GeoLog

sea level

GeoTalk: To understand how ice sheets flow, look at the bedrock below

GeoTalk: To understand how ice sheets flow, look at the bedrock below

Geotalk is a regular feature highlighting early career researchers and their work. In this interview we speak to Mathieu Morlighem, an associate professor of Earth System Science at the University of California, Irvine who uses models to better understand ongoing changes in the Cryosphere. At the General Assembly he was the recipient of a 2018 Arne Richter Award for Outstanding Early Career Scientists.  

Could you start by introducing yourself and telling us a little more about your career path so far?

Mathieu Morlighem (Credit: Mathieu Morlighem)

I am an associate professor at the University of California Irvine (UCI), in the department of Earth System Science. My current research focuses on better understanding and explaining ongoing changes in Greenland and Antarctica using numerical modelling.

I actually started glaciology by accident… I was trained as an engineer, at Ecole Centrale Paris in France, and was interested in aeronautics and space research. I contacted someone at the NASA Jet Propulsion Laboratory (JPL) in the US to do a six-month internship at the end of my master’s degree, thinking that I would be designing spacecrafts. This person was actually a famous glaciologist (Eric Rignot), which I did not know. He explained that I was knocking on the wrong door, but that he was looking for students to build a new generation ice sheet model. I decided to accept this offer and worked on developing a new ice sheet model (ISSM) from scratch.

Even though this was not what I was anticipating as a career path, I truly loved this experience. My initial six-month internship became a PhD, and I then moved to UCI as a project scientist, before getting a faculty position two years later. Looking back, I feel incredibly lucky to have seized that opportunity. I came to the right place, at the right time, surrounded by wonderful people.

This year you received an Arne Richter Award for Outstanding Early Career Scientists for your innovative research in ice-sheet modelling. Could you give us a quick summary of your work in this area?

The Earth’s ice sheets are losing mass at an increasing rate, causing sea levels to rise, and we still don’t know how quickly they could change over the coming centuries. It is a big uncertainty in sea level rise projections and the only way to reduce this uncertainty is to improve ice flow models, which would help policy makers in terms of coastal planning or choosing mitigation strategies.

I am interested in understanding the interactions of ice and climate by combining state-of-the-art numerical modelling with data collected by satellite and airplanes (remote sensing) or directly on-site (in situ).  Modelling ice sheet flow at the scale of Greenland and Antarctica remains scientifically and technically challenging. Important processes are still poorly understood or missing in models so we have a lot to do.

I have been developing the UCI/JPL Ice Sheet System Model, a new generation, open source, high-resolution, higher-order physics ice sheet model with two colleagues at the Jet Propulsion Laboratory over the past 10 years. I am still actively developing ISSM and it is the primary tool of my research.

More specifically, I am working on improving our understanding of ice sheet dynamics and the interactions between the ice and the other components of the Earth system, as well as improving current data assimilation capability to correctly initialize ice sheet models and capture current trends. My work also involves improving our knowledge of the topography of Greenland and Antarctica’s bedrock (through the development of new algorithms and datasets). Knowing the shape of the ground beneath the two ice sheets is key for understanding how an ice sheet’s grounding line (the point where floating ice meets bedrock) changes and how quickly chunks of ice will break from the sheet, also known as calving.

Steensby Glacier flows around a sharp bend in a deep canyon. (Credit: NASA/ Michael Studinger)

At the General Assembly, you presented a new, comprehensive map of Greenland’s bedrock topography beneath its ice and the surrounding ocean’s depths. What is the importance of this kind of information for scientists?

I ended up working on developing this new map because we could not make any reliable simulations with the bedrock maps that were available a few years ago: they were missing key features, such as deep fjords that extend 10s of km under the ice sheet, ridges that stabilize the retreat, underwater sills (that act as sea floor barriers) that may block warm ocean waters at depth from interacting with the ice, etc.

Subglacial bed topography is probably the most important input parameter in an ice sheet model and remains challenging to measure. The bed controls the flow of ice and its discharge into the ocean through a set of narrow valleys occupied by outlet glaciers. I am hoping that the new product that I developed, called BedMachine, will help reduce the uncertainty in numerical models, and help explain current trends.

3D view of the bed topography and ocean bathymetry of the Greenland Ice Sheet from BedMachine v3 (Credit: Peter Fretwell, BAS)

How did you and your colleagues create this map, and how does it compare to previous models?

The key ingredient in this map, is that a lot of it is based on physics instead of a simple “blind” interpolation. Bedrock elevation is measured by airborne radars, which send electromagnetic pulses into the Earth’s immediate sub-surface and collect information on how this energy is reflected back. By analyzing the echo of the electromagnetic wave, we can determine the ice thickness along the radar’s flight lines. Unfortunately, we cannot determine the topography away from these lines and the bed needs to be interpolated between these flight lines in order to provide complete maps.

During my PhD, I developed a new method to infer the bed topography beneath the ice sheets at high resolution based on the conservation of mass and optimization algorithms. Instead of relying solely on bedrock measurements, I combine them with data on ice flow speed that we get from satellite observations, how much snow falls onto the ice sheet and how much melts, as well as how quickly the ice is thinning or thickening. I then use the principle of conservation of mass to map the bed between flight lines. This method is not free of error, of course! But it does capture features that could not be detected with other existing mapping techniques.

3D view of the ocean bathymetry and ice sheet speed (yellow/red) of Greenland’s Northwest coast (Credit: Mathieu Morlighem, UCI)

What are some of the largest discoveries that have been made with this model? 

Looking at the bed topography alone, we found that many fjords beneath the ice, all around Greenland, extend for 10s and 100s of kilometers in some cases and remain below sea level. Scientists had previously thought some years ago that the glaciers would not have to retreat much to reach higher ground, subsequently avoiding additional ice melt from exposure to warmer ocean currents. However, with this new description of the bed under the ice sheet, we see that this is not true. Many glaciers will not detach from the ocean any time soon, and so the ice sheet is more vulnerable to ice melt than we thought.

More recently, a team of geologists in Denmark discovered a meteorite impact crater hidden underneath the ice sheet! I initially thought that it was an artifact of the map, but it is actually a very real feature.

More importantly maybe, this map has been developed by an ice sheet modeller, for ice sheet modellers, in order to improve the reliability of numerical simulations. There are still many places where it has to be improved, but the models are now really starting to look promising: we not only understand the variability in changes in ice dynamics and retreat all around the ice sheet thanks to this map, we are now able to model it! We still have a long way to go, but it is an exciting time to be in this field.

Interview by Olivia Trani, EGU Communications Officer

Geosciences Column: How fast are Greenland’s glaciers melting into the sea?

Geosciences Column: How fast are Greenland’s glaciers melting into the sea?

The Greenland ice sheet is undergoing rapid change, and nowhere more so than at its margins, where large outlet glaciers reach sea level. Because these glaciers are fed by very large reservoirs of ice, they don’t just flow to the coast, but can extend many kilometres out into the ocean. Here, the ice – being lighter than water – will float, but remain connected to the ice on the mainland. This phenomenon is called an ice shelf or, if it is confined to a relatively narrow fjord, an ice tongue. Ice shelves currently exist in Antarctica as well as in high Arctic Canada and Greenland.

Ice shelves already float on the ocean so that their melting does not affect sea level, but they are a crucial part of a glacier’s architecture. The mass of an ice shelf, as well as any contact points with fjord walls, mean that it acts as a buttress for the rest of the glacier, slowing down its flow speed and stabilising it. When ice shelves melt, therefore, this can lead to the whole glacier system behind them flowing faster and thus delivering more land-based ice to the ocean.

Ice shelves lose mass as icebergs calve off at their seaward end, and through melting on their surface – but, unlike glaciers on land, they are with the ocean below. This ice-ocean interface is an important source of melting for a number of glaciers in northern Greenland; instead of the large volume of icebergs produced by many glaciers further south, the large ice tongues reaching into the ocean mean that a lot of ice is instead lost through submarine melting.

This ice-ocean interface is an environment that was, until recently, very difficult to accurately observe and study, and accordingly there is relatively little data on the impact of submarine melting on ice shelves. But the changes that take place here, at the ice-ocean interface, can have important implications for the entire glacier system, as well as for the ice sheet as a whole.

Over the last 30 years, a number of Arctic ice shelves and ice tongues have dramatically shrunk or disappeared entirely. In the Canadian Arctic, the Ellesmere ice shelf broke up into a number of smaller shelves over the course of the 20th century, most of which are continuing to shrink. In Greenland, meanwhile, the dramatic retreat of the Jakobshavn Glacier’s ice tongue during the 2000s has been particularly well documented.

The largest remaining ice tongues in Greenland are now all located in the far north of the island. But even here, at nearly 80°N and beyond, ice tongues are changing rapidly. Warming air temperatures probably play a role in this development, but submarine melting is thought to be the key driver of these rapid changes.

Submarine melting of ice tongues thus appears to be an important variable in ice-sheet dynamics. A new study in the EGU’s open access journal The Cryosphere has now used satellite imagery to produce a detailed map of submarine melt under the three largest ice tongues in northern Greenland. They are the ones belonging to Petermann and Ryder Glaciers in far northwestern Greenland and 79N Glacier – named after the latitude of its location – in the northeast of the island. Each of these ice shelves extends dozens of kilometres from where the glacier stops resting on bedrock and begins to float (the so-called grounding line) and is up to several hundred metres thick.

The locations of Petermann (PG), Ryder (RG) and 79N Glaciers in northern Greenland. From Wilson et al. (2017).

Previous attempts to estimate submarine melt rates relied on an assumption of steady state: that the ice shelf is becoming neither thicker nor thinner. Given the recent changes in all these ice shelves and the glaciers above them, this is not a tenable assumption in this case. Petermann and Ryder Glaciers, in particular, have recently experienced large calving events that were probably related to unusual melt patterns under the waterline.

Lead author Nat Wilson, a PhD student at MIT and Woods Hole Oceanographic Institution, and his colleagues used satellite images spanning four years to create a number of digital elevation models of the Petermann, Ryder and 79N ice shelves. A digital elevation model, or DEM, is a three-dimensional representation of a surface created – in this case – from satellite-based elevation data. By comparing DEMs from different points in time to each other, the team could deduct changes in the height – and therefore volume – of the ice shelves. This method also allowed them to track visible features of the glaciers between images from different years, providing estimates of how fast the ice was flowing down into the ocean.

However, using digital elevation models in a marine setting is not always a straightforward matter. Tides can affect the elevation of ice shelves by a significant amount, especially as the distance from the grounding line increases, and their effect needed to be accounted for in the results. Similarly, the team had to account for the changes on the surface of the ice shelf, where snowfall and melting can affect its volume.

What Wilson and his colleagues were left with was a map of melt rates across the ice shelves. In some respects, the findings were unsurprising. Melt rates were greatest near the coast, where the ice shelves were thickest, because at these points they would be in contact with the ocean at depths of several hundred metres. At such depths, fjords around Greenland often contain warm, dense water that flows in from the continental shelf and contributes to rapid ice melt. As the ice shelves thin towards their outer edges, they are in contact with shallower, colder water that doesn’t melt the ice as quickly.

Submarine melt rates at Greenland’s largest ice tongues are shown in colour shading; the arrows show the direction of ice flow. PG – Petermann Glacier; RG -Ryder Glacier. From Wilson et al. (2017).

All three ice shelves lost between 40-60m per year to submarine melting at their thickest points, while this decreased to about 10m per year in thinner sections. This equates to billions of tonnes of ice melting in contact with the ocean. Each of the ice shelves lost at least five times as much ice to melting underwater than to melting on the surface. This highlights what an important contribution submarine melting makes to the mass balance of Greenland’s ice shelves, and that this remote environment is deserving of our interest and study.

The team found that at Ryder Glacier’s ice shelf, mass loss from melting (from both above and below) is not significantly greater than the amount of ice entering the ice shelf from land: the ice shelf appears to be relatively stable for the time being. The situation is similar at Petermann Glacier, although its ice shelf has been in rapid retreat and lost some 250 km in the decade leading up to 2010. With the extra submarine melting from that area, melting would likely have exceeded incoming ice! It remains to be seen whether Petermann Glacier and its ice shelf will stabilise in their new configuration.

Finally, at 79N Glacier, the results indicate the ice shelf is losing mass faster than it is replenished from upstream. The ice tongue loses some 1.3% of its mass to melting each year – and that’s before iceberg calving is included in the equation. This finding is consistent with satellite imagery that suggests that the ice shelf at 79N has been thinning in recent decades.

This new study shows that there is considerable variability in submarine melting of ice shelves, both in space and in time. 79N glacier’s ice shelf – the biggest one remaining in Greenland – exhibited the highest mass deficit in this study, suggesting that we may see major changes in this glacier in future. With this type of melt making up for the bulk of mass loss of northern Greenland’s ice shelves, its accurate prediction plays an important role in understanding how these huge glaciers – and the whole ice sheet itself – will change in coming years.

By Jon Fuhrmann, freelance science writer

References

Wilson, N., Straneo, F., and Heimbach, P.: Satellite-derived submarine melt rates and mass balance (2011–2015) for Greenland’s largest remaining ice tongues, The Cryosphere, 11, 2773-2782, https://doi.org/10.5194/tc-11-2773-2017, 2017.

Hodgson, D. A. First synchronous retreat of ice shelves marks a new phase of polar deglaciation. Proc. Natl. Acad. Sci. U. S. A. 108, 18859-18860, doi:10.1073/pnas.1116515108 (2011).

Münchow, A., L. Padman, P. Washam, and K.W. Nicholls. 2016. The ice shelf of Petermann Gletscher, North Greenland, and its connection to the Arctic and Atlantic OceansOceanography 29(4):84–95, https://doi.org/10.5670/oceanog.2016.101.

Reeh N. (2017) Greenland Ice Shelves and Ice Tongues. In: Copland L., Mueller D. (eds) Arctic Ice Shelves and Ice Islands. Springer Polar Sciences. Springer, Dordrecht. https://doi.org/10.1007/978-94-024-1101-0_4

Truffer, M., and R. J. Motyka, Where glaciers meet water: Subaqueous melt and its relevance to glaciers in various settings, Rev. Geophys., 54, 220– 239. doi:10.1002/2015RG000494,  (2016)