Ice Hacker

Researcher works to improve models of glacier movement

Though outwardly fit, UM ice-sheet modeler Jesse Johnson describes himself as “the fat slob of glaciology.”

That’s because the associate professor of computer science contends glaciologists are generally superior athletes. A Dutch scientist he collaborates with, for example, was the first person to ski across Greenland and later skied to Antarctica’s South Pole. Another partner from Alaska recently received the Piolet d’Or (Golden Ice Axe), a coveted French mountaineering award. Even Johnson’s closest research partner at UM, Joel Harper, lives for weeks at a time on glaciers and has kite skied on Greenland’s ice cap.

“These are hyper-fit people,” Johnson says. “I think a lot of glaciologists get into the field so they can do mountaineering all the time. That’s not me. I’m more of a computer jockey. I interpret the data and try to improve our models of ice movement.”

Such improved models are key to helping humanity understand the potential perils presented by climate change, melting ice and sea-level rise. But as a guy who makes models, Johnson believes the public should be more skeptical about their power.

“Models help us understand how processes work, and that’s a good thing,” he says. “Unfortunately, policymakers and social pressures push us into situations where we try using our models to predict the future. To be honest with you, if I could predict the future, I wouldn’t be modeling ice sheets – I would be on Wall Street getting rich.”

Johnson came to UM in 2002 after a Peace Corps stint in Africa and then earning a computer science master’s degree and physics doctorate at the University of Maine. In Maine he worked at the university’s Climate Change Institute and was mentored by prominent ice-sheet researchers James Fastook and Terrence Hughes, who created large-scale models applicable to Greenland, Antarctica and past ice ages.

At UM, Johnson launched his own research program investigating ice movement. He says ice behaves much like a fluid during long timescales – not water, but liquids such as ketchup, paint or blood.

“These nonlinear fluids are weird,” he says. “Take ketchup for example. It forms a little dome on your plate that doesn’t really move around. But to get it on your plate, you had to put enough stress on it – by squeezing the bottle – to make it gush out. And that’s similar to how ice behaves – fluid-like under larger shear stress; solid-like under lower stresses. It’s called shear thinning.”

Last year Johnson worked at a South African mathematical institute studying the attributes of blood. “Ice is more like blood than water, so I was able to learn a lot of new tricks about how to deal with the nonlinearities in the viscosity of ice,” he says. “South Africans don’t know much about ice, but they know a lot about blood. That’s where the first artificial heart was installed.”

Johnson says he actually peer-reviewed one of Harper’s papers before they knew they were on the same campus. The two realized they have similar interests and have been close collaborators since 2004, garnering millions in grants from the National Science Foundation, NASA and even foreign countries.

“I think our work complements one another well,” Johnson says. “I’m strong on the computational side, and he’s very strong at thinking in terms of processes and going out into the field and getting the real-world measurements. I will formulate a model, which I’m quite confident is bad in many ways, and when the model fails to explain what Joel measures, we will change the model so that his measurements are reproduced. That’s how we improve the model.”

Harper says Johnson’s abilities as a modeler have been invaluable.

“I am quite fortunate to have a collaborator who can always bring to the table the most cutting-edge numerical tools and techniques – along with good creative thinking,” Harper says. “Jesse is a major player in the international ice-sheet modeling community, and many researchers seek to get him involved in their work. I’m lucky to have the advantage of being on the same campus.”

Johnson spends much of his time writing programs that access computer libraries developed by national laboratories to compute the state of ice sheets. He uses these libraries, which basically do math, to do the computations he needs for comparing Harper’s data to model output.

“With my background I have an easier time on the programming side, but then there is this deep mathematical side I struggle to understand,” Johnson says. “I think many good scientists would tell you the same thing: You can never know enough math.”

Johnson says scientists make extremely accurate measurements of ice velocity on the surface, but they know little of what happens deep within ice sheets – especially when they are kilometers thick. Harper has drilled into the ice to place sensors deep inside, and this has led to some interesting revelations.

“Conventional wisdom is that the fast motion you witness on top is due to sliding on the bottom, but our newer datasets suggest there is more deformation in the ice than we thought,” Johnson says. “The top ice moves faster, so it’s similar to a deck of cards being spread out by a dealer. Our models will need to be changed in fairly significant ways to accommodate the amount of deformation that we are seeing.”

He says the Harper data also reveal more water coming and going beneath the ice than suspected. The data reveal fluctuations that often change on a daily basis, and these changes occur on smaller scales and in shorter time frames than expected.

“In some ways the new data just presents more problems for our models, but I think these are the types of problems that make life fun,” Johnson says. “What would we do if we had everything figured out?”

For one of their more interesting projects, Johnson and Harper received a $1.08 million NASA grant to support Operation IceBridge, a space agency effort to use lasers aboard airplanes to map the density of the Greenland ice cap. The project will continue until a next-generation mapping satellite, ICESat-2, launches in a few years.

Johnson says they are part of a team that helps determine where the DC8 flights for IceBridge are flown.

“You can’t fly the planes everywhere, but how big of an error do you introduce by not flying everywhere?” Johnson asks. “So we take the surface and bed data that is gathered, and we put it into our models. If the models start failing – there is somehow more or less ice than expected – then we infer that the flights aren’t being done correctly and we advise flying more lines in certain areas.”

Though Johnson generally prefers the office for his glaciology work, his research has whisked him around the globe to interact with other modeling experts. He has fond memories of exploring Storglaciären, Sweden’s largest glacier, and of observing penguins in Antarctic cold from the deck of a ship – clad only in shorts after an extended sauna.

“This work has given me a front-row seat to some really fascinating things that are going on,” he says. “I think that alone makes it worthwhile, and if I can contribute in some modest ways with my science, so much the better.”

— By Cary Shimek

Next >