Antarctic glaciology is understanding the likelihood of rapid change in the Ross Sea sector of the West Antarctic Ice Sheet (WAIS). The WAIS is a "marine" ice sheet, resting on a bed that is well below sea level (map) and discharging into large, embayed, floating ice shelves on the Ross and Weddell Seas. Between 5 and 6 meters of sea level equivalent is stored in the ice sheet and even modest changes in the behavior of the system can have meaningful effects on sea level.

The transition zone between ice grounded on bedrock and ice floating in the ocean is called the grounding line (or grounding zone). Changes in grounding line position are connected to changes in either mass budget or dynamics of the coupled ice sheet/ice shelf system. Because the grounding line is where ice flowing out of the ice sheet makes its contribution to sea level, it is important to understand how and why it changes over time.

ice flow
The WAIS has a distinctive surface shape, concave up rather than convex up, which tells us the rapid ice flow usually found near the margins of ice sheets (as in Greenland or East Antarctica) must extend well into the interior of the ice sheet, allowing it to efficiently move ice from the inland reservoir back out to the sea. Indeed this, is the case.

Before the ice sheet formed, West Antarctica was an archipelago. That history has an important affect on ice dynamics. In some parts of the Ross sector, there is a thin (a few meters) layer of marine sediment between the ice and rock below. That sediment, when water-saturated, allows the ice above it to flow very rapidly. The result is a relatively slow-flowing (10's of meters per year) ice sheet through which fast-flowing "ice streams" (100's of meters per year) course (pdf of a technical paper). The traces of the Ross ice streams can be seen from space because the shear margins that snake along the edges of the streams, marking the boundary between slow and rapid ice flow, are heavily crevassed.

Glacier ice behaves as a viscous fluid, compelled to flow by gravity. Its deformation (flow) is a reflection of the balance of forces acting on the material. The gravitational force is the product of mass and the acceleration due to gravity. The magnitude of the resulting gravitational driving stress within the ice depends on the surface slope and thickness of the ice. The driving stress must be balanced by other stresses (stress is the product of a force and the area over which it is applied). Resistive stresses act at the base of the ice, along the the sides of outlet glaciers and ice shelves, and at the seaward fronts of tidewater glaciers and ice shelves. The basal resistance is relatively large where the ice flows over a rigid substrate and relatively small where the ice flows over a soft substrate (water-saturated sediment beneath ice streams; water beneath ice shelves). The Ross ice streams flow fast despite small driving stresses (small surface slopes) because the water saturated sediments beneath them offer little resistance to the driving stress.

The concern now is that changes in the extent of the Ross Ice Shelf could result in a speed up of the ice streams that flow into it. This idea was first put forward in the late 1970's, by John Mercer (abstract) and by Bob Thomas (abstract), before much was known about ice streams, and hinges on how changes in ice shelf geometry affect stress balance at the grounding line, and thus grounding line position. This is a topic in which I have a keen interest, along with many others, for a variety of reasons.

recent history of flow variation
Back in the year 2000, some friends of mine published an interesting paper, documenting flow features in the Ross Ice Shelf that could be used to infer past ice flow in the region (popular article). The observations were made from space, using weather satellite (AVHRR) data and a novel data processing technique that "stacked" together multiple images of the same area so that subtle variations in surface slope were illuminated by multiple sun angels and in that way amplified. Now we use higher-resolution imagery (the MODIS sensor flying on the Terra and Aqua satellites) and see even more. You can get a feel for this yourself at the National Snow and Ice Data Center MODIS Mosaic of Antarctica website.

What my friends noticed was that flow traces in part of the floating ice shelf (termed streaklines in fluid dynamics) were distorted relative to the present-day flow field. If you've ever seen photographs or video of smoke in a wind tunnel (or dye in the water-filled equivalent), you've seen streaklines in a fluid flow. They are the connected locations of a set of particles released into a fluid flow at discrete intervals from a fixed location. When a particle is released into a fluid flow, the direction it travels follows the azimuth of the velocity of the surrounding fluid. If the flow field is steady, you could use a map of flow directions throughout the fluid to draw a trajectory for the particle. Should the flow be disturbed for some reason, those directions would change and the path of the particle would change accordingly. Where the flow field is steady, streaklines and flow direction fields match. Where the flow field has varied over time, streaklines appear distorted relative the flow field.

So my friends wrote a paper, using mostly geometric arguments to deduce past events from the distortions they saw in the Ross Ice Shelf streaklines. At the simplest level, they could conclude that ice flow from the East Antarctic Ice Sheet (through the Transantarctic Mountains) had been constant over the period of time captured by the ice shelf record (about the last 1000 years) but flow from the Ross ice streams had not been constant. When I saw the work, I recognized that I could improve and expand on their interpretation by using numerical models of ice flow (I always think things like that) and for the next few years I set up a sort of cottage industry, running dozens of simulations trying out various flow histories and comparing model results with the observed ice shelf features.

Six years and a couple of papers later, what we have concluded thus far is that the Ross ice streams stagnate and reactivate on century time scales. These are short time scales for a large ice sheet and thus a somewhat unexpected result. The cause for such variability is not climate, but must instead be related to the thermodynamics of the ice streams. Internal variability is one way to state it. There are a few important follow-on thoughts from this, icnluding: none of the models used to project future ice sheet behavior account for internal variability of this sort (recall my post about Greenland); and the Mercer and Thomas scenarios are too simple for this region.

other views
Colleagues up at the University of Washington and at St. Olaf College in Minnesota have been working on the topic of changes in the flow of the Ross ice streams using very different data: radar imaging of internal layers (and buried crevasses that mark relict shear margins) within the ice (project website). Distortions in the layers record details of past ice flow. The trick of course, is figuring out how to read the record.

When I compare notes with my colleagues on the timing of events near the modern Ross grounding line, sometimes the results of our different approaches agree and sometimes they don't. We're all observing the same system though, so the discrepancies must arise from our imperfect knowledge. The desire to reconcile these discrepancies, and thus improve our understanding of century-scale variability in ice discharge from the Ross Sea sector of the WAIS, led to the project that takes me to West Antarctica next week.

the work at hand
This current project grew from conversations between myself and a doctoral student up at UW who was collecting and working with the radar data. She has her PhD now and has moved on to a research faculty position at the University of Texas. Our big picture ideas match for the most part but some details differ and the timing we identify for some key events doesn't match up. So we decided to write a proposal to the National Science Foundation to work on the problem together, to try to sort it all out.
We're making measurements at several modern grounding line locations in order to use what we observe in the present system as a guide to reading the record of the past. Classic geology, that.

The work includes more radar imaging of internal layers and GPS surveying used to measure spatial patterns in ice flow at key locations. We're way out there, 100's of kilometers from the US base, McMurdo, living in tents (five of us this year, four last year) and fending for ourselves. It's a good life, really. I'll try to post some photographs later this week. I think the project will yield useful results and I am enjoying the opportunity to work with a talented (both with the science and the field logistics) young scientist.

Christina Hulbe :: 12:30 PM :: Comments (15) :: Digg It!