Cryospheric Sciences

Did you know?

Did you know… about the fluctuating past of north-east Greenland?

Did you know… about the fluctuating past of north-east Greenland?

Recent geological data shows that during a very cold phase of our Earth’s climate (between 40,000 and 26,000 years ago), there was a huge expansion of polar ice sheets, yet the north-eastern part of the Greenland ice sheet was less extensive than today. How could this have occurred? In this post we shed light on the potential causes of this ice sheet behaviour.

What do we know about present-day north-east Greenland?

The North East Greenland Ice Stream is one of the most interesting icy features of the present-day Greenland Ice Sheet. Extending for more than 600 km from the ice-sheet interior to the ocean, this ice stream drains almost 12% of the whole ice sheet and terminates through a system of fast-flowing outlet glaciers, known as Nioghalvfjerdsfjorden glacier (also called 79N), Zachariae Isstrøm and Storstrømmen (Fig. 1).

These three outlet glaciers belong to the disappearing category of marine-terminating glaciers, i.e. glaciers flowing into the ocean that possess an ice tongue (or small ice shelf) floating on the seawater. These tongues are connected to the main trunk of the glacier that is grounded over the continent, and the zone where the glacier starts to float is called the grounding line. The grounding line can move forward (or backward) if the glacier gains (loses) sufficient mass and can then expand (or retreat).

But how can a glacier gain mass? This is possible if the glacier mass balance is positive, i.e. if the glacier accumulates more ice (through snow precipitation) than is melted away (through surface and basal melting). Therefore, if surface air or oceanic temperatures increase, the glacier will likely lose mass, reducing its floating tongue and potentially making the grounding line retreat.

Recently, the 79N and Zachariae Isstrøm glaciers have lost a huge amount of ice mass via this mechanism due to rising temperatures. This suggests that this part of the ice sheet is very sensitive to climate change. Under a warmer climate, these glaciers could even lose their whole floating tongue, potentially causing irreversible consequences to the stability of the region.

How did north-east Greenland behave in the ancient past?

On millennial timescales, the north-east Greenland has had a very animated past too. Understanding the history of polar ice sheets requires a lot of effort both collecting and analysing paleoclimatic data (i.e. data archives providing information on the past evolution and climate of the ice sheet, e.g. from ice cores, marine sediments, …), which are usually very sparse, and running numerical simulations. Thanks to both disciplines, we know today that during the last glacial period (when Homo Sapiens were still in the Stone Age), the north-eastern part of the ice sheet expanded by hundreds of kilometers away from the coast into the ocean. During its maximum phase (in the Last Glacial Maximum, 21,000 years ago), this north-eastern part very likely reached the continental shelf break, the edge of shallow waters (see Fig. 1). However, very little information is available on the evolution of north-east Greenland before the Last Glacial Maximum.

One of the most interesting contribution to this gap in knowledge came less than two years ago from a group of researchers headed by Dr. Nicolaj Larsen, who ran an expedition at 79°N for new evidence. They focused their research on marine material found on moraines (eroded material deposited by the glacier). There, they collected and dated fragments of marine shells and were able to pinpoint when the moraines were formed, indicating the timing of glacier retreat. Their new data suggested that these moraines were a few thousand years older than the Last Glacial Maximum, meaning that the edge of Zachariae Isstrøm was located at least 20 km upstream of its current position at that date (Fig. 2, right).

Fig. 2: Left: Sample site at Zachariae Isstrøm [Credit: Supplementary Material of Larsen et al., 2018]. Right: Reconstruction of the North East Greenland Ice Stream grounding line distance from its present-day position for the last 45,000 years in the Marine Isotope Stage 3 (MIS-3) and Last Glacial Maximum (LGM) [Credit: I. Tabone, data from Larsen et al., 2018].

How could north-east Greenland be less extensive than today during a cold climate?

How could it be possible that these glaciers were less extensive than today, despite the cold, dry climatic conditions during the last glacial period? Well, it is a complex story! We are actually talking about a glacial time interval called Marine Isotope Stage 3 (~ 60,000-25,000 years ago), a period studded with several rapid abrupt climate events in which polar latitudes received larger incoming summer radiation than during the rest of the glacial period.

During Marine Isotope Stage 3, summer air temperatures were about 6 to 8ºC higher than during the Last Glacial Maximum (Fig 3, middle panel). This is likely due to long-term variations in incoming solar radiation associated with changes in the Earth’s orbit. Marine Isotope Stage 3 was also very dry, with low accumulation (snowfall) rates in northern Greenland (Fig 3, low panel). These warm temperatures and low accumulation could be a possible explanation for the expansion of the north-east Greenland ice sheet during Marine Isotope Stage 3 (Larsen et al., 2018). However, summer air temperatures at high latitudes were still well below 0°C during this time, so it is unlikely that they could have strongly affected ice mass loss in the region. The conclusion is: the atmosphere alone probably could not be the only climatic driver of this strong ice fluctuation at 79°N.

Fig. 3: Grounding line (GL) distance from the present-day (PD) position (a); summer temperature at 79N (b); accumulation rate at North Greenland Eemian Ice Drilling (NEEM) ice-core site [Credit: I. Tabone, data from Rassmussen et al., 2013, figure inspired from Larsen et al., 2018].

So it’s not the atmosphere… What about the ocean?

As the ocean has been recently argued to have an important role in the present-day retreat of the North East Greenland Ice Stream (see this study, this one and this one), it is reasonable to think about the ocean as a possible player in this large past fluctuation. Thanks to a 3D ice-sheet model, we performed simulations of north-east Greenland evolution in the past to assess how sensitive the ice-sheet margins were to variable oceanic conditions during the last glacial cycle. We represented these different oceanic states in the model by imposing melt at the grounding line and at the ice-shelf bases (also called submarine melt). Past submarine melt rates are inferred from past oceanic temperatures changes, because any increase in oceanic temperature results in increased submarine melting. Over long-term timescales, this means submarine melt rates were higher during Marine Isotope Stage 3 compared to during the Last Glacial Maximum.

Results of these model runs are pretty neat: with sufficiently high submarine melting along its margins, the North East Greenland Ice Stream retreats several tens of kilometres upstream from its maximum glacial position. Importantly, this suggests that increasing oceanic temperatures during Marine Isotope Stage 3 could have driven this large instability in the north-eastern ice-sheet margin during the last glacial. This is likely to have caused large reorganisations of the entire region and major ice discharge into the ocean (here is an animation showing the modelled north-east Greenland evolution during the last 45,000 years).

Fig. 4: Evolution of the north-east Greenland grounding-line (GL) distance from its present-day (PD) position simulated by our ice-sheet model. The model was run with no submarine melting (red line) and with progressively higher melting (other coloured lines). The dashed black line shows the reconstruction by Larsen et al. (2018). The three horizontal black dotted lines show the today’s NEGIS grounding-line position (0 km) and the maximum (300 km ± 50 km) reconstructed advance of the north-east Greenland during the Last Glacial Maximum according to Funder et al. (2011) [Credit: I. Tabone].

More investigation is needed!

Today we are aware that the north-east Greenland ice sheet is one of the most vulnerable regions of the ice sheet to current climate change. Figuring out its past evolution will help to understand its behaviour in a warming world, and its importance in the future stability of the entire Greenland ice sheet. However, as our study is the first attempt to look at the causes of this anomalous retreat from a modelling point of view, further work is needed and many questions are still unanswered. Was the ocean the major player in this past fluctuation? To what extent were surface air conditions also a factor? How much abrupt atmospheric warming events have influenced this margin fluctuation at smaller timescales? Further modelling work and observations at the North East Greenland Ice Stream are needed to unravel this icy riddle…

Further reading

Edited by Jenny Arthur and Clara Burgard

Ilaria Tabone is a Postdoc Researcher at the University Complutense of Madrid (Spain) in the Paleoclimatic Analysis and Modelling (PalMA) research group. She investigates the evolution of the Greenland Ice Sheet in the past glacial-interglacial cycles by working with ice-sheet models of continental scale. Her research focuses on ice-ocean and ice-atmosphere interactions. Contact Email:

Did you know? – Proglacial lakes accelerate glacier retreat!

Hooker glacier and its proglacial lake, Aoraki/Mt Cook National Park. [Credit: Jenna L. Sutherland]

In a global context, New Zealand’s small mountain glaciers often get overlooked and yet they are a beautiful part of New Zealand’s landscape. They are the water towers for the South Island and an essential part of its tourism, thanks to a few undeniable heroes (Frans Josef and Fox Glaciers), but sadly, they may not be as prominent in the future. In this post we review the state of modern glaciation in New Zealand, explain why glacier-lake interactions are important and why we need to turn to the past for answers.

Glaciers in New Zealand are losing mass…

New Zealand is home to over 3000 glaciers. However, their mass is quickly declining. New Zealand has one of the oldest and most continuous records of annual ice volume, provided by the End of Summer Snowline Survey, and pioneered by the late Trevor Chinn. Now undertaken by NIWA, aerial surveys on 51 index glaciers have been carried out every year for more than 4 decades (records began in 1977). The surveys document the snowline on the glaciers at the end of every summer, which give us a timeline of glacier-climate interaction. Glacier snowlines, also known as Equilibrium Line Altitudes (ELAs), provide a direct measurement of a glaciers health. They record how much of the previous winter’s snow remains at the end of summer, contributing to long-term glacier ice accumulation. The higher the ELA, the less winter snow remaining, indicating the glacier has shrunk. If the ELA is lower, a larger amount of winter snow remains and the glacier has increased in size. An almost continuous trend from this 40 year archive reveals that glaciers in New Zealand are retreating rapidly – they are getting shorter and losing mass. The significant retreat of Southern Alps glaciers has accelerated over the last decade and recent studies have found that a third of total ice volume has been lost since records began.

…leading to the development of proglacial lakes

The formation of proglacial lakes in mountainous regions, specifically those in contact with the ice margin, is one of the consequences of glacier recession. As a glacier margin retreats, meltwater is impounded in the topographic low between the ice front and the abandoned moraine ridges. The increasing number and size of proglacial lakes is one of the most visually obvious effects of present deglaciation in New Zealand. Most of the existing ice-contact lakes in New Zealand formed when glaciers began to recede from large moraine ridges constructed during the Little Ice Age (LIA), about 150 years ago.

Proglacial lakes currently represent 38 % of the total number of all lakes in New Zealand, and over a third of the country’s perennial ice is contained within lake-calving glaciers. Back in the early 1970s, the Tasman Lake did not exist and the glacier terminated against its outwash sediments. Today, just 40 years later, it now terminates in a large proglacial lake that is more than 8 km long, and over 200 m deep (Figure 1). Since the lake formed, the Tasman Glacier has retreated at an average of 180 m per year and is now in a state of rapid recession.

Figure 1. The Tasman Glacier and its associated proglacial lake. Don’t be fooled by how thin the glacier terminus looks from this perspective, the ice cliff is 50 m tall. The lower part of the glacier is heavily debris-covered but you can see the exposed glacier surface at the head of the valley [credit: Jenna L. Sutherland]

Lake formation is strongly linked to glacier dynamics (Figure 2). The depth of water at the ice margin determines:

  • the distance underneath the glacier that water can travel
  • the rate at which ice calves (or breaks off) from the terminus

Together, these factors increase the speed of ice flow and increase mass loss from the glacier system. The relatively warmer water of the lake compared to the glacier ice also causes thermally-induced melting. An ice-marginal lake can therefore cause glacier margins to fluctuate back and forth, which in turn can cause the speed of the glacier and its mass balance to become partially separated from the climatic signal.

Figure 2. Interactions between the glacier and its proglacial lake. Forces acting upon the glacier are shown in italicised text and with black arrows. The processes between the lake and the glacier are in black text and grey dashed lines [Credit: Jonathan Carrivick, modified from Carrivick and Tweed, 2013]

New Zealand is witnessing unprecedented glacier recession together with lake expansion. It is easy to link one to the other, but the relationship between ice, the development of the lakes, and climate are much more complicated than this. The shift from a land-terminating to a lake-terminating glacier is a defining point in deglaciating environments. This transition represents a threshold and/or tipping point that is critical for the future evolution of the glacier and therefore crucial for us to understand.  Could such glaciers become unstable and catastrophically collapse? Has this happened before? Where and what is this threshold? The current state of knowledge in New Zealand lies in monitoring temporal and spatial evolution of both glaciers and their proglacial lakes. Despite the remarkable record, the short duration of observations in New Zealand (just over 40 years) means that it is difficult to differentiate between natural cycles and occurrences, and dynamic behaviour that is beyond the norm. The answer to such questions lies in turning to the palaeo-record to find out how the Southern Alps ice field has behaved over the last few thousand years or more. It is vital to determine what thresholds control glacier-lake behaviour, and whether these have been crossed in the past. By gaining a deeper understanding of past processes, rates of change, thinning and retreat, as well as previous temperatures and environmental conditions, we will be better placed to understand how the Southern Alps could behave in the future.

Using the past to better understand the future….

The Quaternary record in New Zealand bears witness to the existence of proglacial lakes associated with retreat since the Last Glacial Maximum (LGM; the period between 21,000 and 18,000 years ago). During the LGM, the ice extent and volume was much larger; outlet glaciers advanced beyond the present coastline along the west coast of the South Island (Figure 3). Just like we see in the modern day, the glaciers receded when temperatures began to warm and meltwater ponded, forming large and deep proglacial lakes in contact with the ice margin. Past glacier behaviour in New Zealand has been derived from mapping and dating former ice extents. The maximum extent of ice during previous glaciations is now well constrained, allowing us to determine the speed of glacier recession and thinning. A study has shown that rapid ice recession in the first few millennia saw glacier trunks thin by at least several hundred metres, with implied terminal recession by as much of 40% of the overall glacier length.

Figure 3. The South Island of New Zealand with modern-day glaciers mapped in white and LGM ice extent in light grey. Present day lakes (blue) would have once been in contact with retreating ice margins [Credit: Sutherland et al., 2019]

It has been suggested that the retreat of glaciers in the Southern Alps, immediately after the LGM, was relatively rapid not only because of a warming climate, but also because of the widespread formation of large proglacial lakes accelerating recession. However, despite knowing that large proglacial lakes existed, we have no understanding as to what effect they had on glacier retreat and how much influence these proglacial lakes had on glacier dynamics in combination with climatic warming. The interactions between proglacial lakes and glacier dynamics have not yet been quantified. This is largely a consequence of poor accessibility to modern glacier-lake margins but also because proglacial lakes are currently ignored in ice sheet models (Figure 4). They are treated as an entirely separate component, or, as is more often the case, not at all.

Figure 4. Main components of most ice sheet models (blue), often coupled to other models (green), e.g. ocean and atmosphere models. Note the absence of proglacial lakes as a component of the ice sheet model despite their importance in influencing glacier dynamics [Credit: Jenna. L Sutherland]

…But not without some computer modelling too!

Owing to numerous studies that have dated the recession of ice since 18,000 yrs ago we have a pretty good idea that glacier recession was fast. However, as a concern for the future, what we are more interested in is constraining previous ice volumes and, more specifically, the mechanisms of loss. As well as rates of change, we also need to understand processes of change and how the ice evolved through time. In order to examine and better understand the external (or internal, as the case may be) forcing that drove this rapid recession, we need to resort to relating glacier and climatic changes in numerical ice sheet or glacier simulations.

The physics that govern a glaciers mass balance (the difference between snow gain and snow melt) as well as ice flow are complex. The equations depend on ice thickness, speed, ice and air temperature, elevation, as well as many other factors. Fortunately, these relationships are reasonably well understood and provide the basis of many numerical ice sheet models. We then feed the model with input data, such as topography and past climate, to drive a numerical ice sheet simulation. This enables us to investigate detailed mechanisms driving ice sheet change, such as those between a glacier and its lake and we can become confident in such models when they provide a good fit to the geological record (what we see on the earth’s surface).

The future

Understanding the formation and evolution of proglacial lakes and their outlet glaciers through time can provide insights into the behaviour of glaciers and ice sheets to help us anticipate some of the impacts of present and future deglaciation. Although my research is concerned with just one small valley glacier in a specific region, it is a small step towards a wider understanding of the glacier-lake interaction phenomenon. Of course, every proglacial lake is different, just as every mountain glacier is, but if we can get a handle on what effects a lake had on its glacier in the past, the importance of proglacial lakes might be realised for other regions and more seriously considered when it comes to interpreting glacier response to climatic change.

Further Reading

Edited by Andy Emery

Jenna Sutherland is a final year PhD student in the Department of Geography at the University of Leeds, UK. Her research is focused on the interaction of proglacial lakes and their outlet glaciers during the Last Glacial Maximum in New Zealand, specifically by simulating the presence of proglacial lakes in a numerical ice sheet model and relating these experiments to the sediment-landform record. Her broader interests lie in palaeo-glacial environments. She tweets from @Jennalo13




Did you know? – Storms can make Arctic sea ice disappear even faster

Did you know? – Storms can make Arctic sea ice disappear even faster

The increase in air and water temperature due to climate change drives the retreat in the Arctic sea-ice cover. During summer, when sunlight reaches the Arctic, the absorption of heat by the dark ocean water enhances the sea-ice melt through the ice-albedo feedback. During winter, when sunlight does not reach the Arctic, another feedback is at work, as storms enhance the energy transfer between air, ice and water…

How can storms enhance sea-ice melting?

In summer, when the Arctic sea-ice cover is close to its minimum extent, a large storm can rapidly lead to a further decrease in the ice cover. In the case of a storm, the ice breaks up and is pushed together due to the wind-induced waves that quickly develop in the vast areas of open water. Once the storm is over, the resulting small ice pieces drift apart and melt faster than the larger ice pieces would have melted before the storm.

During winter, there is hardly any open water in the Arctic Ocean. An exception is the Barents Sea, where warm ocean water flows in from the North Atlantic and sinks under the surface when it meets the sea ice. Because this inflowing water got warmer over the past couple of decades, the Barents Sea has become nearly ice free in the winter (Polyakov et al, 2017). The “Whaler’s Bay“, north of Svalbard, is experiencing a similar evolution. In this area of open water, where the banks of the bay are formed not by land, but by sea ice, warm ocean water brought by currents sinks under the ice as well. While the effect of the water warming leads to a clear retreat of the ice in both the Barents Sea and Whaler’s Bay, another phenomenon can lead to an even faster retreat: episodic winter storms (e.g. Boisvert et al, 2016) that bring in warm air and push the sea ice northwards.

Until now, scientists observed both processes mainly using ocean moorings and satellite remote sensing sources as winter in-situ observations in this area are very rare.

Fig. 2: N-ICE2015 was based in an ice camp set around RV Lance. The ship was assisted to approx. 84N by an ice breaker and then left to drift out towards the Whaler’s Bay slowly with the sea ice and ocean surface current. [Credit: Paul Dodd, Norwegian Polar Institute].

Observations from the N-ICE 2015 campaign

In early 2015, the Norwegian Polar Institute led a half-a-year-long international expedition in the sea ice north of the Whaler’s Bay (Norwegian Young Sea Ice Cruise, N-ICE2015). During the winter part of the expedition, when air temperatures were typically below -20ºC, six powerful storms brought strong winds and mild temperatures into the region. The atmosphere-ice-ocean measurements recorded by the expedition revealed complex processes of energy transfer resulting in complex combination of thin and snow-covered sea ice, numerous leads, and pressure ridges (see this previous post about sea-ice dynamics).

The observations from N-ICE2015 show that early winter storms deposited a thick snow cover on the ice. Because snow is an excellent insulator, heat from the ocean cannot escape into the atmosphere and the water right under the ice does not cool enough to form further sea ice. This way, the ice stays relatively thin. So thin, that it yields the weight of the deep snow cover and gets submerged under the sea level. Generally, ice floes can be kilometers-wide and have thick steep edges built of pressure ridges that prevent water seeping into the snow from the sides of the floes.

However, each new storm during the observation period came with strong southerly winds that pushed the ice floes northwards with such force that they cracked in much smaller pieces. After the center of the storm passed over the sea ice, the wind direction reverted to southward and the ice stretched again towards Whaler’s Bay. The cracked floes were then less protected from water seeping into the snow from the cracks. Some large cracks even developed into leads with open water. After the storms, northerly winds brought back cold temperatures and cracks, leads, and flooded snow froze rapidly.

At the same time, the motion of the ice floes led to vertical mixing at the ocean’s surface. If this mixing happened right above the sinking warm water, some of the warm water was brought up to the surface and melted the ice from below.

Figure 3: The Arctic Winter Storm Cycle [Adapted from Graham et al. (2019)].

The shallow warm ocean water currents are unique to the Barents Sea and Whaler’s Bay in the Arctic. Further in the Central Arctic, the cold ocean has much less potential to melt sea ice during winter. Still, many powerful storms can reach beyond the North Pole and cross the whole Arctic Ocean. This means that the abundant snowfall, breaking-up of the ice floes, flooding of snow and opening of the leads can be common in large parts of the Arctic Ocean in the winter (for details see Fig. 3 and Graham et al, 2019). This way, a row of passing storms cannot only build a deep snow cover, but also create a broken-up ‘ice-scape’ that is susceptible to melt faster during summer. Moreover, abundance of areas of thin ice which transforms into leads can let enough light into the ocean to trigger large algae bloom earlier than usual in the season (Assmy et al, 2017). In summary, winter storms can make a large cumulative impact that lasts far beyond the short duration of a single storm.

More observations are needed!

Although it provided a lot of new understanding on the importance of storms for the sea-ice evolution, the N-ICE2015 brought some limitations. For example, it ended in June, before we could observe summer melt processes and it was limited to the area north of the Whaler’s Bay. To confirm and further explore the Arctic atmosphere-ice-ocean-ecosystem processes, the international scientific community is launching an even larger expedition in 2019 and 2020.

The Multidisciplinary drifting Observatory for the Study of Arctic Climate (MOSAiC) will be a huge international effort with hundreds of scientists involved in field work, data analysis and research. Its main goal is to collect the data from the ‘new Arctic’ where the sea ice is much thinner than earlier decades. Such data can help improve the climate model projections that still under-represent the recent decline in sea ice extent and volume. MOSAiC will provide observations spanning over the full annual cycle of sea ice, from freeze-up in fall 2019 to melt in summer 2020. Geographically, the expedition will cover vast distances by drifting with the sea ice from the Central Arctic towards the North Atlantic. MOSAiC observations will build on the results from N-ICE2015 and will measure the effects of storms also in the Central Arctic.

Further reading

Edited by Clara Burgard

Polona Itkin is a researcher at the UiT The Arctic University of Norway, Tromsø. She investigates the sea ice dynamics of the Arctic Ocean and its connection to the sea ice thickness and snow depth. In her work she combines the information from in-situ observations, remote sensing and numerical modeling. Polona was a post-doctoral researcher at the Norwegian Polar Institute and one of many early career scientists involved with N-ICE2015. The expedition was highlighted also in their social media project ‘oceanseaiceNPI’:,,, contact Email: