Coupling of basal hydrology and basal sliding within Elmer/Ice

Olivier Gagliardini, Mauro Werder

Corresponding author: Olivier Gagliardini

Corresponding author e-mail: olivier.gagliardini@ujf-grenoble.fr

The basal sliding of glaciers is modulated by the presence of basal water and more precisely its pressure. On one hand, where the basal water pressure increases, the ice friction on the bedrock is reduced and the basal sliding velocity increases. On the other hand, an increase in basal sliding will influence the development of the basal hydrology system. These processes together with the short-term meltwater forcing lead to non-steady ice velocities and have a seasonal or even diurnal signature. In this study, a basal hydrology model is coupled with an ice flow model through a water-pressure-dependent friction law and through a sliding-dependent cavity opening rate. The basal hydrology model includes an inefficient cavity-type water sheet and a network of efficient discrete channels. Both systems are connected and evolve in time in response to the water inputs. The prognostic equations for ice flow and the hydrology model are implemented in the open-source, finite-element, ice-sheet/ice-flow model Elmer/Ice. The coupling of basal hydrology and ice flow is studied for different glacier geometries and external forcing having different amplitude and temporal signatures. The three levels of coupling, i.e. (1) no coupling, (2) hydrology and ice velocity coupling and (3) hydrology, ice velocity and free surface evolution coupling, are compared in terms of the evolution of the hydrological network and the modulation of the mean velocities.


Heterogeneous status of glacial ice-contacted lakes in the central Himalaya

Qiao Liu, Yong Nie, Shiyin Liu

Corresponding author: Qiao Liu

Corresponding author e-mail: liuqiao@imde.ac.cn

Widespread expansion of glacial lakes around the Himalaya, which are sources for hazard risks in their downstream valleys due to the potential of glacial lake outburst flood (GLOF), has been widely reported during the past decades. Among all types of glacial lake, those lakes contacted with the terminals of modern glaciers are generally found experiencing the most remarkable area increases. One of the most important reasons of that is due to the coupled processes, such as calving, between the lake growth and ice tongue retreat. Thermal absorption and convection of lake water are important factors impacting on the calving processes at the ice cliff or sub-marine melting under the supra-ponded water bodies. However, the current status of the ice-contacted lakes shows great differences, e.g. in their area expansion histories and potentials, due to the complexes between the lake-glacial interactions and/or the local topographic limitations. After experienced the phases of rapid lake growth and terminal retreat, despite the contacting and calving still exist, the position of the calving line may be balanced by the positive advances of the ice tongue or, in some cases, further lake growth is limited by the steep terrain behind its upper margin. We investigate the heterogeneous status of the ice-contacted glacial lakes in the central Himalaya based on a recently completed glacial lake inventory in this region using muti-year Landsat images. From the 1314 investigated glacial lakes (2010) in the central Himalaya, 328 ice-contacted lakes are extracted and their numbers, areas and expansion rates and contributions are statistically analyzed. We will discuss the impacts of different local geomorphology (e.g. slope, lake-basin shape and valley aspect) and glacier characters (e.g. debris cover, velocity and mass balance) on their heterogeneous changes. These related factors are important for both the prediction of lake and glacier changes and the evaluation of GLOF hazards in the future.


Heat balance of supraglacial lakes in western Dronning Maud Land

Matti Leppäranta, Elisa Lindgren, Lauri Arvola

Corresponding author: Matti Leppäranta

Corresponding author e-mail: matti.lepparanta@helsinki.fi

Supraglacial lakes form in blue-ice regions from penetration of solar radiation into the ice in summer. The thermodynamics of one seasonal supraglacial lake was examined based on field data in three austral summers in 2004–15 in western Dronning Maud Land. The lake body consisted of two layers, each about 1 m thick: an upper layer with a thin ice layer on top and main body of liquid water, and a lower layer containing slush and hard ice sub-layers. A sediment-rich slush pocket was found at the bottom. The summer heat budget data collection started when the lake was still solid ice. Albedo of the lake was 0.4–0.6 and light attenuation coefficient was 0.5–0.7 m–1. Latent heat loss and decrease of the surface ice layer by sublimation were important factors of the mass and heat balance. The formation and the depth scale of supraglacial lakes in the region are determined by the light attenuation distance and thermal diffusion coefficient, limiting the growth to less than about 1.5 m in one summer. The potential winter growth is more and thus the lakes freeze-up in winter in the present climatic conditions. The field data showed interannual variations by the factor of two in the liquid water storage of the lake.


The role of isolated drainage in the seasonal evolution of the Greenland subglacial hydrologic system

Matthew Hoffman

Corresponding author: Matthew Hoffman

Corresponding author e-mail: mhoffman@lanl.gov

The Greenland ice sheet speeds up every summer as melt from the surface penetrates km-thick ice through moulins, but observations show that continued meltwater inputs lead to a decay in the speed-up and late summer velocities that can be lower than the winter speed. While this decreasing ice dynamic response to meltwater forcing has often been attributed to channelization of the subglacial drainage system, recent observations of water pressure in moulins indicate that the efficiency of the moulin-channel system does not increase during the second half of summer, failing to explain the concurrent drop in ice velocity. Simultaneous observations of lowering water pressure in boreholes sampling poorly connected regions of the bed suggest that changes in isolated regions of the bed over summer modulate the overall basal traction of the ice sheet. We explore the role of isolated drainage using a coupled model of subglacial hydrology and ice dynamics. The subglacial hydrology model includes components for distributed linked-cavity drainage, channelized drainage, and weakly connected isolated drainage. Passive cavity opening in the isolated drainage region is capable of reproducing observed diurnal variations in borehole water pressure. Further gradual drainage of water from the isolated cavities leads to a seasonal reduction in water pressure there and an associated decrease in ice speed. Finally, the model demonstrates that this summer leakage from the isolated regions of the bed can explain observations of an inverse relationship between summer melt and ice speed in the following autumn, as the timescale for re-pressurizing isolated cavities is slow.


Jökulhlaups at A.P. Olsen Ice Cap, NE Greenland: first results of the geodetic and seismic monitoring in 2012

Daniel Binder, Geo Boffi, Bernd Kulessa, Stefan Mertl, Wolfgang Schöner, Gernot Weyss, Andreas Wieser

Corresponding author: Daniel Binder

Corresponding author e-mail: daniel.binder@zamg.ac.at

A.P. Olsen Ice Cap is a peripheral ice cap in NE Greenland, situated close to the Zackenberg Research Station. Regular flood waves have been monitored for the Zackenberg River, which directly passes the research station. The registered discharge is characterized by a linear increase in between several hours. The origin of the flood waves is an ice-dammed lake impounded by the SE outlet glacier of A.P. Olsen Ice Cap. In spring 2012 geodetic and seismic monitoring was installed in the ablation zone of the outlet glacier to monitor its response to the drainage event. The geodetic network consisted of low-cost single-frequency GPS receivers with a sampling period of 10 s. In the glacier forefield a base station was installed to differentially correct the five ‘on-ice’ stations. During the outburst itself the GPS data exhibit a total reversal of the flow direction. Further, the GPS data revealed a highly dynamic initiation phase 1 week before the drainage event. The dynamic initiation phase was characterized by a vertical uplift of several centimeters and an acceleration of the horizontal flow velocities. The seismic network consisted of five stations on-ice, co-located to the GPS stations. Three stations were equipped with three-component borehole geophones and the remaining two stations were designed as arrays, consisting of three vertical geophones. All seismic sensors were sunk about 3 m into the ice and covered with a geotextile to avoid an untimely meltout. The eigenfrequency of all sensors was 4.5 Hz and a sampling rate of 500 Hz was chosen. The seismic data generally showed numerous discrete events. The spectrogram analysis uncovered three different seismic phases. The first phase was the quiet phase, which was interrupted by discrete single events. The second phase started at the beginning of June, where a stepwise increase of the seismic activity could be observed. The seismicity steadily increased during the following weeks. In the last week before the outburst itself (phase 3), the spectrogram showed besides a general high seismic activity, spectral harmonics that can be linked to the propagation of the floodwater in cracks, respectively channels. The first results of the geodetic and seismic monitoring of the whole fill-and-drain cycle in 2012 suggest the subglacial infiltration and propagation of the floodwater at least about 1 week before the outburst event itself.


The hydrological impact of rapid ice ablation in a small highly glaciated catchment, Virkísjökull, SE Iceland

Verity Flett, Martin Kirkbride, Jez Everest, Andrew Black, Alan MacDonald

Corresponding author: Verity Flett

Corresponding author e-mail: v.t.flett@dundee.ac.uk

Widespread glacial retreat in Iceland will lead to changes in the hydrology of proglacial rivers. An understanding of the connection between glaciers and rivers is necessary in order to predict and manage the anticipated changes. This paper presents a detailed ice ablation study from the Virkísjökull catchment in SE Iceland. This paper provides a better understanding of the connection between enhanced ice ablation, glacier hypsometry and proglacial river run-off during a rapid retreat phase. Degree-day modelling, river discharge monitoring and glacier hypsometric analysis from the 2013 ablation season was used. Results indicate that a 1°C warming will increase ice surface lowering by 8.5 mm d–1 in the lowest 200 m, which translates as 37% more melt available for river flow. Individual glacier hypsometry is responsible for a high sensitivity to fluctuations in the equilibrium-line altitude, with an increase of 100 m resulting in an increase in total meltwater input into the catchment of 13.4% currently levels between May and September. Virkísjökull is at a critical stage where the greatest increase in meltwater input that will be observed during this deglaciation cycle could occur within the next few years due to an increase in the altitude of the late summer snowline.


Meltwater flow through a rapidly deglaciating glacier and foreland catchment system, Virkísjökull, SE Iceland

Verity Flett, Louise Maurice, Andrew Black, Alan MacDonald, Jez Everest, Martin Kirkbride, Andrew Finlayson

Corresponding author: Verity Flett

Corresponding author e-mail: v.t.flett@dundee.ac.uk

This study characterizes the glacial and proglacial hydrology of a rapidly deglaciating system at Virkísjökull in SE Iceland, to determine the water flux through the glacier and proglacial area. This was achieved using dye tracer tests, river discharge measurements and studies of conduits within the foreland and glacier using ground-penetrating radar (GPR). Tracer testing through the glacial system via a moulin demonstrated rapid flow of 0.58 m s–1, which was comparable to the flow rates within the proglacial river, and is at the higher end of velocities observed in glacial conduits. A subsequent test from this moulin at the end of the winter season demonstrated that the conduit system was fully open at this time, suggesting that it may not close during the winter. A tracer test through the proglacial foreland showed that the large proglacial lake does not substantially attenuate flow and revealed the presence of a relic conduit system enabling rapid water transit through this area, with velocities of 0.03 m s–1. The proglacial river is highly responsive to melt as a result of fully developed conduits in both the subglacial and proglacial areas. Given that in the proglacial foreland there are the remains of the previous margin of the glacier, these conduits may be a relic of a previous subglacial pathway that has been abandoned in the foreland. The existence of relic conduits may appear in other deglaciating ice-cored forelands and can have a significant impact on the hydrological connectivity and therefore modelling of river discharge in similar catchments.


Mathematical modelling of melt lake formation on an ice shelf

Samantha Buzzard, Daniel Feltham, Daniela Flocco, Peter Sammonds

Corresponding author: Samantha Buzzard

Corresponding author e-mail: s.c.buzzard@pgr.reading.ac.uk

The accumulation of surface meltwater on ice shelves can lead to the formation of melt lakes. These structures have been implicated in crevasse propagation and ice-shelf collapse; the Larsen B ice shelf was observed to have a large amount of melt lakes present on its surface just before its collapse in 2002. Through modelling the transport of heat through the surface of the Larsen C ice shelf, where melt lakes have also been observed, this work aims to provide new insights into the ways in which melt lakes are forming and the effect that meltwater filling crevasses on the ice shelf will have. This will enable an assessment of the role of meltwater in triggering ice-shelf collapse. The Antarctic Peninsula, where Larsen C is situated, has warmed several times the global average over the last century and this ice shelf has been suggested as a candidate for becoming fully saturated with meltwater by the end of the current century. Here we present results of a 1-D mathematical model of heat transfer through an idealized ice shelf. When forced with automatic weather station data from Larsen C, surface melting and the subsequent meltwater accumulation, melt lake development and refreezing are demonstrated through the modelled results. Furthermore, the effect of lateral meltwater transport upon melt lakes and the effect of the lakes upon the surface energy balance are examined. Investigating the role of meltwater in ice-shelf stability is key as collapse can affect ocean circulation and temperature, and cause a loss of habitat. Additionally, it can cause a loss of the buttressing effect that ice shelves can have on their tributary glaciers, thus allowing the glaciers to accelerate, contributing to sea-level rise.


Reconstruction of glacial erosion patterns at the LGM for the Rhine glacier, Swiss Alps, using a high-resolution Stokes flow numerical model

Denis Cohen, Wilfried Haeberli, Urs H. Fischer, Fabien Gillet-Chaulet

Corresponding author: Denis Cohen

Corresponding author e-mail: denis.cohen@gmail.com

Overdeepened valleys in the Alps are common and attest to strong glacial erosion over the last 2 million years. Yet marked overdeepenings are also found near glacier margins in flatter areas of the Alpine foreland, areas occupied by ice for a only a fraction of time during the Quaternary. Furthermore, when ice occupied these areas, cold and dry climatic conditions with low mass-balance gradients, low flow velocities and low basal shear stress prevailed, a set of conditions not conducive for enhanced glacial erosion. To explore basal conditions at the LGM and the potential for glacial erosion, we developed a high-resolution full-Stokes three-dimensional thermomechanical model of the Rhine glacier, Swiss Alps. Our model indicates thawed basal conditions in nearly all areas of the polythermal piedmont lobes, with very low basal shear stress (<30 kPa) and low sliding speeds. Our model neglects permafrost that may have reduced or delayed thawed basal conditions, but we hypothesize that high geothermal heat flux in the Swiss plateau (>100 mW m–2 in some areas) caused by deep hydrothermal recirculation may have kept warm basal conditions, with subglacial water production in piedmont lobes advancing over overdeepened and low-lying areas. Low effective normal pressures associated with high water pressures may have favored processes of quarrying and overdeepening formation. Also warmer conditions after glacial maxima melted large volumes of ice causing increased water fluxes near glacier margins and increasing subglacial erosion, including subglacial fluvial erosion.


Seasonality of glacier–sediment interactions at Taku Glacier, Alaska

Jason Amundson, Roman Motyka, Martin Truffer

Corresponding author: Jason Amundson

Corresponding author e-mail: jason.amundson@uas.alaska.edu

Taku Glacier, Alaska, is in the advance phase of the tidewater glacier cycle. The glacier terminus, which is not currently calving, is buttressed by a large moraine composed of soft glaciomarine sediment that rises above sea level. Previous work demonstrated that the glacier is excavating subglacial sediments at rates of up to 3 m a–1 through fluvial processes, while deforming proglacial sediments at a distance of more than 200 m in front of the advancing terminus. The rapid changes that are occurring provide a unique opportunity to directly observe the coupling between glacier dynamics and geomorphology. Here, we present meteorological, GPS and time-lapse camera data collected during 2014 to show that mechanical reworking of sediment at the terminus is an important excavation process during the annual spring speed-up. In spring, the drainage efficiency is low. Water input into the glacier overwhelms the drainage system, the water pressure rises, and the ice flux to the terminus increases. Consequently, the terminus bulldozes into the proglacial moraine and overrides it while pushing it forward by several meters. In summer, when surface melt rates are high, the drainage system becomes more efficient and the ice flux decreases, causing the terminus to retreat from the moraine. Throughout the remainder of the summer, subglacial discharge is high and sediment excavation is likely dominated by fluvial processes. Thus, our observations demonstrate seasonality in the efficacy of geomorphological processes during tidewater glacier advance: processes at the terminus are most effective in early spring, while excavation farther up-glacier is most effective during times of maximum runoff.


Is subglacial channel spacing dictated by meltwater flux and the hydrogeology of the substrate? Evidence from Canadian eskers

Robert Storrar, Chris Stokes, David Evans

Corresponding author: Robert Storrar

Corresponding author e-mail: r.storrar@qmul.ac.uk

It has been suggested that the lateral spacing of subglacial meltwater channels is governed by a balance between basal melt rate (meltwater flux) and the transmissivity of the bed (its ability to ‘scavenge’ this meltwater). If this relationship is correct, the spacing between eskers should be able to provide important insights into the nature of the channelized subglacial meltwater drainage systems of the ice sheets they formed under. We use a recently published map and GIS database of >20 000 Canadian eskers and compare measurements of esker spacing with variations contained within the geological and hydrogeological record. We find that eskers are quasi-regularly spaced, potentially reflecting a large-scale control on subglacial channel spacing. We further explore the groundwater theory by using a published permeability map to provide a first-order estimate of substrate transmissivity which, when combined with the esker spacing measurements, can be used to ‘retrodict’ basal melt rate. We find that the values of subglacial meltwater flux produced are consistent with independently derived estimates of basal melt rates from a variety of sources, with the majority of estimates fitting within the (very broad) range of 10–5 and 10 m a–1. Large uncertainty terms prohibit further elaboration but the broad agreement of the results appears to support the groundwater-modulated spacing of eskers. However, eskers are frequently associated with supraglacially derived meltwater, as opposed to just meltwater derived from basal melting, and so we suggest that subglacial meltwater discharge values back-calculated using this technique reflect meltwater supply from both supraglacial and subglacial sources. The theory therefore provides a promising method for reconstructing subglacial meltwater fluxes in a high level of detail, and we suggest that future work should focus on: (1) refining our understanding of the hydrogeology of the Canadian Shield; (2) making detailed measurements of esker volume and clast size in order to independently estimate meltwater discharge; and (3) detecting/measuring channel spacing, meltwater supply and substrate lithology at contemporary glaciers and ice sheets.


Groundwater–meltwater coupling in an active proglacial sandur aquifer in southeast Iceland

Brighid Ó Dochartaigh, Alan MacDonald, Paul Wilson, Andrew Black, Jez Everest

Corresponding author: Brighid Ó Dochartaigh

Corresponding author e-mail: beod@bgs.ac.uk

The sandur proglacial to Virkisjokull, a rapidly retreating glacier in southeast Iceland, forms a thick (at least 50–100 m), high permeability, high storage aquifer that is actively coupled with the meltwater river. The sandur aquifer has been characterized by a combination of piezometer drilling, permeability testing and geophysics, and groundwater–meltwater coupling has been investigated by regular monitoring of hydrochemistry, stable isotopes and dissolved gases and by continuous monitoring of groundwater levels and river stage over 2 years. The study has demonstrated that active rainfall recharge occurs across the sparsely vegetated sandur, and there is significant river influence on groundwater within at least 50–150 m of the river channel, particularly proximal to the glacier. In this zone the river loses to the sandur aquifer throughout most of the year, with high meltwater river flows maintaining high groundwater levels in summer. At ~300–400 m distance from the river, there is no significant river influence on groundwater: here groundwater recharge and levels are controlled by the timing and volume of precipitation; groundwater level fluctuations are larger than in the zone near the river; and summer groundwater levels are lower than winter levels. There is significant down-gradient groundwater flow in the sandur aquifer, and from ~3 km down-gradient from the upper edge of the sandur there is extensive groundwater discharge via springs that provides baseflow to the river throughout much of the year. This study highlights the important role that groundwater storage will play in regulating meltwater river flows as glaciers retreat globally.


Multi-decadal glacier mass and area changesin the western Nyainqentanglha range during 1970–2009 derived from multi-sources of DEMs and glacier inventories

Kunpeng Wu, Shiyin Liu, Qinghua Ye, Junfeng Wei, Yongli Zhao, Zhen Zhang, Weijia Bao, Junli Xu, Wanqin Guo

Corresponding author: Kunpeng Wu

Corresponding author e-mail: liusy@lzb.ac.cn

We used multi-sources of digital elevation models (DEMs) and glacier inventories to investigate the three-dimensional changes of glaciers in the western Nyainqentanglha range. Mass change of glaciers in the western Nyainqentanglha range was detected by utilizing geodetic method based on DEMs derived from topographical maps (1970), SRTM (2000) and ICESat GLAS data (2009). Glacier area changes were derived from two glacier inventories with one from topographical maps in 1970 and the other from Landsat TM images acquired in 2009, as well as boundaries extracted from Landsat TM images acquired in 1990 and 2000. Results show that glaciers in the region have experienced shrinkage and mass loss since 1970. Glacier ice volume has decreased by 3.31 ± 0.99 km3 from 1970 to 1999, equivalent to 0.11 ± 0.03 m a–1 water. While glaciers in the region have been in an accelerated mass loss trend as revealed from SRTM and available ICESat GLAS data, glacier ice volume has decreased by ~3.79 ± 2.09 km3 from 1999 to 2008, equivalent to ~0.42 ± 0.25 m a–1 water. This mass loss trend has been displayed in the glacier area changes. Our results show that area reduction of glaciers for the period 1970–2009 is 208.43 ± 66.17 km2, or 24.3 ± 7.7%. An increase speed in area shrinkage happened during 2000 and 2009, the shrinking rates were 4.59 ± 0.96 km2 a–1 during 1970 and 2000 to 5.61 ± 2.13 km2 a–1 during 2000 and 2009. The mass or area changes of glaciers in the study region may be the median level compared with glaciers in other regions of China. The mass losses have an obvious correlation with extent changes of glaciers. The retreat and mass wastage of glaciers may be attributable to climate warming in the region during the last decades.


The role of light-absorbing impurities in accelerating glacier melt in Washington State, USA

Susan Kaspari, McKenzie Skiles, Dan Pittenger, Dan Pittenger

Corresponding author: Susan Kaspari

Corresponding author e-mail: kaspari@geology.cwu.edu

Washington State (USA) has the largest concentration of glaciers in the conterminous United States. These glaciers provide an important source of water resources, particularly in glacier-dominated watersheds where glacier melt can provide up to 50% of the May–September runoff. Recent and projected glacier retreat in this region affects water resource availability. Since 2010 we have investigated the role of light-absorbing impurity (LAI) deposition on accelerated snow and glacier melt in the Cascade and Olympic Mountains. LAI include black carbon (BC) from the incomplete combustion of fossil and bio-fuels, dust, and light-absorbing organic material. For glaciers and snowpack with substantial deposition of light-absorbing impurities, the impurities are more important than temperature in driving melt. We measured snow spectral albedo using a field spectrometer, and analyzed LAI concentrations in snow and ice cores. BC was measured using a Single Particle Soot Photometer (SP2), and dust was characterized using a gravimetric method, inductively coupled mass spectrometry (ICP-SFMS), and particle size. Our results indicate that LAI concentrations are greatest during the dry summer due to increased dry deposition and melt-induced enrichment. A shallow ice core retrieved from Mount Olympus demonstrated that LAI deposition was a magnitude higher than normal during the 2011 Big Hump forest fire, resulting in a two- to four-fold increase in radiative forcing and melt. The timing of the LAI deposition from the forest fire corresponds to an observed increase in the catchment river discharge. A 158 m ice core retrieved from the South Cascade Glacier (a benchmark glacier) in the North Cascades provides a record of 20th century LAI deposition. The BC record indicates low BC concentrations in the early 1900s, elevated BC concentrations mid-century associated with industrialization, a subsequent decrease during the 1960s and 1970s, followed by rising concentrations. The rising BC concentrations during the late 20th century may be due to melt-induced enrichment. Dust concentrations are variable in the ice-core record. We are using the LAI ice-core record combined with glacier mass-balance and river discharge data to assess the role of LAI in glacier melt and retreat.


Some problems of a modern hydrology of glaciers

Bulat Mavlyudov

Corresponding author: Bulat Mavlyudov

Corresponding author e-mail: bulatrm@bk.ru

Modern glacier hydrology knows that water on the surface of glaciers is absorbed by crevasses and moulins and disappears in thickness of ice. Further water somehow moves inside a glacier and flows out at the glacier tongue. Thus the drainage systems located between places of water absorption and discharge can be effective and ineffective. The majority of scientists hold the opinion that water from moulins move at once on a glacier bottom and further moves along it. Confirmation of such a situation is revealed only in the extremely insignificant quantity of cases. Therefore in most cases our knowledge of the position of internal drainage channels inside glaciers is only assumption. What really occurs with water in thickness of glaciers we do not know. This is objective absence of knowledge, which it is impossible to overcome by any methods yet. All available methods do not always allow us to make unequivocal interpretation of the received results. Only the speleological method allows us to penetrate directly in the thickness of glaciers. But such research is individual and has been used only in several regions. Speleological research methods of glacier internal drainage have opened some new problems, which are not clear how to solve yet. One problem is that the results of speleological research do not always correspond with other research (e.g. geophysical) and almost never with the models of internal drainage constructed on the basis of equipotential surfaces. It means that there is no full correctness of the theories taken as a principle of construction of geophysical models and models of drainage on the basis of equipotential surfaces. Another problem is that sometimes there is not quite correct interpretation of data received both by hydrological and speleological methods. Now speleological methods have become more frequent in application. But even they yet do not allow us to explain the general structure of the internal drainage of glaciers, to understand the reasons for its formation and to show how it evolves. Being based on the long-term research of some elements of internal drainage of glaciers located in different regions of the world and on the basis of published data, the author offers a hypothesis on the formation and evolution of the internal drainage of glaciers which is more coordinated with the majority of known data than previous ones.


New type of moulin

Bulat Mavlyudov

Corresponding author: Bulat Mavlyudov

Corresponding author e-mail: bulatrm@bk.ru

It is known, that moulins are formed on crevasses. However our research in past years has shown that moulins can be formed on such areas of glaciers where crevasses are absent. Mainly we saw such moulins on polythermal glaciers of the Spitsbergen archipelago. The reason for their formation was not clear. We find the reason for their formation from research on temperate Bellingshausen Dome (Waterloo – King George Island, Southern Shetland Islands, Antarctica) in 2007–09. It has appeared that moulins are formed on space between planes of walls of the buried inactive glacial canyons when streams of meltwater inflow into them from sideways. In this case real moulin is formed. However the depth of such kinds of moulin cannot exceed the depth of a parent canyon. Usually the depth of such moulins does not exceed 15 m. And only in two cases moulins of this type with a depth of about 30 m have been found (Longyearbreen and Tavlebreen, Nordenskiold Land, Spitsbergen). Apparently it is the maximum depth of such moulins. Thus moulins in glaciers can originate in three ways: (1) on the base of crevasses, (2) on the base of ice–rock contacts on the glacier boundary, and (3) on the base of buried glacial canyons when stream of meltwater inflow in it from sideways.


A hypothetical model for the deglacial hydrology of ice sheets in mountainous areas

Trevor Faulkner

Corresponding author: Trevor Faulkner

Corresponding author e-mail: trevor@marblecaves.org.uk

Few studies have attempted to describe a general model of the deglacial hydrology of ice sheets in mountainous areas. This hypothetical model is based on a study of the Late Weichselian deglaciation of central Scandinavia, using parabolic formulae reconstructed from those proposed by Arne Grønlie in 1975. It is assumed that deglaciation was mainly driven by two summer heat fluxes. One was warming by sea water that caused the ice-sheet margin to ‘backwaste’ eastwards as the melting sea encroached up fjords and coastal valleys to the marine limit. This was under the influence of competing sea-level rise and isostatic uplift, which also caused seismic and aseismic neotectonic movements. The second was direct solar warming that caused the ice sheet to melt and sublimate from its upper surface as it ‘downwasted’. When mountain peaks and ridges emerged as nunataks above the ice sheet, static ice-dammed lakes (IDLs) were initially formed alongside them as the rocks warmed each summer. Six types of narrow and wide IDLs are identified that grew in size and reached down to bedrock, perhaps forming subglacial reservoirs. They inundated most parts of the landscape for periods up to ~1000 years, as they lowered into the valleys. Each part of the topography is assigned to one of ten ‘glacial situations’, to which the IDLs are related. IDLs and other bodies of water became part of meltwater flow regimes, of which 13 types are described. These had highly variable flow velocities along englacial conduits and subglacial waterways that formed Nye and Röthlisberger channels towards short or long tidewater glaciers, as deglaciation proceeded. They applied variously in ice that was ‘warm’ or ‘cold’, and above and below the plastic behaviour limit. Above this limit, ice behaves as a solid rock with fractures and crevasses, whereas below it, englacial conduits do not survive. It is anticipated that such a model should also apply to the deglaciation of the west coast of Scotland and the east coast of North America.


The Weichselian deglaciation of central Scandinavia

Trevor Faulkner

Corresponding author: Trevor Faulkner

Corresponding author e-mail: trevor@marblecaves.org.uk

‘Central Scandinavia’ extends from the Norwegian coast to the Caledonide thrust front in Sweden. It is assumed that the Weichselian deglaciation was driven by two summer heat fluxes: warming by the sea as it encroached up fjords and caused the ice-sheet margin to retreat eastwards (accompanied by deglacial seismicity caused by local isostasy); and warming by direct solar insolation that caused the ice sheet to ablate from its upper surface. An earlier empirical parabolic time relationship of H = 0.75t2 for melting height has been reconstructed as H = 1700 + 5(YD isobase–220)–0.75 × 10–4 × (13500 t)2 m, where t = 14Ca BP, using information about the isostatic uplift that increased inland. Deglaciation was modelled by drawing deglacial maps at ~300 year intervals, after the ice began to melt vigorously in the Bølling interstadial below the present 1700 m altitude (at the 220 m YD isobase at Børgefjell) that was already ice-free. The maps show the advance of the sea, the recession of the ice-sheet margin, the synchronous thinning of the ice sheet, and the evolution of ice-dammed lakes (IDLs), After ~11 ka BP, the melting height lowered at roughly 0.5 m a–1 at all isobases. The formula agrees with many marine dates to within 300 years in the central and northern part of the area, but downwasting by calving was much faster along the coast, as the sea submerged large areas of the isostatically depressed strandflat. Across the Swedish border, there is evidence that large IDLs and glaciers remained longer than predicted. Most of the area was probably systematically flooded by glacial meltwater for periods of 800–1200 14Ca during deglaciation as local IDLs lowered and coalesced (commonly at jökulhlaups). Six types of deglacial IDL are described, with hydrological flows initially into englacial conduits and subglacial reservoirs, and later into subglacial waterways that ran to the sea. The ice margin lost its sharp, morainal definition as it retreated under the increasing influence of topography. The IDLs reduced in size as the ice sheet shrank into separate valley glaciers, which, in Norway, eventually became tidewater glaciers at the deglaciation marine limit.


Direct monitoring of the subglacial system: how does it differ from borehole measurements?

Pierre-Marie Lefeuvre, Thomas Zwinger, Alexandra Messerli, Miriam Jackson, Gaute Lappegard, Jon Ove Hagen

Corresponding author: Pierre-Marie Lefeuvre

Corresponding author e-mail: p.m.b.e.lefeuvre@geo.uio.no

Measurements of pressure at the glacier bed are made from sensors installed either directly at the glacier bed (load cells) or in boreholes. Results from both kinds of instrumentation show a seasonal evolution of the subglacial drainage system. However, the characteristics of the response of each to surface melting and rainfall are significantly different. This study summarizes results from time series analysis of basal pressure and modelling work conducted at the Svartisen Subglacial Laboratory in order to compare the outcome with the understanding of the subglacial system from borehole campaigns. Installed under 200 m of glacier ice, the load cell network shows periods with spatial anti-correlation in summer, which is caused by stress redistribution between the efficient drainage system and isolated parts of the glacier bed. A full stress ice flow model reveals the importance of this mechanical stress transfer around subglacial channels that is responsible for causing daily pressure events at the glacier bed during stable weather. A characteristic event recorded at a load cell is a sudden drop-and-peak in pressure, usually corresponding with the peak in subglacial water pressure. These typical events are not observed in borehole measurements, they instead measure sinusoidal fluctuations in water pressure. Different response characteristics are also observed during spring speed-up events. During these events, pressure sensors in boreholes generally become connected to an efficient drainage system and then exhibit daily pressure fluctuations for the rest of the melt season. Instead of presenting an increase in connection and activity, the load cells show a damped response after the subglacial drainage system has accommodated the first input of meltwater. The observations from load cells and boreholes are not contradictory, instead they complement each other and help characterize the mechanical and hydrological dynamics occurring at the glacier bed.


In situ constraints on the relationship between basal water flux, water pressure and ice velocity from borehole experiments in western Greenland

Toby Meierbachtol, Joel Harper, Neil Humphrey

Corresponding author: Toby Meierbachtol

Corresponding author e-mail: toby.meierbachtol@umontana.edu

Subglacial water pressure is commonly invoked as the primary link between input of meltwater to the bed and ice sliding speed over daily to seasonal timescales. The complexity of the subglacial system, however, is known to often preclude local water flux as the only factor influencing basal pressure. Further, interpreting observations of water pressure in terms of flow at the bed is complicated by the difficulty in measuring water flux. Here we directly constrain basal water flux using in situ dye-tracing techniques in two boreholes having concurrent measurements of basal water pressure and located in 670 m thick ice in western Greenland. The ~2 week time series indicate that diffusional processes primarily dominate dye concentration changes at the bed, and stagnant conditions generally persist in the local subglacial system to which the boreholes are connected. Despite the lack of water flow, basal pressure measurements in the same boreholes vary by up to tens of meters. Diurnal pressure variations are apparent, and generally out of phase with surface velocity measurements. With water flow constrained at the bottom of the boreholes, we hypothesize that mechanical processes exert the dominant influence on measured water pressure. Pressure variations of the observed magnitude may be induced by very small volume changes in such an isolated, coupled cavity-borehole system.


Liquid water storage in snow and ice of 80 eastern Alpine basins

Michael Kuhn, Kay Helfricht, Martin Ortner

Corresponding author: Michael Kuhn

Corresponding author e-mail: michael.kuhn@uibk.ac.at

The retention and release of liquid water in glacierized basins was modelled with OEZ, a conceptual, semi-distributed model of the water and ice balance designed for long-term averages with monthly resolution for 100 m elevation bands. Here we present the components of the liquid water balance of 80 mostly glacierized basins on either side of the Main Alpine Divide between 10 and 13°E in the period 1998–2006 and compare them with the records of 30 basins monitored from 1969 to 1997. Records of daily runoff, long-term changes of glacier volume in three glacier inventories and parameterized values of evaporation yield basin precipitation as a residual. They are combined with records of temperature and precipitation T and P, Histalp values of dT/dz, and values of dP/dz and degree-day factors constrained to close the water balance, yielding monthly values of rain, meltwater, evaporation and runoff for a liquid water balance. Liquid water is stored in the mature snowpack, in glaciers, in soil and in the reservoirs of power plants. Meltwater retention in snow has maxima in excess of 100 mm month–1 in May, often followed by maximum release when the retaining snow matrix melts. Glacier storage peaks in August due to ice melt and the ensuing filling of the englacial reservoirs. These two components combined to a common maximum of storage in summer in the first period and developed two distinct maxima in the warmer period 1998–2006. A further maximum of liquid water storage that was often found in October is most likely due to a peak in precipitation in the southern part of that region. The fraction of liquid water stored in soil is uncertain – a minimum value is suggested by the base flow of 10–30 mm month–1 from November to March. Examples are given for the storage in and release from power plant reservoirs.


Controls on the initiation of hydrofracture beneath supraglacial lakes on the Greenland ice sheet

Andrew Williamson, Alison Banwell, Neil Arnold, Ian Willis

Corresponding author: Andrew Williamson

Corresponding author e-mail: agw41@cam.ac.uk

Supraglacial lakes form annually within the ablation zone of the Greenland ice sheet (GrIS). Many of these lakes drain in the mid- to late melt season, with drainage classified into fast (<24 hour) or slow (>1 day). Fast drainage occurs by hydrofracture, delivering large meltwater pulses to the bed, subsequently affecting subglacial drainage system development, basal water pressures and ice-sheet dynamics. Thus, studying rapidly draining lakes is crucial for the development of coupled ice-sheet hydrology and ice-dynamics models, which will be critical for quantifying the future mass loss from the GrIS. Although much is known about the GrIS’s hydrology, a fundamental unknown surrounds the precise factor(s) initiating hydrofracture beneath lakes. Several studies suggest that rapid drainage occurs at a critical water-volume threshold, which varies with ice thickness, with larger volumes necessary for hydrofracture in thicker ice. Following observations of clusters of rapid lake-drainage events, other studies suggest that synoptic trigger mechanisms may be responsible: e.g. when one lake drains, it may alter local ice velocity and basal stress fields, initiating the drainage of nearby lakes. A combination of the ice-volume threshold and synoptic factors may be responsible for rapid drainage, but this requires verification. Here we conduct a high-resolution study on the land-terminating Paakitsoq region of the GrIS to investigate a range of specific factors that may control rapid lake drainage. We analyse Moderate Resolution Imaging Spectroradiometer (MODIS) imagery with lake area, depth and volume algorithms, and determine the drainage dates of lakes. For each lake, the water volumes immediately prior to drainage are compared with other local data, notably ice thickness, slope angles, surface velocities and strain rates, and lake-filling rates to elicit the statistically significant controls for initiating hydrofracture beneath lakes. Ultimately, we aim to use this methodology to investigate how the importance of these different factors varies over the GrIS both temporally and spatially (e.g. for land- and marine-terminating catchments, for areas with compressive and extensive strain regimes, and for the lower and upper ablation areas). This knowledge will be used to improve the physical basis of our existing lake-filling and fracture model, generating improved estimates of water delivery to the base of the GrIS.


A model of drumlin growth at Múlajökull, Iceland, based on contrasting subglacial hydrology in its quiescent and surging states

Neal Iverson, Rebecca McCracken, Lucas Zoet, Anders Schomacker, Ivar Benediktsson, Mark Johnson, Thomas Hooyer

Corresponding author: Neal Iverson

Corresponding author e-mail: niverson@iastate.edu

Drumlins exposed by the recession of Múlajökull in central Iceland provide an unusual opportunity to study their formation in the forefield of a modern surge-type glacier. Observations there indicate the following: (1) drumlin relief is a consequence of till deposition on drumlins and erosion between them; (2) till deposition occurred during surges; (3) longitudinally compressive strain in till of the drumlins, such as that associated with negative flux divergence in a deformable bed, is absent; (4) till was under higher effective normal stresses between drumlins than within them; (5) crevasses on the glacier surface after surging are coincident with drumlins; and (6) water drainage beneath the glacier margin during the current quiescent state is through channels in low areas between drumlins. On the basis of these observations, we present a model of drumlin growth guided by the expectation that subglacial sediment transport will be closely tied to subglacial hydrology through its control on patterns of effective normal stress on the bed. During periods of slow flow between surges, those patterns are set by diffusive flow of water through and along the bed toward R-channels and by the distribution of crevasses on the glacier surface, which decreases the total normal stress on the bed above drumlins. Basal melting is controlled by fluxes of frictional and geothermal heat, with the distribution of the latter affected by the variation in ice temperature associated with bed topography and basal motion. A model of regelation infiltration then predicts significant entrainment of sediment (decimeters to a few meters) in ice near R-channels, where effective stresses on the bed are highest. In contrast, during surges, high water pressure and distributed subglacial drainage are expected to prevent the intimate contact between ice and the bed that regelation infiltration requires. High frictional heat fluxes are, in that case, sufficient to deposit a layer of till on the bed of a few decimeters to a few meters in thickness. Thus, in the model, differential entrainment of sediment in ice between surges and deposition from it during surges cause drumlins to grow, over length scales set by the spacing of R-channels. Sediment transport by bed deformation and channelized water could be included in the model, at the expense of adding some poorly constrained parameters.


Glacier motion in relation to jökulhlaups from western Vatnajökull in Iceland observed with continuous GPS measurements

Bergur Einarsson, Eyjólfur Magnússon, Matthew J. Roberts, Tómas Jóhannesson, Finnur Pálsson, Thorsteinn Thorsteinsson

Corresponding author: Bergur Einarsson

Corresponding author e-mail: bergur@vedur.is

Jökulhlaups (glacier outburst floods) from subglacial geothermal areas, marginal lakes and subglacial volcanic eruptions are common in Iceland. These floods overfill the subglacial hydrological system and have a substantial effect on the dynamics of the glacier. The effect of jökulhlaups on glacier dynamics is controlled by how the floodwater propagates at the glacier bed and properties of the subglacial flood path. Many aspects of the propagation mechanism of jökulhlaups at the glacier bed are still uncertain. Campaigns with continuous GPS instruments on the glacier surface can give valuable information on how jökulhlaups affect glacier motion and shed light on the propagation mechanism. Three such campaigns have been carried out on western Vatnajökull, SE Iceland, over known jökulhlaup flood paths from the subglacial lake Grímsvötn and the subglacial lakes under the Skaftá cauldrons. Four jökulhlaups have been captured in these campaigns. Two were slowly rising jökulhlaups from Grímsvötn with maximum discharge of 4000 m3 s–1 and 3000 m3 s–1 in 2004 and 2010. The other two took place in 2008 and drained from the western and the eastern Skaftá cauldrons. They were both rapidly rising and with maximum discharge of 240 m3 s–1 and 1290 m3 s–1, respectively. Glacier surface movements measured in these campaigns as well as discharge curves for the corresponding jökulhlaups are presented and interpreted. Features identified in these measurements are interpreted as indicating (1) propagation of rapidly rising jökulhlaups with a subglacial pressure wave, (2) decreased glacier basal friction during jökulhlaups, (3) subglacial accumulation of water in slowly rising jökulhlaups, and (4) lifting of the glacier caused by subglacial water pressure exceeding overburden both in rapidly and slowly rising jökulhlaups. Different propagation mechanisms of a rapidly rising jökulhlaup, depending on the development of the subglacial hydrological system, are also inferred. Some of these observations, such as subglacial water accumulation and glacier lifting caused by subglacial water pressure exceeding overburden in slowly rising jökulhlaups, are inconsistent with assumptions that are typically made in theoretical and numerical modelling of jökulhlaups.


The importance of extreme events in release of suspended yield from Werenskioldbreen catchment (Spitsbergen)

Elżbieta Majchrowska

Corresponding author: Elżbieta Majchrowska

Corresponding author e-mail: elzbieta.majchrowska@us.edu.pl

Sediment yield (mm a–1) can be used as an indicator of hydrological response to climate change. Hodson and others (1998) suggested that glaciofluvial sediments dominate in marine sedimentation in high latitude and those occurring over the last 10 k years (Syvetski and others, 1986). Therefore, positive trend in suspended sediment delivery from glacierized area is expected, but Bogen and Bønsnes (2003), who looked at a long series of measurements for different glaciers, indicate wide variation in the supply of sediment yield and show that it is controlled by bedrock susceptibility to erosion and by glaciological parameters as a second factor, especially in the case of the glaciers of the polythermal or temperate structure. These authors also emphasize the importance of large floods in release of suspended materials. In view of the research undertaken in the High Arctic, I present a study of transport dynamics of fluvial suspended sediment matters in glacierized catchment in southern Spitsbergen (Werenskioldbreen). Hydrometerological conditions and sediment transport were measured over six seasons (2007–2012) in order to characterize the seasonal course of suspended sediment concentrations (SSC), the magnitude of suspended sediment loads (SSL) and mechanical denudation rate. Fine materials in a glacier catchment have different origins, e.g. materials from mass movement processes on slopes and moraines, which are delivered mostly as a result of snowmelt or intensive precipitation. Furthermore, sediment materials come from the proglacial sandur with increasing active layer depth, which are given potentially more materials to elute. One of the most important sources for polythermal glacier as Werenskioldbreen is material from the glacier bed. In this context, I determine the role of the subglacial sediment sources in this glacier and examine how extreme events such as heavy precipitation (especially in the second part of the ablation season, in the autumn) modify sediment transport in longer time series.


Parameter estimation with a lumped model of subglacial hydrology

Douglas Brinkerhoff, Colin Meyer, Ed Bueler

Corresponding author: Douglas Brinkerhoff

Corresponding author e-mail: dbrinkerhoff@alaska.edu

We formulate a model of the subglacial hydrologic system as a pair of coupled equations representing the evolution of englacial and subglacial storage through their respective proxies of basal water pressure and average subglacial conduit size. We non-dimensionalize this model and find that its behavior is governed by seven parameters. Three of these parameterize the Darcy–Weisbach flux relation, while the other four represent geometric factors relating subglacial to englacial reservoir size, englacial storage to water pressure, the rate of turbulent channel opening, and the dependence of sliding velocity on water pressure. We apply the model to Kennicott Glacier, Wrangell Mountains, AK, using measurements of area-averaged input flux, terminus outflow, and surface velocities to invert for the values of model parameters. This simple model is able to reproduce observations, including the amplitude and frequency of evident diurnal cycles in velocity and discharge. We find that for Kennicott Glacier, observations are best described when opening of channels due to basal sliding is much larger than that due to turbulent melt. Also, at pressures less than overburden, englacial storage accounts for most of the total storage of water. Finally, the model predicts non-negligible timescales for pressure changes, implying that the value of englacial porosity contributes substantively to the dynamics of the subglacial system and to those of the overlying glacier itself.


Spatial extent and temporal variability of the Greenland firn aquifer detected by ground and airborne radars

Clément Miège, Richard Forster, Ludovic Brucker, Lora Koenig, Kip Solomon, John Paden, Sivaprasad Gogineni

Corresponding author: Clément Miège

Corresponding author e-mail: clement.miege@gmail.com

Widespread perennial firn aquifers are found in an elevation range of ~1200–2000 m within the lower percolation zone of the Greenland ice sheet. Aquifers are predominantly found in the high snow-accumulation regions of southeast and south Greenland. Prior to melt onset, we map the aquifer extent and quantify the water table interannual variations using surface-based and airborne radars. The aquifers are mapped along a linear distance of 2420 km over three NASA Operation IceBridge (OIB) campaigns (2011–13) and the average depth to the water table is 22 ± 8 m. At our southeast Greenland field site, we found that the water table was spatially stable between 2011 and 2014, fluctuating less than 2.5 m vertically. For the upper part of Helheim Glacier, a spatially stable aquifer is inferred from missing bed echoes from 1993 to 2008. After 2008, this aquifer expands to higher elevations and after spring 2012 drainage of its lower elevation portion is suspected. Combining radar data with surface elevation profiles, we find that in general firn aquifers are initiated by strong surface slope gradients. The water flows laterally in an unconfined aquifer, topographically driven by ice-sheet surface undulations until water encounters crevasses. Our hydraulic simulations suggest local flow cells are formed within the aquifer and some water is discharged at the steep-to-flat transitions. The firn is saturated with water at the firn–ice transition, which delivers heat downward. Water drainage into crevasses is suspected from repeated radar profiles and visible imagery, but the volume and volumetric flow rate of water are still unknown.


Potential effects of dust deposition on snow and glaciers in the Andes of central Chile

Sophie Biskop, Guillermo Azócar, Alexander Brenning

Corresponding author: Sophie Biskop

Corresponding author e-mail: sophie.biskop@uni-jena.de

The deposition of particulate matter emitted into the air/atmosphere by natural and anthropogenic sources has the potential of increasing melt rates of snow and ice. These possible effects are of particular concern in dry regions such as the Andes of central Chile, which are affected by mining and other industrial activities. In Chile, open-pit mines often tend to be located next to glaciers that can be an important water resource for municipal supplies and irrigated agriculture, in particular during dry seasons. In the context of mine expansion projects, the objective of this study is to assess potential changes in glacier melt rates that can be attributed to the effect of mining-generated dust on glaciers. Therefore, an enhanced temperature-index glacier melt model is applied which includes albedo parameterization that ‘captures’ the influence of particulate matter. Spatially distributed output data of existing dust deposition models are used as input for the estimation of snow and ice albedo. Based upon model-simulated dust deposition rates, expected best- and worst-case scenarios were evaluated, which attempt to represent, as a rough approximation, climatic variability, uncontrolled emissions and multi-annual dust accumulation. The estimated possible changes of annual losses of snow and glacier ice are primarily due to the extended ice melt period as an indirect effect, and to a smaller extent due to increased snowmelt in the accumulation areas. Recommendations for future work include mainly the in situ monitoring of meteorological, glaciological and particulate variables that are relevant for the snow and ice melt model. In addition, field experiments are very useful to empirically determine the possible impacts of particulate matter with different characteristics. These additional data would allow better model calibration and validation.


Subglacial hydrology has a control on ice-stream margins: the role of pore pressure distribution and basal strenghtening

Thibaut Perol, James R. Rice, John D. Platt, Jenny Suckale

Corresponding author: Thibaut Perol

Corresponding author e-mail: tperol@seas.harvard.edu

Ice streams are drainage routes that control the discharge rate of the West Antarctica ice sheet. The fast-flowing ice streams are separated from the nearly stagnant ice in the adjacent ridge by zones of highly localized deformation known as shear margins. It is still unclear what mechanisms select the location of shear margins, or how this relates to observations of margin migration. We study the structure of a shear margin using a two-dimensional thermomechanical model in a cross-section perpendicular to the direction of downstream flow. We show that intense straining at the margins can partially melt the ice, and introduce a drainage channel at the base of the shear margin that collects the melt produced. As shown in Röthlisberger (1972), the channel locally decreases the pore pressure in the subglacial till, leading to an effective normal stress just outside the channel that can be up to a hundred times higher than that inferred under the majority of the ice stream. We couple the ice flow with a subglacial hydrology model for the basal shear strength of the slipping bed assuming that the till has a Coulomb-plastic rheology, showing that the additional basal resistance produced by the channel water suction lowers the stress concentrated on the locked portion of the bed adjacent to the stream. For a given location of the basal slipping-to-locked transition we constraint the transmissivity of the subglacial hydrology system required to have a stable margin – where the basal shear stress resolved on the bed is always less than the yield strength of the till when the bed is locked. Matching the ice deformation data at Dragon margin of Whillans Ice Stream B2 (Echelmeyer and others, 1994) we find that margins are stable when the transition in basal slip conditions occurs less than 500 m away from a channel operating at 200 kPa and for a hydraulic transmissivity consistent with a thin water-film at the bed of 0.2 mm thickness.


Modelling heterogeneous meltwater percolation on the Greenland ice sheet

Stefan Ligtenberg, Michiel van den Broeke

Corresponding author: Stefan Ligtenberg

Corresponding author e-mail: s.r.m.ligtenberg@uu.nl

Surface melt on the Greenland ice sheet (GrIS) has increased over the last decades, with the latest record set in the summer of 2012. To assess current and future ice-sheet mass balance, it is important to quantify what part of this meltwater reaches the ocean and contributes to sea-level change. Meltwater produced at the surface has several pathways to the open ocean. It can infiltrate the local firn pack, where it is either stored temporarily or refrozen, or it can run off along the surface or via englacial drainage systems. In this study, we focus on the first; more specifically, in which manner meltwater percolates the firn column. Over the past years, GrIS research has shown that meltwater does not infiltrate the firn pack homogeneously (i.e. matrix flow), but that inhomogeneities in horizontal firn layers causes preferential flow paths for meltwater (i.e. piping). Although this process has been observed and studied at a few isolated sites, it has never been examined on the entire GrIS. To do so, we use the firn model IMAU-FDM with new parameterizations for preferential flow, impermeable ice lenses and subsurface runoff. At the surface, IMAU-FDM is forced with realistic climate data from the regional climate model RACMO2.3. The model results are evaluated with temperatures and density measurements from firn cores across the GrIS. By allowing for heterogeneous meltwater percolation, the model is able to store heat and mass much deeper in the firn column. This is, however, in part counteracted by the inclusion of impermeability of ice lenses, which causes part of the meltwater to run off horizontally.


Hydraulic controls on the thermal surge mechanism? Direct velocity measurements from Basin-3, Austfonna, Svalbard, during surge initiation

Thorben Dunse, Thomas V. Schuler, Thomas Schellenberger, Andreas M. Kääb, Jon Ove Hagen, Carleen H. Tijm-Reijmer

Corresponding author: Thorben Dunse

Corresponding author e-mail: thorben.dunse@geo.uio.no

Glacier surges are commonly explained by either thermal or hydraulic mechanisms. In the first case, changes in the subglacial temperature conditions inhibit or enable basal motion of cold or polythermal glaciers. Alternatively, surge behavior of temperate glaciers has been attributed to hydraulic lubrication, where predominance of inefficient subglacial drainage involves high basal water pressure and thus causes enhanced basal motion. The polythermal glaciers on Svalbard are believed to belong to the first group and a posteriori observations from winter satellite images suggest a gradually increasing acceleration over several years, attributed to an increasing proportion of the glacier bed that becomes temperate. Since 2004, a wealth of ground data have been collected on Austfonna Ice Cap, Svalbard. Basin-3 is a marine-terminating drainage basin that has surged in the past. We present 5 year, continuous GPS records (2008–13) from Basin-3, which display hitherto unknown detail of the surge initiation. Winter velocities largely follow the known pattern of increasing acceleration, but the time series reveal a step-wise rather than a continuous acceleration. The bulk of the acceleration takes place in a short period during summer and is highly correlated to the surface meltwater production. These observations are difficult to explain using common surge theory and hence suggest the existence of a so far overlooked process. The close relation between glacier acceleration and surface melt rates suggests a hydraulic control on basal thermal conditions. Here we propose an extension to the cryo-hydrologic warming mechanism where we focus on the role of meltwater in warming the glacier bed, in addition to the significance of meltwater for ice rheology. The still ongoing (February 2015) surge of Basin-3 caused drastically enhanced ice discharge from Austfonna, and estimated frontal ablation rates from this surge alone are comparable with those of entire Svalbard.


Superimposed ice at perennial snowfields and glacierets – a modelling approach

Rune Strand Ødegård

Corresponding author: Rune Strand Ødegård

Corresponding author e-mail: rune.oedegaard@hig.no

Numerous archaeological finds from perennial snowfields and glacierets have uncovered an urgent need for a better understanding of the physical characteristics of these features and how they link to the Holocene climate history and glacier mass balance. There are about 3000 known artefact finds globally at perennial snowfields and glacierets, ~2000 of these finds are in central southern Norway. Archaeologists have so far identified more than 65 sites with finds in this area, but many sites with potential finds have not been checked. From a natural/cultural management perspective, there is particular interest in the sensitivity of perennial snowfields to future climate change. The finds are mostly related to reindeer hunting, the oldest finds so far date back to the Bronze Age. Radiocarbon dating suggests that small perennial snowfields have existed continuously since the Holocene thermal maximum (HTM) without disappearing or developing into a glacier with basal sliding. There are indications of internal deformation. Results from a case study at a perennial snowfield called Juvfonne (more than 600 registered wooden artefacts) show a cold-based ice underlain by permafrost. Juvfonne is presently in the superimposed ice zone following the terminology for glaciers. The temperature measurements at 5–10 m depth in the ice are similar to what is measured in a nearby permafrost borehole at Juvvasshøe. The surface meltwater does not percolate through the level of the winter cold wave. The heat flow into the ice is gradually decreasing during the melt season. Superimposed ice is formed at the level of impermeable ice. Cold ice combined with moderate deformation is obviously a prerequisite for the preservation of wooden/leather artefacts for thousands of years. A modelling experiment was set up to obtain a better understanding of how the mass balance and thermal regime relates to long-term changes in surface melt, with particular focus on superimposed ice. The model follows previous 1-D modelling of superimposed ice. The only required forcing data are hourly air temperatures. Snow accumulation rates are simulated in the model with realistic interannual variability. The fact that these features have existed for thousands of years with surface altitude changes within a few tens of meters gives interesting constraints to the long-term mass-balance modelling. One interesting outcome of the model is the probability of subaerial exposure of artefacts.


Spatial and temporal evolution of basal friction coefficients inferred from surface horizontal velocities and comparison with surface water production in Kronebreen, Svalbard

Dorothée Vallot, Rickard Pettersson, Adrian Luckman, Martina Schäfer, Thomas Zwinger, Jack Kohler, Ward Van Pelt, Björn Claremar, Doug Benn

Corresponding author: Dorothée Vallot

Corresponding author e-mail: dorothee.vallot@geo.uu.se

Kronebreen is a grounded tidewater glacier situated in northwest Svalbard that has been recorded as one of the fastest glaciers in the archipelago. A large accumulation area drained into a narrow channel associated with sliding at the base are believed to play a role in controlling these high velocities. Recent observations have shown a sudden thinning and retreat of the front, which indicate a complex dynamical behavior that is important to understand and be able to estimate the glacier’s contribution to sea-level change. In order to understand the evolution of basal sliding through time, we infer the spatially and temporally varying basal friction coefficients by inversing observed surface ice velocities using the open-source finite-element model Elmer/ICE developed at the CSC, Finland. The model is constrained by the digital elevation models of surface and bedrock topography and observed surface velocities (TerraSAR-X) for 11 day intervals over 2 years. We then analyze the spatial and temporal evolution of these coefficients and relate it to other parameters influencing motion such as surface water.


The role of water in actively forming subglacial landforms using geophysical data from Rutford Ice Stream, Antarctica

Edward King, Andrew Smith

Corresponding author: Edward King

Corresponding author e-mail: ecki@bas.ac.uk

Rutford Ice Stream has many elongate landforms at the bed, and erosion and deposition on decadal timescales has been observed. Geophysics also suggests that the sediments are soft, high porosity (up to 60%) tills. Similar porosities have been found in soft till exposed on the sea bed in the NE Antarctic Peninsula and beneath Bindschadler Ice Stream. Both seismic and laboratory measurements indicate very low shear strength. Seismic monitoring indicates movement of ice over soft till generates no micro earthquakes, implying that the flow of the ice stream is taken up by deformation within the soft till. Soft-sediment deformation is common in soft tills, and micromorphological analysis shows that high water content and high pore-water pressures can lead to varying degrees of liquifaction, hydrofracturing and homogenization of basal sediments. Hence we interpret the sediments comprising the drumlins and MSGL beneath Rutford Ice Stream as weak till with very high water content, very low effective pressure and undergoing active shear and deformation. The role of water within the sediment is likely to be significant and the till may be close to liquifaction in places. Little is known about the movement of water through subglacial landforms or the distribution of pressure within them, e.g from flanks to crest. Hydraulic conductivity from a static laboratory test on samples from Kamb Ice Stream was very low, but the experiment did not reproduce subglacial churning of sediment, a process that may allow water to move more freely through the till. Also, the clay content of the Kamb Ice Stream till was particularly high (35%) and there are no known MSGL or drumlins in that region. Where MSGLs have been directly sampled, clay content is generally much lower and the diamict forming the landforms has a sandy/silty matrix, likely to enable higher levels of hydraulic continuity. We postulate that the Rutford till has a relatively high hydraulic continuity due to the sediment composition and the dynamic deformation of the very weak material. We further postulate that these properties allow the effective pressure to vary through the body of the landforms. Here we take the dimensions of a drumlin beneath Rutford and determine what the distribution of water pressure within the feature might be under different assumptions of hydraulic continuity, pressure transfer and the degree to which the system is sealed or open. We then consider the impact on theories of drumlin/MSGL formation.


Discharge and hydrochemistry of an outlet of the Greenland ice sheet

Joseph Graly, Joel Harrington, Neil Humphrey

Corresponding author: Joseph Graly

Corresponding author e-mail: jgraly@uwyo.edu

In order to assess subglacial processes of the Greenland ice sheet, outlet stream discharge variations and major element hydrochemistry where measured at a large, land-terminating outlet in western Greenland. Measurements occurred over the course of several days in July 2013. The outlet is at the terminus of Isunnguata Sermia, ~25 km northeast of Kangerlussuaq. Discharge was assessed from hourly photography of the outlet from multiple angles, including where a mid-stream naled provided a natural gauge. pH and alkalinity were measured in samples of stream water collected every 3 hours. Major cation, anion and dissolved Si concentrations were measured in the same samples in a laboratory setting. Discharge in the outlet peaks between 19:00 and 21:00 hours and is lowest between 7:00 and 9:00 hours. In interior surface streams, discharge typically peaks at 15:00 hours and is lowest at 4:00 or 5:00 hours. This implies a 3–6 hour offset between the melt of the interior surface and discharge at the terminal outlet. This offset is longer for the peak discharge in the afternoon than for the minimum discharge in the morning. The concentration of dissolved solutes in the stream shows a complex pattern of a diurnal cycle with a spike during the stream’s waning flow (around 2:00 hours). The magnitude of the spike varies between 20 and 60% of the average of the two adjacent samples. This variability suggests that the 3 hour sampling schedule may have not been frequent enough to fully characterize the spike. Diurnal minima range from 242 to 318 μM TDS. Diurnal maxima range from 310 to 440 μM TDS, implying 30–40% changes over the course of a day. The maximum is found at the 11:00 hour sample and the minimum at the 23:00 hour sample. This is 3–5 hours offset from peak and minimum flow. The discharge and chemical data suggest that the rate of chemical weathering under the ice is influenced by the diurnal flux of surface meltwater and that the dissolved solutes are more than simply diluted by the changing flux of water. The delay between the lowest flow of water and the peak of dissolved solute concentration suggests that total chemical weathering is increasing along with increased flow, though not to an extent that totally negates a dilution effect. The expansion of the subglacial network into seldom accessed regions during high flow could account for this phenomenon. The spike of partial silicate reaction products during waning flow suggests a flushing effect.


Using stable isotopes and continuous meltwater river monitoring to investigate the hydrology of a rapidly retreating Icelandic outlet glacier

Alan MacDonald, Andrew Black, Brighid Ó Dochartaigh, Verity Flett, George Darling, Jez Everest

Corresponding author: Alan MacDonald

Corresponding author e-mail: amm@bgs.ac.uk

Virkisjökull is an outlet glacier draining the western flanks of Öræfajökull on the southern extremity of the Vatnajökull ice cap in SE Iceland. The glacier is highly sensitive to environmental change and has been shown to be in rapid retreat since 2005. Since 2011 there has been continuous measurements of flow in the proglacial meltwater channel, daily photographs of the snout and regular campaigns to sample water chemistry and stable isotopes from the meltwater, ice and groundwater. Ice- and snowmelt dominate the summer flows, but significant flow is also observed in winter due to periodic snowmelt and winter storms. The data indicate a highly efficient glacial drainage system, which makes use of a series of permanent englacial channels within active and buried ice throughout the year. Groundwater contributions to river flow can also be distinguished using stable isotopes and forms an important source of baseflow to the river at low flows. The efficient and year round drainage network may be a contributory factor to the grounding and subsequent rapid retreat of the glacier.


Applying GlaDS hydrology model to analysis of sub-Antarctic lakes

Christine Dow, Mauro Werder, Sophie Nowicki, Ryan Walker, Greg Babonis, Bea Csatho, Mathieu Morlighem

Corresponding author: Christine Dow

Corresponding author e-mail: christine.f.dow@nasa.gov

The presence of water at the bed of the Antarctic ice sheets is known to be a first-order control on ice dynamics. In many regions, distribution and flux of this water is complicated by multi-year storage in subglacial lake basins. With more lakes and pockets of stored water being identified every year from surface altimetry measurements and radio-echo sounding it is apparent that constraining the impact of this water is an important step for determining the drivers of Antarctic ice dynamics. We examine the role of sub-Antarctic lakes in subglacial hydrological development and ice dynamics. In particular we focus on hydrological systems downstream of draining lakes in addition to the controls on lake drainage timing and stability. Our method utilizes GlaDS, a 2-D subglacial hydrology model, which incorporates development of a coexisting distributed and efficient drainage system. We apply the model both to a synthetic system and to Recovery Ice Stream, with bed topography developed from mass conservation inversion techniques. The model outputs of lake growth and drainage are compared with ICESat surface altimetry measurements. Here we present our method and initial results from the hydrology model.


Subglacial hydrology as a control on emergence, scale and spacing of ice streams

Teresa Kyrke-Smith, Richard Katz, Andrew Fowler

Corresponding author: Teresa Kyrke-Smith

Corresponding author e-mail: terkyr@bas.ac.uk

Observations have long associated ice streams with the presence of meltwater at the bed. More recently, theoretical models have been able to reproduce ice-stream behaviour as a consequence of the coupled dynamics of ice and subglacial meltwater. In this presentation we analyse the properties of ice streams that form in a coupled model of ice flow and subglacial hydrology. We see that there is a natural length-scale defining ice stream width and seperation. This arises as a result of the balance between effective pressure gradients driving meltwater away from ice streams and gradients in hydrostatic pressure driving water towards ice streams. We further discuss how the model interacts with topography; small topographic variations can have a strong effect on where ice streams emerge in the model. However, streams evolve to be closer to the dimensions defined by the natural length-scale of the unperturbed system. The non-dimensional parameter that defines this length-scale is therefore of fundamental importance in the model.


Tributary interaction during speed-up events of Black Rapids Glacier

Martin Truffer, Lee Petersen, Christopher Larsen

Corresponding author: Martin Truffer

Corresponding author e-mail: truffer@gi.alaska.edu

Black Rapids Glacier, Alaska, is a surge-type glacier currently in its quiescent phase. Its flow is dominated by two main tributaries. The tributary that is less active during quiescence becomes the dominant contributor of ice during surge events. This interplay causes the characteristic looped moraines. We used annually occurring lake drainage events as natural experiments to document this flow switch and instrumented the confluence area with high-precision GPS instruments. The velocity measurements show that the speed-up events behave like mini-surges. They occur on the usually slower branch, where the speed increases fivefold or more over the course of several hours. The other branch only shows a minor reaction to these events. At the conclusion of these speed-up events the velocities revert back to normal. We hypothesize that the usually dominant branch creates a ‘dam effect’ and allows the slower tributary to build up to surge geometry by allowing unusually high stresses, which can eventually lead to a surge. However, decreasing mass balances, particularly since the mid-1990s, have resulted in a glacier surface in the reservoir area that is no longer thickening. It thus appears that Black Rapids Glacier is not capable of building up to surge geometry in the current climate.


Stress and strength immediately adjacent to a subglacial drainage channel in an ice-stream shear margin

John D. Platt, Thibaut Perol, Jenny Suckale, James R. Rice

Corresponding author: John D. Platt

Corresponding author e-mail: jplatt@dtm.ciw.edu

The mass loss from the West Antarctic ice sheet is dominated by numerous rapidly flowing ice streams, which are separated from the stagnant ice in the adjacent ridge by a zone of concentrated deformation known as a shear margin. Since the discharge from an ice stream is thought to depend sensitively on the ice-stream width, determining the physical processes that control the shear margin location is crucial to understanding how ice streams may respond to a changing climate. The transition from a deforming to an undeforming bed within the shear margin concentrates large stresses on the undeforming bed beneath the ridge. Thus, for a stable shear margin configuration – where the yield strength of the bed is always greater than the stress applied to the bed whenever the till is undeforming – to exist there must also be a mechanism that raises the basal strength within the shear margin. Several papers have looked at how freezing of the subglacial till that the ice stream rests upon may provide this strengthening, and recent work has also studied how the development of a subglacial drainage channel could strengthen the bed by lowering the pore pressure in the till. In this presentation we investigate how the presence of a drainage channel at the transition from a deforming to an undeforming bed alters the stress field within the shear margin. We show that the channel limits the maximum shear stress resolved on the undeforming bed, and alters the yield strength of the till by changing the normal stress applied to the ice–till interface. Comparing the maximum stress on the bed with the till strength we show that the transition from a deforming to an undeforming bed across a channel is a stable configuration if the flux in the channel exceeds a critical value. Using parameter values thought typical for major ice streams we find that this critical flux is unrealistically large, and thus the transition from a deforming to an undeforming bed does not occur at a drainage channel. To conclude we investigate the sensitivity of our model to several key parameters. These results show that the large uncertainty in determining the appropriate rheology for the highly stressed ice adjacent to the channel may lower our prediction for the critical flux by several orders of magnitude, possibly allowing a transition from a deforming to an undeforming bed across the channel to be a mechanically stable configuration.


Monitoring water bodies underneath glacier cauldrons with radio-echo sounding

Eyjólfur Magnússon, Finnur Pálsson, Magnús T. Gudmundsson, Þórdís Högnadóttir, Laurent Mingo, Baptiste Gombert

Corresponding author: Eyjólfur Magnússon

Corresponding author e-mail: eyjolfm@hi.is

A large part of the active volcanic zones of Iceland are covered by glaciers. This includes some of the most active volcanoes and many powerful geothermal areas. Volcanic and geothermal activity melts the glacier from below. The meltwater often accumulates at or near the source before flowing towards and from the glacier margin in jökulhlaups of variable magnitude, from minor increase in river flow (~m3s–1 peak flow) to catastrophic jökulhlaups (~105–106 m3s–1 peak flow). Persistent geothermal activity produces depressions in the glacier surface commonly referred to as cauldrons. These depressions often cause a local minimum in the subglacial water pressure, resulting in water accumulation at the glacier bed under the cauldron. To monitor the hazard from a cauldron, measurements showing whether water is accumulating under a cauldron, and ideally how much has accumulated, are needed. Such monitoring has, until recently, mostly been based on repeated measurements of surface elevation across the cauldrons. Decrease in the cauldron depth is generally considered as a sign of water accumulation. Such measurements do not always reveal water accumulation, particularly if the meltwater is produced directly under the cauldron, resulting in increased cauldron depth since the volume of the meltwater is less than that of the melted ice. In this study we present a method to monitor water accumulation underneath cauldrons using radio-echo sounding (RES) survey. RES profiles are measured repeatedly across the cauldron. Since 2012 over 12 cauldrons on Mýrdalsjökull ice cap, covering Katla volcano, have been surveyed with RES biannually. Beneath many cauldrons we observe changes in the strength or in the derived altitude of the measured backscatter. Under some cauldrons we observe a development in the backscatter that we interpret as a cycle: from absence of water under the cauldron then followed by water accumulation and consequent drainage in a jökulhlaup. We also present studies on subglacial water accumulation at two other areas applying the same method: (1) a cauldron in W-Vatnajökull, discovered when draining a substantial jökulhlaup (peak flow >2000 m3s–1) into a hydropower reservoir near the glacier margin; (2) at Bárðarbunga volcano where the ongoing caldera formation may cause water to accumulate within the caldera, a possible threat of a large jökulhlaup in the future.


Spatial variation of Greenland ice sheet meltwater export inferred from river discharge observations

Asa Rennermalm, Andreas Mikkelsen, Irina Overeem, Vena Chu, Laurence Smith, Dirk van As, Tom Mote

Corresponding author: Asa Rennermalm

Corresponding author e-mail: arennerm@rci.rutgers.edu

Greenland ice sheet mass loss has accelerated since the 1990s. However, the fraction of meltwater retained within the ice versus released to surrounding oceans is not well known. Proglacial discharge measurements between 2009 and 2012 at two sites along the Watson River in southwest Greenland are used to partition meltwater exported from the ice-sheet margin and interior. Meltwater export from the margin is disproportionally large relative to its small catchment area fraction (0.04–0.06). Regardless, meltwater export originating from the interior regions dominates, despite almost 50% being retained/refrozen, presumably in firn, melt ponds, crevasses and macropores. Large interannual variability in total Watson River meltwater export is controlled by the interior, which demonstrates that Greenland ice sheet interior meltwater export to global oceans cannot be ignored.


Short-term velocity variations and sliding sensitivity of a slowly surging glacier

Gwenn Flowers, Alexander Jarosch, Patrick Belliveau

Corresponding author: Gwenn Flowers

Corresponding author e-mail: gflowers@sfu.ca

Glacier surging is understood as an episodic flow phenomenon closely linked to basal hydrology. There is some evidence, even from temperate glaciers, that the dramatic fast-flow phase of a surge cycle may be preceded by a decades long acceleration. The term ‘slow surge’ has been coined to describe the occurrence of this accelerating phase in the absence of a subsequent fast phase. We use daily surface velocities measured over two summers on a slowly surging glacier in Yukon, Canada, to examine the interplay between ordinary melt-season dynamics and secular changes associated with the surge. Three GPS stations were installed in 2007 and 2008 in the ablation area of the polythermal study glacier. Previous work has shown that the highest measured flow rates and the largest seasonal flow anomalies occur within a ~1.5 km long zone occupied by the middle and upper GPS stations, where sliding accomplishes 50–100% of the annual displacement. GPS data were processed in an absolute reference frame as daily statics using GAMIT/GLOBK, and velocities produced by numerical differentiation of these records corrected for observed station tilt associated with ablation. A comparison of these daily static receiver positions with relative ‘moving baseline’ processed (RTKLIB) sub-daily solutions resulted in good overall agreement between the two processing strategies. We find that horizontal velocities are often in phase between receivers, except during larger velocity events that seem to initiate within the ~1.5 km long zone of fastest flow. Velocity variability within this zone is significantly higher than just below it, where summer speeds are close to the annual mean. Horizontal and vertical strain rates estimated from the GPS data imply thickening within the instrumented region, consistent with previously modelled drawdown above. The estimated thickening rate at the surge front, marked by an abrupt transition from fast to slow flow, is comparable with the annual net balance, suggesting that the surge front may not propagate downstream from its current position. Using simple models of springs, dashpots and frictional elements, we explore longitudinal coupling within the fast-flow zone to test the hypothesis that this zone drives the dynamics of the ice above and below it. We speculate that the ongoing slow surge enhances the glacier sensitivity to ordinary seasonal processes, including short-term summer sliding events.


A multi-sensor approach to continuous monitoring of a glacier outflow channel: Virkisá River, SE Iceland

Andrew Black, Jez Everest, Alan MacDonald, Verity Flett, Heiko Buxel, Tom Shanahan

Corresponding author: Andrew Black

Corresponding author e-mail: a.z.black@dundee.ac.uk

Continuous monitoring of meltwater river flows provides valuable data in understanding glacier response to changing climatic conditions. Where glacial catchment runoff converges into a single channel, continuous flow measurement enables total catchment outflow to be quantified, comprising total melt of snow and ice, plus rainfall-derived runoff. However, in practice, the derivation of continuous flow data is often constrained by practical considerations – such as ice affecting the continuous monitoring of water levels and frequently shifting channel features. A multi-sensor approach has been adopted in monitoring the Virkisá River, SE Iceland, to attempt to address these challenges. A road bridge is used as the gauging site, given the otherwise braided nature of the channel, and the structure of the bridge as a robust monitoring platform. Water level sensors have been installed in the centre and margin of the river channel, to generate independent measurements of water level. Water temperature is recorded by one of the sensors, and contributes to the quality assurance of the water level time series. The site is calibrated by periodic flow gaugings obtained either by wading or, in higher flows, by bridge gaugings. A water-surface velocity sensor facilitates a continuous, independent and indirect check on the calibration. This permits the detection of any rating shifts which may occur between infrequent site visits. Finally, two cameras provide visual information on conditions, allowing assessments of the presence or absence of ice, evidence of control changes, and unexpectedly detecting the temporary diversion of the channel while bridge maintenance was undertaken. In the first 3 years of operation, the resultant flow record after quality control is 83% complete, yielding a catchment-average annual runoff of 5.7 m. The multi-sensor approach has proved successful in generating reliable data and allowing new insights into the melt behaviour of the glacier, and surface water/groundwater interactions in the adjacent sandur.


Subglacial conduits in sediments

Ian Hewitt

Corresponding author: Ian Hewitt

Corresponding author e-mail: hewitt@maths.ox.ac.uk

Conduits that transport water at the base of an ice sheet or glacier may form through a combination of melting upwards into the ice and erosion downwards into the bedrock or subglacial sediments. The case of a conduit on hard bedrock, when melting causes a roughly semicircular incision upwards into the ice, is often referred to as an R-channel. A relatively standard set of equations has developed to model the evolution and efficiency of such channels. The opposite case of a conduit eroded predominantly into soft sediments is often referred to as a canal; these are generally considered to be wide and shallow features. Models for this type of conduit are much less well established; previous models assume the canals are in steady state, and make ad hoc assumptions about the balance of processes controlling the position of the walls. Much of the current thinking around subglacial hydrology is predicated on the R-channel type model, but the prevalence of R-channels beneath the Greenland and Antarctic ice sheets is poorly known. Moreover the different types of conduit are often thought to have quite different characteristics, particularly in their control of water pressure. Since most regions of dynamic ice activity are thought to be underlain by sediments, a better understanding of if and how conduits develop in this setting is necessary. In this study, a relatively simple set of equations is developed to model the evolution of sediment floored channels in an analogous way to an R-channel. The model requires consideration of the energy balance that results in melting of the ice roof, and also the erosion, deposition and creep of the sediments. Implications for the evolution of large-scale drainage systems over subglacial sediment will be discussed, and for subglacial floods. It is also hoped that this work will pave the way to further process-based considerations of subglacial erosion and landform development.


Detection of the ~12 year cyclic surge-like event at Donjek Glacier, Yukon, Canada

Takahiro Abe, Masato Furuya

Corresponding author: Takahiro Abe

Corresponding author e-mail: abetaka@frontier.hokudai.ac.jp

Glacier surge exhibits orders-of-magnitude speed-up, resulting in km-scale terminus advance. Although there are many surge-type glaciers near the border of Alaska and Yukon, Canada, their recurrent intervals are quite uncertain due to the lack of continuous observations. To better understand the surge dynamics and predict the next surge, it is essential to examine the entire surge cycles. Here we use Landsat optical imageries acquired between 1986 and 2014 to derive the spatial-temporal changes in the velocity fields, and report three surge-like episodes in 1989, 2001 and 2013 at Donjek Glacier in Yukon, Canada. Donjek Glacier is located in the Donjek River Valley system in the northern Yukon, which consists of Steele, Spring, Donjek and Kluane glaciers; all these are surge-type (Clarke and Holdsworth, 2002). Although some previous studies reported the surge event in the 1970s (Johnson, 1972; Clarke and Holdsworth, 2002) and Donjek Glacier is known as a polythermal glacier (Clarke and Holdsworth, 2002), there is no report about the recent surge. Abe and Furuya (2014) reported the spatial-temporal changes in the ice velocity during its quiescent period from 2006 to 2011, but their long-term evolution remains uncertain. We apply the feature-tracking in frequency domain with orientation images (Heid and Kääb, 2012) to the B4 images for Landsat 5 and the B8 images for Landsat 7/8. We also examined the terminus area fluctuations associated with the surge events using the composite false images, taking advantage of the clear contrast between ice and rock (McNabb and Hock, 2014). In the 2001 event, the surface speed significantly increases by up to 5 m d–1; during the quiescent phases it was ~0.5 m d–1 at the terminus. While the duration of active phase is about ~3 and ~1 year in the 2001 and 2013 events, the phase in the 1989 event is unclear because of the lack of high temporal resolution data. Moreover, the terminus area cyclically expanded by about 3–4 km2. Given the previous report of surging in the late 1970s (Clarke and Holdsworth, 2002), it turns out that the surge-like episode has been repeating quite regularly every ~12 years, which indicates the next episode is very likely to occur in 2025. The recurrent interval is relatively shorter than those at other surge-type glaciers in Yukon. We discuss the dynamics of Donjek Glacier in terms of glacial hydrology and geomorphology, comparing the glacier pulses reported by Turrin and Foster (2014).


Mass balance and ice dynamics on the northern slope of Mýrdalsjökull, Iceland

Christoph Mayer, Ludwig Braun, Christof Völksen, Julia Jaenicke, Ulrich Münzer

Corresponding author: Christoph Mayer

Corresponding author e-mail: Christoph.Mayer@kfg.badw.de

Icelandic ice caps are characterized by high accumulation rates during winter and strong surface melting during summer. Therefore, it is rather difficult to establish and maintain mass-balance monitoring programs on such glaciers. With the development of a new generation of radar satellite systems, e.g. the TanDEM-X mission, seasonal time series of bulk geodetic mass-balance calculations seem possible. DEM differencing allows the analysis of spatial volume change patterns, but for the derivation of mass-balance values also the contribution of the dynamic mass redistribution is required. In our approach we use localized GPS measurements for the calibration of the horizontal velocity field derived from feature-tracking of optical satellite imagery. The vertical ice velocity is also determined by GPS measurements at individual stakes, while the spatial distribution requires additional information. We utilize the geometry change of volcanic depositions to derive the missing information. This allows us to determine local mass changes (loss and redistribution) for a considerable part of the ablation zone.


Water-saturated sediments beneath Pine Island Glacier, West Antarctica

Andy Smith, Alex Brisbourne, Edward King, David Vaughan, Damon Davies, Robert Bingham

Corresponding author: Andy Smith

Corresponding author e-mail: amsm@bas.ac.uk

Pine Island Glacier is a rapidly changing glacier in West Antarctica which, of all the world’s glaciers, currently makes one of the greatest contributions to sea-level rise. These on-going changes are almost certainly driven from the ocean offshore and its interactions with the floating ice shelf, but understanding what controls the glacier’s response to variations in the ocean is still unclear. iSTAR is a UK NERC funded programme studying the whole ocean, ice shelf, glacier and drainage basin system of Pine Island Glacier. As part of iSTAR, two ‘over-snow traverses’ have completed journeys across Pine Island Glacier and its drainage basin, conducting a range of glaciological and geophysical work, both en route and at specific locations. On the most recent of these traverses, seismic data were acquired on the main glacier trunk and on some of its tributaries to determine the nature of the bed material, and to understand both the basal conditions and their potential influence on the ice flow. The different tributaries are responding to the forcing variations in different ways and this work is aimed at understanding the role that basal conditions play in controlling and changing the ice flow. Initial results from the seismic surveys show that Pine Island Glacier and its tributaries are largely underlain by water-saturated sediments. Sediment porosity and hence water content varies, implying different degrees of active sediment deformation in different places. Complementary ground radar surveys show a range of different landscapes beneath the ice, with numerous drumlins and elongated bedforms being common in some places, and rare elsewhere. Broad correlations exist between this subglacial landscape and the sediment characteristics and water content interpreted from the seismic data. Earlier work on the main glacier trunk led to the hypothesis that active erosion of the water-saturated sediments beneath the ice could itself trigger ice flow changes in the near future, irrespective of any external forcing from further downstream. Analysis of one previous seismic line repeated on the iSTAR traverse will be used for an early test of this hypothesis.


Iceberg distribution in the Amundsen Sea (Antarctica)

Aleksandra Mazur, Anna Wåhlin, Adam Krężel

Corresponding author: Aleksandra Mazur

Corresponding author e-mail: aleksandra.mazur@ug.edu.pl

Icebergs are masses of freshwater ice that are calved from the margins of ice shelves and glacier tongues. They account for a significant part of the freshwater flux and may affect hydrological and biological processes in the basin. The West Antarctic is one of the fastest-warming areas on Earth. Ice shelves in the Amundsen Sea which drain the West Antarctic ice sheet are thinning rapidly. Although large quantities of glacial meltwater have been found in the water column on the Amundsen Sea shelf, it is not known whether this meltwater comes from the ice shelves or from the icebergs already detached from the glaciers. Therefore to better understand processes that occur in the basin, iceberg studies are essential. Visible and infrared sensors are often blinded with clouds or darkness, and for this reason radar data are the most suitable for iceberg analysis. In the studies 431 ASAR (advanced synthetic aperture radar) wide swath mode images acquired between 1 January and 31 December 2011 were used. Classification of icebergs was performed with the usage of object-based image analysis (OBIA) methods and eCognition Developer software. The verification was carried out on four polygons at each season. In 12 of 16 polygons calculated kappa coefficient (K) were higher than 0.86. Based on the classification results the probability of iceberg occurrence in the Amundsen Sea was calculated. It can be clearly seen that iceberg distribution follows the bathymetry of the basin. It can be also noticed that the majority of icebergs that are formed in the Amundsen Sea tend to melt within the basin. As a next step the size and number of icebergs in the region were estimated. In areas with the highest probability of iceberg occurrence the average number of icebergs present in a 15 × 15 km grid is between 10 and 15 with average total area between 20 and 25 km2. The majority of icebergs present in the Amundsen Sea have a size less than 5 km2. In general the biggest icebergs are formed close to Thwaites ice tongue and in the western part of the Getz Ice Shelf. Based on the sizes of detected icebergs, the average total ice mass present in the Amundsen Sea was also estimated.


Submarginal drumlin formation at Fláajökull, Iceland

Sverrir Jónsson, Ólafur Ingólfsson, Ívar Örn Benediksson, Anders Schomacker, Helga Lucia Bergsdóttir, William Jacobson

Corresponding author: Sverrir Jónsson

Corresponding author e-mail: saj7@hi.is

Fláajökull is an non-surge type outlet glacier draining the southeastern part of Vatnajökull ice cap. Like most glaciers in Iceland, Fláajökull had a stand still or an advance in the latter part of the 20th century. During this period the glacier built up a prominent end moraine. After 1995 the glacier has been continuously retreating, revealing a small cluster of drumlins. This study investigates the morphology and sedimentology of these drumlins. The drumlins have a core of glaciofluvial sediment or till that is draped by 1–3 m thick, massive subglacial traction till showing sheared microfabrics in the direction of ice flow. Sedimentological data indicate the drumlins were formed by a combination of erosion and sedimentation. The location of the drumlins just proximal, and partly under, the 1995 end moraine suggests that they were formed submarginally during the ~25 year long advance phase in the late 20th century.


Assessing the subglacial lake coverage of the Antarctic ice sheet

Sebastian Goeller, Daniel Steinhage, Malte Thoma, Klaus Grosfeld

Corresponding author: Sebastian Goeller

Corresponding author e-mail: Sebastian.Goeller@awi.de

The existence of water at the base of the Antarctic ice sheet has been proven by the observation of several hundred subglacial lakes within the last five decades. They accumulate basal meltwater from their upstream catchment areas and release it continuously or in surges towards other downstream located subglacial lakes and eventually to the ocean. Both subglacial lakes and basal water streams are recognized to decrease traction at the ice base and thus have a great impact on ice dynamics. However, the total extent of Antarctic subglacial lakes is still unknown, since particularly the detection of smaller lakes depends on in situ survey campaigns. We approach this question by combining modeling and remote-sensing strategies. Following Shreve’s hydraulic potential equation, we simulate possible lake locations and can thus recall the majority of the currently observed lakes. Additionally, we find many new and so far unexplored possible lake locations. In order to validate these predictions, we analyze ice-penetrating radar profiles from our radio-echo sounding flights of the entire last decade in Dronning Maud Land. This leads to the identification of 31 new subglacial lakes. Based on these findings, we estimate the total number of Antarctic subglacial lakes to be about 1300, a factor of three more than what has been discovered so far. Their overall extent is assessed to cover about 0.6% of the Antarctic ice–bed interface. Respective to assessments of the current inventory, we find that only about 30% of all Antarctic subglacial lakes and about 65% of the total estimated lake-covered area are discovered at present.


Hydrothermal state of glaciers in Nordenskiöld Land, Spitsbergen

Ivan Lavrentiev, Andrey Glazovsky, Yuri Macheret, Ilya Marchuk

Corresponding author: Ivan Lavrentiev

Corresponding author e-mail: ilavrentiev@gmail.com

Data of ground-based radar studies conducted in 1999, 2007 and 2010–13 at 15 glaciers in Nordenskiöld Land show that most of these glaciers are polythermal. This conclusion is based on radio-wave velocity (RWV) estimates, obtained from common midpoint (CMP) measurements, analysis of hyperbolic reflections from water inclusions in temperate ice, and power reflections from cold-temperate surface (CTS). Using these data, we estimate the volume of cold ice and temperate ice in these glaciers, water content and water volume in temperate ice and their spatial distribution. The study shows the differences in water abundance at CTS and changes in the hydrothermal structure of glaciers in the area.


Melt-pond formation on the Larsen C ice shelf, Antarctica

Suzanne Bevan, Adrian Luckman, Bernd Kulessa, Bryn Hubbard, David Ashmore, Peter Kuipers Munneke, Martin O’Leary

Corresponding author: Suzanne Bevan

Corresponding author e-mail: s.l.bevan@swansea.ac.uk

The Larsen C ice shelf (LCIS) is the largest ice shelf on the Antarctic Peninsula and is gradually being exposed to conditions that preceded the collapse of its northerly neighbour, the Larsen B ice shelf. Collapse preconditions include increased surface melt and firn densification, melt ponding, rift propagation and ice flow acceleration. Close to the grounding line and particularly on the northern part of the ice shelf, Föhn winds descending from the Antarctic Peninsula mountains can lead to positive surface air temperatures and increased summer surface melt. When the meltwater generated exceeds the storage capacity of the firn layer surface melt ponds may form. Melt ponds are therefore a clear indicator of a saturated firn layer, and have been implicated in crevasse hydrofracture during ice-shelf collapse. Optical televiewing of a borehole drilled in Cabinet Inlet on the LCIS in November 2014, as part of the MIDAS project, shows a solid ice body commencing within 3 m of the surface extending to a depth of 45 m. Structure within the ice body indicates interannual variability in the formation processes; from percolating meltwater refreezing within the firn or snowpack above an existing ice layer, to bulk refreezing of melt ponds. Melt ponds are visible in optical satellite imagery but are not observed every year; this may be because they only form episodically or because the cloud cover prevents sufficiently frequent observations. A further indication of recent firn saturation can be detected in the backscatter signal from active microwave winter images, where a drop in backscatter, compared with the high volume scattering indicative of distributed ice pipes and lenses, points to the presence of a solid ice layer with specular reflection characteristics. Here we will review the optical satellite record, including MODIS and Landsat images, and demonstrate, using high spatial resolution TanDEM-X DEMs, that when melt ponds form they occupy shallow troughs in grounding zone flow stripes. We also investigate the relationship between surface melt intensity and pond formation using Envisat ASAR data between 2006 and 2012, and explore the feasibility of continuing the analysis with the newly online Sentinel-1a SAR data.


Potential subglacial basins in the central Barents Sea: implications on the dynamics and retreat patterns of Sentralbankrenna

Mariana S.R. Esteves, Lilja R. Bjarnadóttir, Denise C. Rüther, Monica C.M. Winsborrow, Karin Andreassen

Corresponding author: Mariana S.R. Esteves

Corresponding author e-mail: mariana.esteves@uit.no

Subglacial lakes play an important role within the subglacial hydrological system, directly influencing ice flow velocity, basal meltwater discharge and thus the ice-sheet dynamics and stability. Due to logistical difficulties, little is known about the processes and environments occurring within contemporary subglacial lakes, however, investigations of palaeo-subglacial lakes and the sediments within these basins can provide us with some clues. Using high-resolution bathymetric data provided by the MAREANO Programme, several potential subglacial basins have been located in the central Barents Sea. There are several channels leading to and from the basins and ultimately into Sentralbankrenna and we infer that these basins likely influenced the dynamics and retreat patterns of Sentralbankrenna ice stream. A transect of sediment cores will be collected from the potential lake basins in July 2015. Sedimentological, micropalaeontological and biogeochemical analyses will be undertaken in order to: firstly, determine the palaeo-processes and -environments within these basins during the last deglaciation; and, secondly, to further understand the role of these basins in the deglaciation and retreat of Sentralbankrenna ice stream.


Meltwater landforms in the central Barents Sea and their implications for former subglacial drainage

Calvin Shackleton, Lilja R. Bjarnadóttir, Monica C. M. Winsborrow, Karin Andreassen

Corresponding author: Calvin Shackleton

Corresponding author e-mail: calvin.s.shackleton@uit.no

The drainage of meltwater at the bed of an ice sheet has considerable influence on the dynamics of overlying ice. Many factors influence the supply and drainage of subglacial melt, which has important consequences for ice–bed interaction and flow velocities. Understanding the temporal and spatial scales of such variation and their implications both locally and at the ice-sheet scale are crucial to improving predictions of ice-sheet stability and change. Contemporary subglacial environments are remote and difficult to study, making investigation of palaeo-ice-sheet hydrology through the study of the erosional and depositional products of meltwater activity appealing. Bathymetry data acquired by the MAREANO Programme, covering an area of ~24 000 km2 at 5 m horizontal resolution, has revealed a number of potential meltwater features in the central Barents Sea, providing valuable insights into the nature of hydrological systems operating at the base of the former Barents Sea Ice Sheet (BSIS). Large, sinuous features up to 3.5 km wide and ~40 km long, containing numerous single channels with depths up to 40 m, are observed and interpreted to be braided tunnel valleys. Dendritic meltwater channels are also observed, up to 42 km long and 24 m deep, along with a small number of anastomosing meltwater features. Detailed mapping of the meltwater features and associated glacial landforms provides the basis for a comprehensive reconstruction of subglacial drainage systems and related ice-sheet dynamics in the little-studied central region of the BSIS. This yields useful insights into the organization of subglacial drainage in a marine-based ice sheet.


Geological evidence for past subglacial conditions: dry-based vs wet-based ice

Johann Philipp Klages, Gerhard Kuhn, Alastair G.C. Graham, James A. Smith, Claus-Dieter Hillenbrand

Corresponding author: Johann Philipp Klages

Corresponding author e-mail: johann.klages@awi.de

Amongst the geological community there is now consensus that the West Antarctic ice sheet reached outer continental shelf positions in most sectors during its last maximum extent. This conclusion is mainly formed on the basis of extensive palaeo-glaciological work that has been conducted during the past decade in various drainage sectors of the ice sheet. In these regions of the continental shelf, reconstructions of ice-sheet extent, relative flow velocities and retreat have mainly focused on deeper cross-shelf depressions, generally interpreted as pathways of fast-flowing palaeo-ice streams. Within these troughs, the prevalent glacial landforms consist of continuums of streamlined bedforms (e.g. mega-scale glacial lineations and drumlins) whose progressive increase in elongation is indicative of formation beneath wet-based streaming ice. However, extensive shelf areas surrounding these spatially restricted deeper corridors largely lack well-preserved glacial landforms, thereby inhibiting reliable shelf-wide ice-flow reconstructions. Recently, by analysing new high-resolution bathymetric data from these inter-ice stream ridges, former basal ice conditions could be illuminated for the first time. Subglacial landforms such as hill-hole pairs, sediment rafts and crevasse-squeeze ridges clearly indicate cold/dry-based, slow-flowing or even stagnant ice on the shelf areas outside the troughs. Here we present a new compilation of bathymetric datasets from the eastern Amundsen Sea Embayment covering both palaeo-ice stream troughs and adjacent inter-ice stream ridges. It allows a clear distinction between different basal ice regimes (dry-based vs wet-based) across the former ice-sheet bed. However, we additionally identified regions of apparent changes in basal conditions, i.e. regions where hill-hole pairs overprint mega-scale glacial lineations, thereby suggesting cold/dry-based basal conditions prior to final retreat rather than streaming flow. Generally, the ability to reconstruct the width and basal form of the West Antarctic ice sheet in this manner, across other parts of the Antarctic shelf, will greatly aid numerical ice-sheet models that aim to simulate past configurations of the ice sheet and its evolution to the present day.


Englacial channel knickpoint formation and migration from direct observations and modeling

Kiya Riverman, Richard Alley, Rudy Slingerland, Nathan Stevens

Corresponding author: Kiya Riverman

Corresponding author e-mail: kiya.l.wilson@gmail.com

Wintertime mapping of englacial channels in Svalbard shows that knickpoint migration dominates channel evolution. We present results from 2 years of glacial-speleological mapping efforts of englacial and subglacial channels on Larsbreen Glacier, Svalbard. Flat channels formed through cut-and-closure are interrupted by clustered knickpoints ranging in height from 1 to 10 m, which constitute ~66% of the overall elevation drop along the channel length. From former knickpoint scars we derive relative horizontal/vertical erosion rates, and find that horizontal migration of knickpoints is a dominant incision mechanism by which water cuts to the bed of the glacier. Observed englacial channels reach the bed by falling over a knickpoint, giving the streams particularly large bed erosive capacity. From observations, we conjecture that steep slopes are unstable and rapidly form knickpoints, creating the observed pattern of flat slopes punctuated by knickpoint clusters. We test these hypotheses by modeling knickpoint growth and migration using 3ST1D, a 1-D finite-difference model of open streamflow, capable of handling transcritical flow, coupled to our own derivation of ice melt and erosion. Modeled erosion rates are comparable with observations of about 0.6 m a–1, and increase with larger slopes. This drives a thermal instability of steep slopes, promoting knickpoint formation. We find that small perturbations in flow rapidly diffuse, but larger steps can persist and migrate. We suggest that knickpoints form through two processes: differential channel incision between debris-laden and adjacent clean ice, as well as erosion rate variations with water flow over sufficiently large bumps. This work demonstrates the importance of incorporating knickpoint processes into glacial hydrologic modeling.


Subglacial drainage at Paakitsoq, West Greenland: insights from two hydrological models

Alison Banwell, Ian Hewitt, Ian Willis, Neil Arnold

Corresponding author: Alison Banwell

Corresponding author e-mail: afb39@cam.ac.uk

Uncertainty remains about the ways in which the surface hydrology of the Greenland ice sheet influences its subglacial drainage system, affecting basal water pressures and ice velocities. Two main modelling approaches have emerged in the literature. The first uses real boundary conditions and inputs to simulate subglacial drainage for specific catchments in particular years. For example, Banwell and others (2013) use a surface energy-balance model and a surface water routing and lake filling/draining model to generate meltwater input hydrographs to their subglacial drainage model for Paakitsoq, a marginal (<20 km) region of West Greenland, for the year 2005. However, their subglacial drainage model is somewhat idealized as it is only composed of channels, and the locations of these are prescribed. The second approach uses synthetic boundary conditions (e.g. simplified ice geometry) and idealized water inputs to simulate generic conditions for typical years (e.g. Schoof, 2010; Hewitt, 2013; Werder and others, 2013). The representation of subglacial processes is more realistic in these models in that conduit locations are able to develop in response to variable meltwater inputs and pressure gradients, and can take the form of a linked-cavity/poroelastic sheet or channels, with interaction and switching between the two dependent upon water flux. Here to benefit from the advantages that each modeling approach offers, we run the Hewitt (2013) model using the Banwell and others (2013) boundary conditions and meltwater inputs for Paakitsoq. The combined model is used to quantify how the subglacial drainage system evolves during the melt season in response to lake drainage events and subsequent meltwater inflow to moulins. Sensitivity tests reveal how drainage evolution, including spatially and temporally varying water pressures, strongly depends on the pattern and density of ‘open’ moulins. A comparison of the original results of Banwell and others (2013) with the new results shows strong similarities; however, there are some key differences, particularly during the early and late melt season when the assumption of channelized drainage is not valid.


Ice-shelf fracture due to viscoelastic-flexure stress induced by fill/drain cycles of supraglacial lakes

Alison Banwell, Douglas MacAyeal

Corresponding author: Alison Banwell

Corresponding author e-mail: afb39@cam.ac.uk

Using a previously derived treatment of viscoelastic flexure of floating ice shelves, we simulate multiple years of evolution of a single, axisymmetric supraglacial lake when it is subjected to annual fill/drain cycles. Our viscoelastic treatment follows the assumptions of the well-known thin-beam and thin-plate analysis, but crucially also covers the Glen flow-law rheology. As the ice-shelf surface does not completely return to its un-flexed position after a 1 year fill/drain cycle, the lake basin deepens with each successive seasonal cycle. We evaluate the timescale over which a typical lake can reach a sufficient depth such that ice-shelf fracture ‘at a distance’ can occur in response to lake filling/drainage. We show that although this is unlikely during one fill/drain cycle, after multiple years, fracture is possible, assuming that surface meltwater availability is not limited, and that lake drainage at the end of each melt season is also possible. This demonstration is motivated by the proposition that flexure in response to supraglacial lake filling/drainage may cause neighbouring lakes to drain, which, in turn, may cause farther removed lakes to drain. Such self-stimulating behaviour may have contributed to the fragmentation process that accompanied the rapid break-up of the Larsen B ice shelf in 2002. Ultimate proof for our hypotheses, however, depends on observations, which are currently lacking.


Heat transfer in englacial channels: new insights from explicit numerical simulations of turbulent flow

Alexander H. Jarosch, Thomas Zwinger

Corresponding author: Alexander H. Jarosch

Corresponding author e-mail: alex@hi.is

We simulate turbulent flow and heat transfer of water in englacial channels and compare our numerical results to the most commonly used heat transfer parameterization in glaciology, i.e. the Dittus–Boelter equation. The three-dimensional flow is simulated by solving the incompressible Navier–Stokes equations utilizing a variational multiscale method (VMS) turbulence model and the finite-element method (i.e. Elmer-FEM software), which also solves the heat equation. Studying a wide range of key parameters of the system, e.g. channel diameter, Reynolds number, water flux, water temperature and Darcy–Weisbach wall roughness (which is explicitly represented on the wall geometry), we find the Dittus–Boelter equation to be inadequate for glaciological applications and present a new, highly suitable heat transfer parameterization for englacial/subglacial channels. This new parameterization utilizes a standard combination of dimensionless numbers describing the flow and channel (i.e. Reynolds number, Prandtl number and Darcy–Weisbach roughness) to predict a suitable Nusselt number describing the effective heat transfer and thus can be readily used in existing englacial/subglacial hydrology models.


Benchmark experiments for the comparison of subglacial hydrological models

Basile de Fleurian, Mauro A. Werder, Olivier Gagliardini, Eric Rignot

Corresponding author: Basile de Fleurian

Corresponding author e-mail: basile.defleurian@uci.edu

The recent emergence of a number of subglacial hydrological models allows us to obtain theoretical insights on the coupling between water pressure and the sliding of glaciers. In ice flow models, it is relatively clear what the simulated physics ought to be. Conversely, the physical processes incorporated into subglacial hydrology models are diverse as it is as yet unclear which ones are of relevance. An intercomparison of hydrology models will therefore need a somewhat different approach to the one used in the many ice flow model intercomparisons (EISMINT, ISMIP, etc.). Here we present a set of experiments that will allow the comparison of the behavior of different hydrology models. The design of the benchmark aims at allowing the participation of a wide range of models based on different physical approaches. We aim at evaluating the models with a focus on the effective pressure which has the most impact on the dynamics of glaciers. The final aim of this benchmark is to provide a tool to the ice dynamics modelling community which will help them select a hydrology model for coupling to their ice flow models.


Viscoelastic behavior of subglacial water chambers

Douglas MacAyeal, Grant Macdonald, Alison Banwell, Christine Dow, Olga Sergienko

Corresponding author: Douglas MacAyeal

Corresponding author e-mail: drm7@uchicago.edu

Sudden inflation of subglacial water chambers is a phenomenon that is common to a number of aspects of supra- and subglacial hydrology. Models of these phenomena have often relied exclusively on either elastic or viscous treatment of ice deformation. The length scales of flexure associated with elastic and viscous rheological treatment are different, due to the fact that elastic plate rigidity and viscous plate rigidity have different parameters (although both depend on H3, where H is ice thickness). This means that to estimate the surface and basal stresses associated with inflating or deflating water chambers, a viscoelastic treatment may have the most general applicability. Accordingly, we will present a viscoelastic treatment of subglacial water chambers using simple viscoelastic plate theory, and apply the treatment to an analysis of subglacial lake phenomena in Greenland and Antarctica.


Progressive formation of drumlins within the active drumlin field of Múlajökull, surge-type glacier, Iceland

Ívar Örn Benediktsson, Sverrir A. Jónsson, Anders Schomacker, Mark Johnson, Ólafur Ingólfsson

Corresponding author: Ívar Örn Benediktsson

Corresponding author e-mail: ivar_orn.benediktsson@geol.lu.se

In order to shed light on drumlin formation, the stratigraphy and morphology of drumlins within the active drumlin field at the Múlajökull surge-type glacier in Iceland were studied. A total of 110 drumlins were mapped and measured, and their internal stratigraphy described in four exposures. All exposures reveal several till units where the youngest till commonly truncates older tills on the flanks and the proximal side of the drumlins. Drumlins proximal to the 1992 surge moraine are relatively long and narrow while drumlins distal to the moraine are wider and slightly shorter. A conceptual model is proposed to explain the formation of the drumlins at Múlajökull. Radial crevasses formed in the glacier when it initially spilled onto the flat foreland. These crevasses led to spatial differences in normal pressure at the base of the glacier so that deposition was favoured beneath the crevasses and erosion between them. Consequently, the original crevasse pattern of the glacier controlled the location of proto-drumlins. Once the proto-drumlins were formed, a feedback mechanism was established leading to continued crevassing and increased sedimentation at the location of the proto-drumlins. The drumlins are then maintained and their relief and elongation ratio increases as the glacier erodes the sides and drapes a new till layer over the landform. Our observations of this only known active drumlin field may have implications for drumlins within well-known Pleistocene drumlin fields, and our model may be tested on modern drumlins and drumlin fields that possibly become exposed upon future ice retreat.


Architecture and structural evolution of a Little Ice Age end moraine at Múlajökull, surge-type glacier, Iceland

Ívar Örn Benediktsson, Anders Schomacker, Mark Johnson, Alessa Geiger, Ólafur Ingólfsson, Esther Ruth Guðmundsdóttir

Corresponding author: Ívar Örn Benediktsson

Corresponding author e-mail: ivar_orn.benediktsson@geol.lu.se

The internal architecture and structural evolution of the Arnarfellsmúlar end moraine at the Múlajökull surge-type glacier, central Iceland, is described in order to elucidate glaciotectonic processes during glacier surging. The moraine is 4–7 m high, 50–100 m wide and generally single-crested and composed of a highly deformed sequence of loess, peat and tephra that is draped by subglacial traction till up to the crest. The internal architecture is dominated by steep, high-amplitude overturned folds and thrusts in crest zone but open, low-amplitude folds on the distal slope. Section balancing suggests a décollement depth of 1.4 m and a total horizontal shortening of around 69%. This suggests that the glacier coupled to the foreland about 70 m up-glacier from its terminal position to initiate the formation of the moraine, most likely due to changes in submarginal pore-water pressure. The structural evolution is polyphase in that the formation commenced with a series of low-amplitude open folding of the foreland before piggyback thrusting set in. The first thrust developed some 40 m in front of the ice, inducing ~6 m displacement of the hanging wall. The hanging wall of the first thrust was later distorted by a second and a third thrust, with a displacement of ~7 and ~10 m, respectively. Overfolding occurred under or immediately in front of the ice margin and was partly synchronous to the thrusting, i.e. the strata got folded during the upward transport of the hanging wall along a thrust. Radiocarbon datings and analysis of tephra layers, along with historical references, bracket the formation of the Múlar moraine between AD 1717 and 1760, which suggests that Múlajökull had its LIA maximum and greatest surges earlier than many other surge-type glaciers in Iceland.


Spatio-temporal character of fog on Arctic glaciers from time-lapse photography

Hester Jiskoot, Amy Fossheim, Benjamin Fox, Mariah Hierath, Brad Danielson, Matt Nolan

Corresponding author: Hester Jiskoot

Corresponding author e-mail: hester.jiskoot@uleth.ca

This study uses time-lapse imagery and weather station data to present the first results of fog spatio-temporal patterns over two Arctic glaciers. For tidewater-terminating Belcher Glacier, Devon Island, Canada, 1–6 hourly images from three off-glacier time-lapse cameras were analysed over May–August 2007–09. For land-terminating McCall Glacier, Brooks Range, Alaska, hourly images from an on-glacier camera over May–September 2011–14 were analysed. From thousands of images on both glaciers we extracted over 5000 images with fog, which allowed us to analyse fog frequency, extent, height, temporal and spatial patterns, as well as concurrent cloud types, precipitation and sea-ice break-up dates. On both glaciers fog occurs on 25–35% of the images in all years. Seasonal patterns are comparable, with the highest fog frequency on Belcher Glacier in June–July (>30%) and on McCall Glacier in June–August (25–45% of images). On both glaciers fog migrates up-glacier, from either the proglacial fjord or the Alaskan North Slope coastal tundra. Maximum extent usually occurs between noon and evening, but daily timing patterns differ. During significant fog extent, which occurs roughly a quarter of the time, a large proportion of the ablation area is under dense fog that can be up to 100s of meters thick. Results for both glaciers suggest that Arctic fog has the potential to significantly affect surface ablation. Consequently, weather and energy-balance data below and above the fog layer were analysed to quantify the influence of fog on temperature and radiation at the glacier surface. Fog influence was separated into three conditions: fog under clear skies; fog under scattered cirriform clouds; fog under broken to overcast stratocumulus. These results help inform us about general influence of fog on glaciers. Many Arctic and Antarctic glaciers terminate in coastal regions where sea fog is a common cloud type during summer. Fog corresponds in part to sea-ice break-up and it is unknown how its occurrence will change with sustained warming. Little is known about the local and regional effects of coastal fog on glaciers, though fog has the potential to significantly influence glacier ablation due to its associated surface temperature inversions and radiative effects. The net effect of fog on glacier melt is still an enigma, largely because no analysis of concurrent factors has been done and no systematic analysis of fog over any glacier had been performed before the current study.


Basal conditions of the Greenland ice sheet from internal layering in radio echograms

Christian Panton, Nanna Karlsson

Corresponding author: Christian Panton

Corresponding author e-mail: panton@nbi.ku.dk

In this study we use information from an extensive radio-echo sounding dataset to identify areas of the Greenland ice sheet influenced by changing basal conditions. Based on an automatic algorithm for calculating the slope of the internal radio-echo layering, we identify areas where flow over a non-homogeneous basal environment has modified the layering. We find that large parts of the ice sheet have been or still are influenced by processes acting near the bed. This is particularly evident at the onset of ice streams, although less dynamic areas close to the ice divide also contain imprints of non-uniform basal conditions. Our results underscore the importance of basal dynamics for the flow of the Greenland ice sheet and the need to understand the dynamics of its subglacial environment.


Constraints on the magnitude and variability of basal water pressure, western Greenland ice sheet margin

Patrick Wright, Joel Harper, Neil Humphrey, Toby Meierbachtol

Corresponding author: Patrick Wright

Corresponding author e-mail: patrickjwright@gmail.com

Efforts to calculate basal hydraulic potential gradients and to model groundwater hydrology beneath the Greenland ice sheet (GrIS) require broad assumptions of basal water pressure values in addition to surface and bedrock topography datasets. To provide accurate constraints on basal water pressure for the western GrIS margin, we present a statistical analysis of a suite of water pressure records from 11 boreholes collected during 2010–13. The borehole sites span a W–E transect from near-margin locations with ice depths of 100–150 m to sites 34 km inland with ice depths >800 m, and include four continuous over-winter records. Data from all boreholes with ice depths greater than 100 m show an annual mean basal water pressure of approximately 95% overburden (OB), which is largely influenced by a winter mean pressure of approximately 96% OB. Despite the larger pressure range for periods of summer diurnal variability, pressure values during the melt season remain high, with a mean value of approximately 92% OB. The overall magnitude and range of pressure does not show a significant trend as a function of ice depth throughout the transect, with minimum pressure values never below 75% OB, and maximum pressures reaching >100% OB. We present analysis of the time distribution of pressure throughout characteristic diurnal cycles, and perform Fourier analysis on summer diurnal periods, with evidence suggesting that semi-diurnal frequencies are forced by Earth tides. These results can provide initial and boundary conditions for multi-scale modeling studies of basal hydrologic processes.


Numerical modelling of bedrock erosion by sediment-transporting subglacial water

Flavien Beaud, Gwenn E. Flowers, Jeremy G. Venditti

Corresponding author: Flavien Beaud

Corresponding author e-mail: fbeaud@sfu.ca

The carving of tunnel valleys has been attributed to the action of water flow under former continental ice sheets. In Alpine settings, recent studies suggest that inner gorges (narrow canyons carved at the bottom of U-shaped valleys) persist through several glacial cycles. The relative importance of glacial erosion, glacial meltwater erosion and interglacial fluvial erosion remains largely unexplored. Numerical modelling studies of glacial erosion often neglect subglacial water flow or only consider its control on effective pressure and sliding speed. We examine the role of bedrock erosion by subglacial water flow in order to determine whether it makes a significant contribution to overall glacial erosion. We adapt the latest physics-based models of fluvial abrasion to the subglacial environment in assembling the first model that aims to quantify bedrock erosion caused by transient subglacial water flow. The subglacial drainage model consists of a one-dimensional network of cavities dynamically coupled to one or several R-channels. The bedrock erosion model is based on the tools and cover effect, whereby particles entrained by the flow impact the bedrock where it is exposed. We explore the dependency of glacial meltwater erosion on the structure and magnitude of water input to the system, the ice geometry and the sediment supply. We find that erosion is not a function of water discharge alone, but also depends on channel size and water pressure. Modelled glacial meltwater erosion rates are one to two orders of magnitude lower than the expected rates of total glacial erosion required to produce the sediment supply rates we impose, suggesting that glacial meltwater erosion is negligible at the basin scale. Nevertheless, due to the extreme localization of glacial meltwater erosion (at the bottom of R-channels), modelled incision of Nye (bedrock) channels several centimetres deep and a few meters wide can occur in a single year. Such incision rates argue for the ability of subglacial water flow to enhance the relief or even initiate the carving of an inner gorge and explain the gradual carving of a tunnel valley.


Short duration event recorded in multiple instruments at a field site located 35 km inland from the western margin of the Greenland ice sheet: possible downgradient effects of a supraglacial lake drainage?

Nathan Maier, Neil Humphrey, Joel Harper

Corresponding author: Nathan Maier

Corresponding author e-mail: nmaier@uwyo.edu

Continuous monitoring of an in situ vertical strain sensor array, GPS surface array, and basal water pressures recorded a short duration event ~2.5 days (DOY 209–212) during the summer of 2014, at a field site located 35 km inland from the western margin of the Greenland Ice Sheet. Despite its small magnitude, the event triggered a unique response in all field instrumentation. The event had two distinct phases. During the initial phase, GPS velocities deviated from diurnal cycles, rapidly peaking at a velocity double the mean velocity (DOY 204–213). This subsequently was followed by an 8.4 hour decline before returning to normal diurnal velocity cycles. Coincident with the rapid velocity increase, basal water pressures from a borehole deviated from diurnal pressure cycles and recorded a head elevation decrease from 604 to 587 m. Basal water pressures then fluctuated irregularly without returning to diurnal cycles for the remainder of the phase. Compressional vertical strain rates increased 9.5 hours prior to the event, but remained constant through the initial phase. The subsequent second phase of the event exhibited a change to extensional strain recorded by 15 in situ vertical strain sensors installed in a full-depth borehole. Surface velocities during the second phase were elevated. Basal water pressures fluctuated irregularly, with an increase in head elevation from 587 to 630 m recorded during the final 30 hours of the second phase, which terminated with the return to compressional vertical strain. Basal water pressures did not return to diurnal pressure cycles for the following day, which is the remainder of the measurement record. During the event, eight supraglacial lakes were identified using MODIS imagery to have drained approximately 50 km upgradient from the instrumented field site, with no observable supraglacial lake drainages occurring during the previous 7 days. It remains unclear how supraglacial lake drainages influence ice-sheet dynamics at distances >10 km away from the drainage site. We discuss how the event could be the downgradient propagation of the disturbance generated through local uplift and ice acceleration at the drainage site. Our results suggest high intensity events, such as localized speed-ups caused by supraglacial lake drainages, can influence ice dynamics at distances greater than longitudinal coupling length scales of 4–10 ice thicknesses generally considered for ice sheets.


Modeling Canadian Rockies glacier melt response to climate change

Samaneh Ebrahimi, Shawn Marshall

Corresponding author: Samaneh Ebrahimi

Corresponding author e-mail: samaneh.ebrahimi@ucalgary.ca

Changes in glacier mass balance govern the hydrology of catchments through contribution of meltwater to river flows. Meteorological and glaciological observations are not always available to determine regional ice mass changes, so modelling of glacier sensitivity to climate variability is necessary to understand glacier response to ongoing climate change. This paper aims to find a relationship between mass-balance sensitivities and climate variables which can be used for glaciers in a similar climatic setting where mass-balance observations are not available. Here two automatic weather stations (AWS) were established near the equilibrium-line altitude and two in the forefields of Haig and Kwadacha Glaciers in the Canadian Rocky Mountains. The sites are about 1000 km apart but occupy similar environments, as outlet glaciers that flow to the southeast from icefields that sit atop the continental divide. Our study focuses on mean energy budgets for the ablation months May to September for the two sites. We first calculate the 1, 2 and 5 day melt using a full distributed surface energy balance and compared it with the measured snow/ice surface height at each AWS site, to calibrate the surface roughness. We then project the effect of hypothetical changes in meteorological conditions such as changes in temperature, wind speed, humidity, cloud cover and albedo on the energy balance. Glacier melt is modeled for the end of this century based on meteorological perturbations applied in isolation and in combination, based on CMIP5 RCP4.5 climate change scenarios for the Canadian Rockies. Projected changes in glacier mass balance and runoff are estimated for the major river systems that drain eastward from the Canadian Rocky Mountains.


A stationary ice-penetrating radar system for detection of transient glacio-hydraulic conditions

Laurent Mingo, Gwenn Flowers

Corresponding author: Laurent Mingo

Corresponding author e-mail: laurent.m@bluesystem.ca

The vast majority of ice-penetrating radar (IPR) studies have been carried out as spatial surveys on land, by air and from space. By contrast, few examples of fixed-position temporal IPR studies are found in the literature. Here we describe the design and deployment of a new stationary IPR tested in summer 2014 on a large outlet glacier of the St Elias icefields. The radar system was deployed on Kaskawulsh Glacier ~0.5 km from an ice-marginal lake dammed by the glacier flank. The lake is known to drain annually in a subglacial jökulhlaup. The radar system was programmed to perform a series of soundings every 4 hours, while the lake level was monitored with a pressure transducer. A dual-input simultaneous sampling scheme was implemented to allow one channel to capture the full airwave while the other captures the remainder of the signal at higher resolution. This feature can detect fluctuations related to glacier surface conditions or transmitter power. It also permits a trigger threshold set above the level of potential ringing while maintaining high resolution on the second channel. For stand-alone operation, a microcontroller-based timing engine, transmitter sleep mode and fiber-optics transmitter pulse drive were also implemented in the system. In its first deployment, the stationary radar system operated autonomously for 6 weeks from late July to early September. At the beginning of the deployment period when the lake was filling, strong and stable internal reflectors are clearly visible. Over the course of the following 2–3 days, coincident with the most rapid phase of lake drainage, these reflectors fade dramatically until nearly all internal reflectors disappear. The weeks that follow are characterized by a stable signal nearly free of internal reflectors. The synchronicity of the lake drainage, as detected by the pressure transducer and time-lapse imagery, and the marked change in radar signature demonstrate the ability of the system to capture time-variable changes in englacial properties. We speculate that significant changes in englacial water storage, several hundred metres from the margin of the draining lake, are responsible for the temporal change in the radar signal. We conclude that stationary IPRs can be used to monitor temporal changes in hydro-glaciological conditions, and are particularly well suited to situations where the appearance or disappearance of englacial water can produce clear time-varying radar signals.


Climate change impacts on snow, glaciers and hydrology in the Hindu Kush Himalaya and Brahmaputra basin for long-term effects on Bangladesh

Shahadat Hossain, Tawhidur Rahaman, Shakil Khan

Corresponding author: Shahadat Hossain

Corresponding author e-mail: shlitonbuet@gmail.com

The Hindu Kush Himalayan (HKH) region, sometimes called the ‘Third Pole’, is one of the largest sources of snow and ice outside the Polar regions. Over 1.3 billion people living in the region and the downstream plains depend on the meltwaters from the HKH mountains to meet their needs for drinking, irrigation, hydropower, navigation and other uses. Snow and glacial melts contribute to the flow downstream and are of particular importance in the low-flow season. Brahmaputra basin is one of the important basins of the HKH, about 651 335 km2 with China, India, Bhutan and Bangladesh, which serves almost 120 million people around it. The Brahmaputra River basin approach strikes a balance between the existing natural functions of the river system and societal expectations for livelihoods, industry, recreation, nature management and agriculture. Climate change, population growth, infrastructure development, land use change, rapid urbanization and inefficient water use are all having an impact on water resources for Bangladesh. Of all the drivers affecting the hydrological regime in the HKH region, climate change is of particular concern, as the warming trend in the region is significantly higher than the global average. The most visible impact of climate change in the region is the rapid retreat of glaciers, which affects water availability in the major river basins of the HKH, with serious implications for the lives and livelihoods of millions of people. The main objective of this research is to identify the relation of integrated river basin management to reduce physical vulnerabilities and improve food and energy security for downstream communities of Bangladesh in the HKH region, while recognizing upstream interests. The final outcome will come as actionable recommendations for integrated water resource management practices and policies, including measures for risk management and for equitable access to water for energy and food security, formulated, shared and used at basin and community levels.


Supraglacial lake evolution on Petermann Glacier, northwest Greenland

Grant Macdonald, Alison Banwell, Neil Arnold, Douglas MacAyeal

Corresponding author: Grant Macdonald

Corresponding author e-mail: gjmacdonald@uchicago.edu

Supraglacial lakes are important for their role in supraglacial and subglacial hydrology and associated influence on ice dynamics. Additionally, their drainage has been linked to the collapse of floating ice. Petermann Glacier, with a floating ice-tongue terminus and situated in the north-facing northwest of Greenland, grants an opportunity to study supraglacial lakes on floating and grounded ice in close proximity, sharing the same fjord and climate, and in a less-studied region of the ice sheet. Here we present a study that attempts to improve understanding of lake evolution in general and in particular how lakes on grounded and floating ice compare. We use Landsat imagery to monitor spatial and temporal evolution on Petermann Glacier and its catchment. We apply a lake identification and area algorithm and a lake-depth algorithm to create a catalogue of lakes and their geometry for the years 2009–2013. Using this, lake geometry and volume are assessed in the context of velocity and topography data collected from NASA IceBridge to compare lake development to elevation statistics, ice flow and location in relation to the grounding line and a large subglacial canyon. Furthermore, we present analysis of systematic differences in lake geometry and change between lakes on floating and grounded ice and higher-resolution analysis of specific lakes of interest. The implications of the results for lake evolution and ice dynamics, particularly in a warming climate, are assessed.


Modeling of glacier mass changes in the High Mountain Asia region since the1900s

Weijia Bao, Shiyin Liu

Corresponding author: Weijia Bao

Corresponding author e-mail: baowj@lzb.ac.cn

Glaciers in High Mountain Asia (HMA) have experienced quite a heterogeneous change of patterns in recent decades. Besides local factors, the spatial and temporal heterogeneity of climate and climate change also play a role in these differences. In this paper, forcing by high-resolution (0.5° by 0.5°) gridded monthly precipitation and air temperature datasets (the CRU TS3.20 dataset), more than 60 000 glaciers in the HMA region were reconstructed for mass change in the period 1901–2013 by using an enhanced degree-day mass-balance model. Seven parameters (e.g. gradients of air temperature and precipitation, degree-day factors) in this model were calibrated for 17 HMA glaciers of which mass-balance glacier data are available in the World Glacier Monitoring Service (WGMS) and then expanded to the rest of the glaciers in this whole region. Model uncertainties and climate sensitivities of glacier mass balance were also analyzed, and the spatial distribution pattern of glacier mass change since the 1900s in the HMA region was illustrated.


Mass and energy balance of Icelandic ice caps: influence of the 2010 and 2011 subglacial eruptions

Sverrir Guðmundsson, Finnur Pálsson, Guðfinna Aðalgeirsdóttir, Helgi Björnsson, Simon Cascoin, Þröstur Þorsteinsson, Hannes H. Haraldsson, Andri Gunnarsson

Corresponding author: Sverrir Gudmundsson

Corresponding author e-mail: sg@hi.is

The mass balance of Vatnajökull ice cap in Iceland has been monitored since 1991 and of the Langjökull ice cap since 1997. Automatic weather stations (AWSs) have been used to estimate the full energy balance of the glacier surfaces; around ten stations on Vatnajökull since 1996 and two on Langjökull since 2001. Following the subglacial eruptions in the Eyjafjallajökull volcano 14 April to 22 May 2010 and Grímsvötn volcano 21–28 May 2011, airborne tephra was deposited on the surface of all the major ice caps in Iceland. The mass balance and AWS records provide valuable data to study the impact of aerosol deposition on albedo and radiative forcing. Here we compare the mass and energy balances of the ice caps during the exceptional circumstances of 2010 and 2011 to the mass and energy balances both to the years before (1997–2009) as well as the years after the eruptions (after 2011). The 2010 Eyjafjallajökull eruption produced a thin layer of fine-grained dark trachyandesite tephra particles on the surface (fraction of a mm). This highly reduced the surface albedo and greatly enhanced melting, especially within the accumulation area, where up to 3 years of net accumulation melted. The resulting negative net balance in 2010 of Langjökull and western Vatnajökull was up to threefold the average during the preceding warm decade. The intensive melting and hence runoff rates are consistent with observed discharges of the rivers draining the ice caps. In 2011, basalt tephra dust from the Grímsvötn eruption in Vatnajökull was deposited on the ice caps. This resulted also in high albedo drop and intensive melting rates at areas close to the eruption site, especially on western Vatnajökull. However, the weather during summer 2011 was not favorable to ablation on northern Vatnajökull and Langjökull; an exceptionally cold period from early May to mid-June in 2011 with occasional snowfall delayed the start of the ablation season by 3 weeks compared with typical years. Thus, despite the lower albedo, causing fast melt rates after 11 June, the total summer ablation was close to the average for the years 1997–2009. Also, an area of ~420 km2 south of Grímsvötn was insulated by thepra (no melting). Our observation show that in the consecutive years the surface conditions were back to normal and hence mass and energy balance controlled by climate.


A comparative study of subglacial and supraglacial hydrology for the Kangerlussuaq sector, West Greenland

Katrin Lindbäck, Rickard Pettersson, Sam Doyle, Alun Hubbard, Dirk van As

Corresponding author: Katrin Lindbäck

Corresponding author e-mail: katrin.lindback@geo.uu.se

Meltwater plays a vital role in governing ice-sheet dynamics. Although considerable research has been undertaken to quantify surface melt on the Greenland ice sheet, knowledge of the basal hydrological regime at sufficient resolution and spatial coverage is currently lacking. Here we present subglacial hydrological drainage catchments, flow networks and sinks calculated using a high-resolution (150 m) basal digital elevation model of the 6500 km2 land-terminating Kangerlussuaq section of the Greenland ice sheet. The results indicate that surface meltwater delivered to the bed via moulins is predominantly routed towards the margin in well-defined trough systems. The abundance of small subglacial sinks (median area = 0.58 km2) in the study area mirror supraglacial lakes at the surface and indicate that there may be interannual water storage in ephemeral subglacial ponds. However, comparison of subglacial hydrology with supraglacial runoff, ice flow and proglacial discharge illustrates the complex spatial relationships of the routing of meltwater. Sensitivity tests, in which subglacial water pressure was varied by ±10%, indicate that the extent of subglacial drainage catchments may vary by up to 50% (1000 km2) in size due to lateral competition by adjacent subglacial catchments. These findings suggest that subglacial water routing may vary spatially on short timescales as the subglacial pressure regime changes during the year. Static delineations of drainage catchments are therefore likely to be to simplistic. A better understanding of the subglacial hydrological processes is therefore needed before conclusions can be drawn on the retention capacity of the Greenland ice sheet. The results presented in this study add to the currently limited knowledge of the basal hydrological regime of the Greenland ice sheet.


Seasonal ice-flow variability surrounding three Greenland rapid supraglacial lake drainages

Laura A. Stevens, Mark Behn, Sarah Das, Ian Joughin, Jeffrey McGuire, Tom Herring, Michiel van den Broeke

Corresponding author: Laura A. Stevens

Corresponding author e-mail: stevensl@mit.edu

Across much of the ablation region of the western Greenland ice sheet, hydro-fracture events related to supraglacial lake drainages rapidly deliver large volumes of meltwater to the bed and create conduits providing efficient surface-to-bed drainage networks for the remainder of the summer melt season. At a lake on the western Greenland ice sheet, we have observed that drainage events are preceded by a 6–12 hour period of ice-sheet uplift and/or enhanced basal slip. Our observations from a dense GPS network surrounding this lake allow us to determine the distribution of meltwater at the ice-sheet bed before, during and after three rapid drainages in 2011–13, each of which generates tensile stresses that promote hydro-fracture beneath the lake. These observations suggest that stress transients associated with enhanced meltwater transport to the bed beneath lakes are required to initiate surface-to-bed hydro-fractures in compressional lake basins. However, the context of these basally triggered rapid lake drainages have not been explored in relation to seasonal ice-flow variability in the region. Here we examine seasonal ice-flow and surface melt variability surrounding these three mid-June drainage events, investigating melt and speed-up relationships across the GPS network in the weeks leading up to and following each lake drainage. We investigate differences in the timing of melt onset and magnitude, seasonal ice velocities and post-drainage basin response to place the transient speed-up and uplift sourced from the lake drainages in context within overall melt-season ice flow. We find that early season (mid-June) rapid lake drainages (2011–13) induce uplift that persists for multiple days post-drainage, in contrast to a short-lived (<1 day) surface uplift event observed during a late-season (mid-July) rapid drainage at the same lake in 2006. This differential post-drainage uplift suggests that a less efficient subglacial hydrologic system during the early summer season is unable to dissipate the meltwater pulse quickly into the subglacial hydrologic network. These results examine the large uplift and speed-up events produced by lake drainage events within the seasonal evolution of ice flow, and indicate that ice-sheet response here is modulated, at least in part, by the seasonal evolution of the subglacial hydrological system.


Neoglaciation and Little Ice Age (LIA) glaciers and ice caps on Svalbard

Wesley R. Farnsworth, Ólafur Ingólfsson, Anders Schomacker

Corresponding author: Wesley R. Farnsworth

Corresponding author e-mail: WesleyF@unis.no

Holocene climate has displayed a range of variability, fluctuating between warm and cold as well as humid and dry conditions on the multi-decadal to multi-centennial timescale. Arctic glaciers and ice caps are particularly sensitive to changes in high-latitude winter precipitation and summer temperatures. Thus the reconstruction of Svalbard glacier chronologies through the Holocene will provide valuable insight regarding the response of glaciers on Svalbard to antecedent changes in climate. At present, knowledge of the spatial changes of glaciers and ice caps on Svalbard through the Holocene lacks key details. For example, the LIA is the best documented climate event during the Holocene and considered to be the main event driving maximum glacier extent during this period. But the spatial extent of glacial expansion is poorly understood as well as the onset, duration and causal factors remain underdeveloped. Additionally numerous glacial landforms and deposits that pre-exist the LIA features are present around Svalbard, yet little is known about the spatial and temporal relations of these landforms. This abstract introduces a 4 year doctoral project that will create detailed geomorphological maps of glacier forelands on Svalbard by analyzing aerial images in digital-stereo-view. Also, extensive sedimentological mapping in the field, coring of threshold lakes and retrieving samples for dating will be conducted. This project will strive not only to distinguish spatial trends in ice expansion of Svalbard glaciers and ice caps through the Holocene, but to correlate phases of ice expansion with climate periods, internal glacial processes, sea-level fluctuations, or some form of combination.


Geochemical and geological evidence for a palaeo-subglacial lake in Pine Island Bay, Antarctica

Gerhard Kuhn, Claus-Dieter Hillenbrand, Sabine Kasten, James A. Smith, Frank-Oliver Nitsche, Johann Philipp Klages

Corresponding author: Gerhard Kuhn

Corresponding author e-mail: gerhard.kuhn@awi.de

Subglacial meltwater facilitates rapid ice flow beneath concurrent ice sheets, and there is widespread evidence for a dynamic subglacial water system beneath the Antarctic ice sheet. It steers and affects the pattern of ice flow and is a direct result from boundary processes acting at the base of the ice sheet, i.e. pressure-induced basal melting. Consequently, the occurrence of subglacial meltwater plays an important role in bedrock erosion, subsequent re-sedimentation, and in shaping the topography of ice-sheet beds. Here we present new geological and geochemical data from sediments recovered on the West Antarctic continental shelf in Pine Island Bay that we interpret as reliable indicators for deposition in a palaeo-subglacial lake beneath the formerly expanded West Antarctic ice sheet, presumably during or following the Last Glacial Maximum. Characteristic changes of sedimentary facies and geochemical profiles within these cores taken on RV Polarstern expeditions ANT-XXIII/4 (2006) and ANT-XXVI/3 (2010) support the presence of an active subglacial lake system during the late stages of the last glacial period. These findings have important implications for palaeo ice-sheet dynamics, suggesting there was considerable water available to lubricate the bedrock–ice interface and deposit water-saturated subglacial sediments (soft tills). Based on our investigations performed so far, we suggest that the transition from subglacial lake to contact with the ocean took place in the early Holocene. During this time we speculate that the ice sheet thinned and successively transformed into an ice shelf with sub-ice cavities flushed by tidal currents. Based on bathymetric maps and relative sea-level curves we will aim to estimate ice thickness as the grounding line retreated across the subglacial lake threshold further inland. Our findings may also have implications for ice-sheet models, which have to consider the predominantly non-linear effects related to subglacial hydrology.


Reconstruction of a palaeo-subglacial lake network in Alberta, Canada

Stephen Livingstone, Daniel Utting, Chris Clark, Alastair Ruffell, Steven Pawley, Nigel Atkinson, Gunnar Mallon

Corresponding author: Stephen Livingstone

Corresponding author e-mail: s.j.livingstone@sheffield.ac.uk

Subglacial lakes have been widely documented since first being identified beneath the Antarctic ice sheet in the 1960s and comprise a significant component of the subglacial hydrological system. However, their investigation is largely limited to contemporary ice masses despite critical information that could be gleaned from palaeo-subglacial lake studies, including: (1) their influence on meltwater drainage, ice flow and ice streams; (2) details about how they relate to palaeo-floods, ice dynamics and sub-Milankovitch-scale climate events; and (3) as archives of long-term Quaternary climate change. They are also readily accessible, we can sample the sediments and map the landforms with ease and we have comprehensive information on lake-bed properties. Output from numerical ice-sheet models and the simple Shreve equation approach has been used to diagnose where subglacial lakes are likely to have occurred in the geological record. However, their identification remains controversial due to the difficulty in distinguishing their signature from proglacial lake deposits. Here we present new geomorphological, geophysical, sedimentological and dating evidence for the existence of a palaeo-subglacial lake network beneath the suture zone of the former Cordilleran and Laurentide ice sheets. These relatively small (~1 km diameter) palaeo-subglacial lakes manifest as flat-spots in a drumlin field and are perched in upland areas behind small ridges. The flat-spots, which comprise basins in-filled with glaciolacustrine deposits (radiocarbon ages >22 cal. ka BP), are associated with subglacial meltwater channels eroded into the underlying substrate and which trend downstream into eskers, proven from ground-penetrating radar. The channel-esker landsystems are interpreted as the signature of subglacial lake outburst events, which has resulted in a number of new insights regarding the style and evolution of lake drainage.


Roughness modulation of Greenland ice sheet hydrology and ice-flow dynamics

Stephen Livingstone, Andrew Sole, David Rippin, Jon Hill, Malcolm McMillan, Duncan Quincey

Corresponding author: Stephen Livingstone

Corresponding author e-mail: s.j.livingstone@sheffield.ac.uk

The contribution of the Greenland ice sheet (GrIS) to future sea-level rise is uncertain. Observations reveal the important role of basal water in controlling ice flow to the ice-sheet margin. In Greenland, drainage of large volumes of surface meltwater to the ice-sheet bed dominates the subglacial hydrological system and provides an efficient means of moving mass and heat through the ice sheet. Ice surface and bed roughness (amplitude and wavelength of topography) largely control where meltwater can access the bed, and the nature of its subsequent flow beneath the ice. However, no systematic investigation into the influence of roughness on Greenland hydrology and dynamics exists. Thus, physical processes controlling storage and drainage of surface and basal meltwater, and the way these affect ice flow are not comprehensively understood. This presents a critical obstacle in efforts to predict the fate of the GrIS. Here we present repeat high-resolution mapping of the ice surface drainage network (e.g. lakes, channels and moulins), quantify bed and ice surface roughness from existing radar data, and calculate ice velocities at monthly, seasonal and annual time intervals to investigate ice-flow variability. We focus on two neighbouring land-terminating regions of the western GrIS characterized by different bed roughnesses, the Russell/Leverett glacier catchments and the area directly to its south. The results allow us to identify and quantify the influence of bed roughness on the spatial and temporal pattern of surface and basal meltwater drainage and Greenland ice-flow dynamics.


Rapid and large-scale, subglacial erosion by a high-magnitude jökulhlaup, Skeiðarárjökull, Iceland

Andy Russell, Jonathan Carrivick, Daniel Parsons, Helgi Björnsson, Finnur Pálsson

Corresponding author: Andy Russell

Corresponding author e-mail: andy.russell@ncl.ac.uk

Subglacial erosion rates are virtually impossible to measure directly and are usually inferred from fluvial sediment yields or from net glacier bed elevation change. The November 1996 jökulhlaup from Grímsvötn subglacial lake propagated through Skeiðarárjökull, southern Iceland, as a high pressure englacial and subglacial flood wave. It reached a peak discharge of ~50 000 m3 s–1 within 14 hours. The snout of Skeiðarárjökull occupies a large subglacial overdeepening which caused floodwater to initially burst supraglacially from hydrofractures with a transition to subglacial efflux from numerous conduits across the entire 23 km wide ice margin. As the jökulhlaup progressed, discharge became focussed on the central part of the glacier margin where large-scale mechanical erosion of glacier substrate has been inferred from the presence of copious intraclasts (rip-ups) present within ice-proximal jökulhlaup deposits and the net deposition of 50–96 × 106 m3 of sediment in the ice-proximal proglacial area. Ice radar data collected in 1993 provides a baseline for the pre-1996 jökulhlaup bed elevation of this sector of Skeiðarárjökull. Since 1996, glacier margin stagnation, thinning and retreat resulted in the formation of a large proglacial lake. In 2008 and 2009, surveys of proglacial lake bathymetry were made from a small boat equipped with a 200 kHz single beam echo-sounder linked to real-time-kinematic GPS (RTK-GPS) permitting full differential corrections of position (to within ±0.02 m). When the 2008 and 2009 bathymetry was compared with the 1993 glacier bed elevation data 20–35 m depth of net bed elevation lowering was revealed over a distance of 400 m. Between 1992 and 2008 the margin of Skeiðarárjökull receded by ~1 km, thinned by up to 200 m and had negligible horizontal velocity. As the November 1996 jökulhlaup was the only major discharge of water from this sector of the glacier margin between 1992 and 2008 we ascribe measured bed elevation lowering to hydraulic scour during the rising stage of the November 1996 jökulhlaup. This study provides the first direct evidence of large-scale subglacial erosion by a well-constrained, short-lived high-magnitude jökulhlaup within a contemporary glacial system. Our findings demonstrate the importance of meltwater as a potent agent of subglacial erosion at glacier margins, enhancing glacial overdeepenings with subsequent implications for glacial and meltwater processes.


Spatial distribution changes of land-based glaciers and calving processes of tidewater glaciers in Admiralty Bay (King George Island, Antarctica) based upon satellite images since 1995

Joanna Sziło, Robert J. Bialik

Corresponding author: Joanna Sziło

Corresponding author e-mail: jszilo@igf.edu.pl

There is growing evidence that glacial systems in the Polar regions respond rapidly to climate change. Mostly, it is due to its location where the largest and most rapid changes occur. Therefore, the researchers have observed, with high interest, the spatial distribution of glaciers and calving processes, as they give information about recession or advance, which can be indirectly translating into the climatic conditions. For this reason, satellite remote sensing has been used to observe spatial distribution and geometry changes of land-based glaciers and estimation of calving processes of tidewater glaciers in Admiralty Bay (King George Island, West Antarctica). In the presented paper the focus is on the satellite images taken since 1995 in order to indicate ablation and accumulation zones of these glaciers. The second goal is to get information about suspended sediment concentration and difference in sea-water temperature, which gives an indication of localization of the actual outflows of fresh water in tidewater glaciers that may impact on the planctonic and benthic coastal biosphere. It is strongly suggested in the paper that the tidewater glaciers (i.e. Lange Glacier, Domeyko Glacier and Glacier Vieville) require continuous observation; in particular the prediction of calving processes should be of huge interest. This information is required to assess the rate of sea-level rise (SLR) as a result of calving processes in Admiralty Bay, Antarctica. The high-resolution images were used to recognize actual and previous glacial situation. Thus, the current images were compared with older ones, taken during the period of the last 20 years. This analysis allowed for the removal range fronts of the glaciers and can be useful to predict future range fronts. The current images were processed to distinguish places with fresh snow or ice, as well. Moreover, the suitability of the use of remote sensing in the glacial region of Admiralty Bay is also confirmed in this study. This kind of research could matter in the discussion of climate change and also in an attempt to predict the future situation of considered glaciers and draw conclusions about main trends. Due to the complexity of the issue, it is suggested that the research should include longer timescales and be complemented by measurements in the field (i.e. dynamics of the glaciers).


Röthlisberger channel model in ice-stream shear margins and mountain glaciers accounting for anti-plane shear loading and undeforming bed

Matheus Fernandes, Colin Meyer, Thibaut Perol, James Rice

Corresponding author: Matheus Fernandes

Corresponding author e-mail: fernandes@seas.harvard.edu

Ice-stream margins in West Antarctica form from the transition between fast-flowing and stagnant ice. Observations and modelling studies suggest that intense lateral deformation of ice at the margins generates internal meltwater within the ice sheet. Röthlisberger channels (R-channels) may form in which water flows in the downstream direction along the margin. Standard theory argues that the steady-state channel diameter comes from the balance of in-plane creeping and melting rates. However, at the margin, large shear strains acting in the direction parallel to their cross-sectional axis (in the anti-plane direction) can develop. In view of understanding how R-channels evolve within ice streams, we construct a finite-element model (FEM) of an R-channel as considered by Nye (1953) with superimposed anti-plane shear stresses applied along the bed of the ice sheet on the boundary outside of the channel. This is similar to the model presented by Weertman (1972) to describe R-channels formed at the base of mountain glaciers. Weertman’s analysis suggested that the presence of large anti-plane shear stresses, as compared with in-plane stresses, makes the in-plane flow law effectively linear viscous (pseudo-Newtonian). Using our FEM, we examine the limitations of Weertman’s solution as well as when his linear viscous model may be adequate. Furthermore, we examine the effects of shear stresses acting transverse to the channel axis as a result of the interlocking of the bed and the ice. This, as pointed out by Weertman, will change the in-plane solution from being axisymmetric; ultimately making the assumption of a Nye solution near the bed invalid. In addition, we attempt to modify the basal boundary conditions to incorporate the hydrological model proposed by Perol and others (2014), where the basal shear stress varies as a function of the distance from the channel. Ultimately, this makes a comprehensive model for R-channels with anti-plane shear stress hypothesized to exist in ice-stream shear margins and mountain glaciers.


Accumulation season activity of the subglacial drainage system of Werenskioldbreen, Svalbard: an analysis of glacier-derived naledi based on GPR/dGPS surveying and water chemistry

Lukasz Stachnik, Jacob C. Yde, Dariusz Ignatiuk, Marta Kondracka

Corresponding author: Lukasz Stachnik

Corresponding author e-mail: lukasz.stachnik@gmail.com

Glacier naledi (also referred to as icings and aufeis) are extrusive ice masses formed in front of some glaciers due to release of meltwater during the accumulation season. Although glacier naledi provide an opportunity to gain insights into the activity of the subglacial drainage system during the accumulation season, there have only been a few detailed studies on their characteristics. In this study, we examine glacier-derived naled assemblages in the proglacial zone of the polythermal glacier Werenskioldbreen (27 km2) in southwest Svalbard. The spatial distribution of the naled is determined by GPR/dGPS surveying and shows that naled genesis is due to release of water from four subglacial sources along the glacier terminus. One of these sources is related to subglacial portal that is active during the ablation season, and two sources may be linked to a multi-layered drainage network within a medial moraine. The GPR profiles show distinct differences between naled ice, sub-naled water lenses and bedrock. Sub-naled water samples were collected in order to determine the major ion concentrations. The ion association between Ca2+ Mg2+ and SO42– reflects a close coupling between carbonate dissolution and sulphide oxidation (r = 0.98; slope = 1.6) consistent with the local lithology, which is dominated by schist with carbonates and amphibolites. This is also consistent with high saturation indices of carbonates in whitish ice layers within the naled. We conclude that sulphide oxidation coupled to carbonate dissolution operates in the subglacial drainage system during the accumulation season. In general, the composition of the sub-naled water chemistry reflects the water chemistry in meltwater emanating from the subglacial drainage system during the ablation season.


An exploratory investigation of basal crevasses on Byrd Glacier, East Antarctica

Sarah Child

Corresponding author: Sarah Child

Corresponding author e-mail: s021c518@ku.edu

Ice sheets lose the majority of their mass through calving. Basal crevasses are believed to be linked to calving processes and in order to develop accurate calving models, it is first necessary to better understand basal crevasses. Basal crevasses are fractures found at the base of an ice sheet that propagate upwards. Several of these features are located on the floating trunk portion of Byrd Glacier, East Antarctica. The crevasses were identified from radar data, as hyperbolic returns, which were collected by the Center for the Remote Sensing of Ice Sheets (CReSIS) in 2011–12. The heights of the crevasses were measured from the radar data and these were combined with ice flow velocity data to provide a better sense of the stress regime acting on the ice at the time of the radar data collection. The purpose in this study of better understanding the stress regime is to explore the possibility of the role subglacial flooding plays in forming basal crevasses on Byrd Glacier. Specifically, exploring the idea that basal crevasses form when flooding has ceased and ice flow momentum at the grounding line not in equilibrium. When the floodwater supply at the bed of the glacier is ‘shut off’, bed lubrication is reduced, which causes grounded ice flow speeds to decrease, while the floating ice is still maintaining the same momentum for a period of time. During 2005–07, two lakes located 200 km from the grounding line of Byrd Glacier flooded causing the ice flow velocity to increase by ~100 m a–1. To investigate the possibility of flood-induced fracturing, a linear elastic fracture mechanics model is implemented using tensile stresses derived from the 2005–07 velocities. Tensile stresses derived from non-flooding years are then used in the model and the results are compared. A qualitative assessment of this theory of crevasses caused by subglacial flood is conducted by calculating the ice flow momentum equilibrium at the grounding line both at the end of the 2005–07 flood and non-flooding years. The findings from both investigations will show the likelihood of crevasse formation and/or if other external forcings acting on Byrd Glacier require more exploration in their role of basal crevasse development.


A virtual laboratory for understanding shear margin dynamics

Cooper Elsworth, Jenny Suckale

Corresponding author: Cooper Elsworth

Corresponding author e-mail: coopere@stanford.edu

The dynamics of ice-stream shear margins remain an important unknown in ice-sheet mass flux due to the margin’s control of stream width and velocity. On the Siple Coast in particular, ice-stream dynamics lack distinct topographic control and depend instead on characteristics of the unconsolidated subglacial till. This indicates a strong influence of subglacial hydrology on margin position and stability. Three potential controlling mechanisms are the transition from thawed to frozen bed, the formation of subglacial drainage channels, and the existence of sticky spots. This study aims to evaluate these mechanisms with comparisons to ice-stream observables and comment on hydrological implications of these basal strength signatures. The distinct behavior of these ice streams depends upon the strong coupling between till rheology, ice rheology and shear margin position. We propose a 2-D, time-dependent, thermomechanical model that resolves the margin location self-consistently through force balance. We assume a Mohr–Coulomb till, with basal slip occurring where the shear stress exceeds the local yield strength of the till. Yield strength is strongly dependent on pore pressure resulting in variations of basal stress across the width of the stream with the presence of subglacial water. The model can therefore provide constraints on basal strength distributions based on surface observations of margin position and stream velocity. Furthermore, we link melt production in the ice column to pore pressure in order to determine possible feedback between glacial melt and basal sliding. In addition, realistic ice rheologies are implemented to provide insight on rheological effects on model results. We present consistent basal strength distributions for velocity observations spanning the Siple Coast ice streams, with particular emphasis on margin migration. Implications of basal strength distributions on pore pressure variation are discussed and comparisons of margin stabilization mechanisms are performed. We determine that velocity distributions along the length of a stream require different basal strength profiles to match data, suggesting the development of different hydrological regimes. Time-dependent simulations provide insight into the evolution of subglacial conditions and effects on margin migration and stability. The proposed model allows for the comparison of different hydrological models with constraints provided by surface ice-stream data.


Variegated Glacier surge video

Hermann Engelhardt

Corresponding author: Hermann Engelhardt

Corresponding author e-mail: engel@caltech.edu

Variegated Glacier is a 20 km long surge glacier in Alaska with a surge recurring about every 20 years. The 1983 surge has been studied in detail showing the correlation of ice velocity mostly sliding and basal water pressure. A documentary of this surge will be presented in a video taken at the height of the surge.


Fractures as conveyors of englacial water flow

Andrew Fountain, Peter Jansson

Corresponding author: Andrew Fountain

Corresponding author e-mail: andrew@pdx.edu

A field investigation over a 3 year period on Storglaciaren showed that 79% of the 48 boreholes drilled encountered englacial hydraulic connections. These connections were inferred by a drop in water level from the glacier surface, during hot-water drilling and subsequent variations in water level over time. A borehole camera showed that in every case the hydraulic connection was intersection of a borehole and a subsurface fracture. Measurements of water movement typically showed speeds of ~1–2 cm s–1. The absence of Röthlisberger conduits was notable. This paper examines the meltwater input to the glacier and estimates whether a fracture network can accommodate the water flux at the flow speeds observed or whether englacial conduits are required to convey excess water. Conclusions from this work are important for models of water flow though glaciers and ice sheets, addresses the relevance of Röthlisberger conduits as the main pathway to convey water through a glacier, and considers the presence and evolution of englacial fractures in temperate glaciers.


Crevasses and supraglacial water as indicators of surge progression: a data analysis perspective

Ute Herzfeld, Thomas Trantow, Lukas Goetz-Weiss, Jessica Bobeck

Corresponding author: Ute Herzfeld

Corresponding author e-mail: uch5678@gmail.com

Glacial acceleration has been identified as the main source of uncertainty in sea-level rise assessment in the current realm of climatic warming, as established in the Fifth Assessment Report of the Intergovernmental Panel for Climate Change. Surging is one of only three types of glacial acceleration and the one that has seen the least amount of research, largely due to relative paucity of observations. The Bering–Bagley Glacial System (BBGS), Alaska, is the largest glacial system outside of Greenland and Antarctica and the largest surge-type glacier on Earth. During the recent surge, we collected airborne observations in four campaigns in 2011–13, including laser altimetry, imagery and GPS data, which are complemented by satellite imagery from WorldView-1,2 and Landsat-8 MABEL (Multiple Altimeter Beam Experimental Lidar) and CryoSat-2 altimeter data. We take an approach that utilizes the two most prominent indicators of surging – heavy and rapidly changing crevassing and presence of large amounts of supraglacial water – to document surge progression and to better understand the mechanisms that control surging. A newly developed form of connectionist-geostatistical classification is applied to associate crevasse patterns observed in airborne video imagery and in satellite imagery to deformation types. Analysis of laser altimeter data yields information on relative crevasse age and crevasse depths. The crevasse patterns serve as indicators of progression of the surge dynamics. Velocity maps are derived to supplement the crevasse classification. As supraglacial water is indicative of a blocked cavity system in the glacier, the appearance and disappearance of water-filled crevasses and supraglacial lakes provides temporal and spatial constraints on surge waves. Supraglacial water types are objectively mapped from remote-sensing data using a combination of spectral and connectionist classification. Together the results provide a better understanding of the spatially and temporally complex dynamics of the surge in the BBGS and surging in complex systems in general, as well as inputs for numerical modeling experiments.


Crevasses and supraglacial water as indicators of surge progression: a modeling perspective

Thomas Trantow, Ute Herzfeld

Corresponding author: Thomas Trantow

Corresponding author e-mail: uch5678@gmail.com

Glacial acceleration has been identified as the main source of uncertainty in sea-level rise assessment in the current realm of climatic warming, as established in the Fifth Assessment Report of the Intergovernmental Panel for Climate Change. Three types of glacial acceleration have been identified by Clarke (1987): ice streams, surge-type and tidewater glaciers. Surge-type glaciers, characterized by their quasi-periodic cycle between long quiescent phases of normal flow and short surge phases of accelerated flow and rapid advancement, have seen the least amount of research and are still incompletely understood. Many marine-terminating glaciers on the periphery of the Greenland ice sheet have either displayed surge behavior or are expected to be surge-type glaciers. Despite their significant number, current ice-sheet models used to predict glacial contribution to sea-level rise, such as the Community Earth System Model, do not account for the surge phenomenon, indeed treating all fast-flowing ice identically, therefore limiting certainty of estimations. The recent surge of the Bering–Bagley Glacial System, Alaska, serves as the prototype of a surge in a large and complex glacier system and provided a rare opportunity to collect observations. To better understand the mechanisms that control surging, methods of data analysis that utilize a number of satellite and airborne measurements together with numerical modeling experiments are employed and form a basis for investigation. A time series of glacier geometry has been derived using measurements from Cryosat-2, Operation Ice Bridge and laser altimeter data from four aerial campaigns by our group during the surge. Velocity maps derived from Landsat imagery have been assimilated into the model for investigations of basal conditions using the Ice Sheet System Model software. Forward modeling experiments simulating glacier dynamics on the order of a full surge-cycle are carried out using Elmer/Ice. New methods of surface detection and classification derived from high-resolution imagery are being developed to derive stress and strain rate fields for alternative ways to initialize and constrain the model for times when linear correlation methods, used to derive velocity maps, fail due to highly nonlinear surge behavior. Results from numerical modeling experiments are compared with results from observations and data analysis, with a focus on surge crevasse fields and glacial hydrology during the surge.


Possible changes in precipitation patterns associated with the retreat and thinning of Vatnajökull ice cap

Hálfdán Ágústsson, Haraldur Ólafsson, Finnur Pálsson

Corresponding author: Hálfdán Ágústsson

Corresponding author e-mail: halfdana@gmail.com

The large Icelandic glaciers have a significant effect on the spatial distribution of precipitation in Iceland. The annual maxima observed near the ice caps at the south coast of Iceland is associated with their high and broad orography as well as with the frequent passage of atmospheric lows and fronts. To quantify the effect of the glaciers on the flow and the precipitation patterns, two sets of high-resolution atmospheric simulations have been performed. The control simulation uses the current land height and glacial cover, while in the sensitivity run the glaciers have been removed and the bottom topography of the glaciers used as model land height. The simulations are done at 8 and 2 km horizontal resolution and are forced with the Interim re-analysis of the ECMWF for the period September 2004 to September 2006. The key results for Vatnajökull ice cap in southeast Iceland indicate up to 25% decrease in annual precipitation on large parts of the ice cap and an overall decrease close to 15% when the glacial cover is removed. There is furthermore greater spillover of precipitation in regions near the west and north edge of the ice cap but little changes further in the lee of the ice cap. The results of this study are of relevance for e.g. planning of hydropower availability and harnessing in a warming climate characterized by retreating glaciers.


Drainage networks, lakes and water fluxes beneath the Antarctic ice sheet

Ian Willis, Ed Pope, Gwendolyn Leysinger Vieli, Sylvan Long

Corresponding author: Ian Willis

Corresponding author e-mail: iw102@cam.ac.uk

The hydrology beneath the Antarctic ice sheet is poorly understood, despite its implications for subglacial lake behaviour, ice dynamics and mass balance. Here we use the BEDMAP2 datasets to calculate the subglacial hydraulic potential and the flow pathways of Antarctica’s subglacial drainage system. Many known subglacial lakes lie close to sinks in the hydraulic potential surface and to major flow pathways. We model patterns of basal melt across the continent, which we cumulate along the major flow pathways to quantify steady-state water fluxes from the main basin outlets at the ocean and from known subglacial lakes. The total flux from the continent is ~21 km3 a–1, equating to 6.0 × 10–5 m a–1 of global sea-level rise. The Byrd Glacier basin has the greatest single basin flux of ~2.7 km3 a–1. The fluxes from subglacial lakes range from 1.05 × 10–4 to 1.49 km3 a–1. If water movement between lakes is episodic rather than continuous, recurrence intervals range between ~1 and ~10 000 years for a 1 km3 flood, or from ~3 to ~50 000 years for a 5 km3 flood.


Vertical thermal structure of the subglacial Skaftá lakes in W-Vatnajökull, Iceland

Thorsteinn Thorsteinsson, Tómas Jóhannesson, Bergur Einarsson, Eric Gaidos

Corresponding author: Thorsteinn Thorsteinsson

Corresponding author e-mail: thor@vedur.is

The subglacial lakes beneath the eastern and western Skaftá cauldrons on W-Vatnajökull have released floodwaters into the river Skaftá at 1–3 year intervals since the start of regular monitoring in 1955. The jökulhlaups are of the rapidly rising type and the total volume released varies between 0.05 and 0.4 km3. The subglacial lakes have been accessed by hot-water drilling through the overlying 300 m thick ice shelves in 2006, 2007 and 2014. In June 2006, the temperature in the lake below the western cauldron was 4–5°C in the top 100 m of the water column, underlain by a distinct ~10 m deep slightly cooler water mass at 3.5–4°C. The water temperature was higher than the temperature of maximum water density at the pressure in the lake, and a temperature increase upward (except for a bottom few meters near the bed) indicated stable stratification at the location where the borehole was drilled. Vigorous heat transport through the subglacial water body was inferred from the near-uniform temperature of the main water body to within ±0.1°C, indicating the presence of circulation cells driven by upwelling of hot water above localized spots of geothermal activity and downwelling of cold water melted from the ice-shelf base. In June 2007, the lake below the eastern Skaftá cauldron was accessed with the hot-water drill, allowing deployment of a thermistor string containing 13 sensors spaced at 15 m intervals. The temperature in this lake was near 4°C throughout the 100 m deep water column and monitoring of the temperature profile (through a satellite link) revealed little change in the 3 month period during which the borehole cable survived. Bottom temperatures up to 13°C were recorded below the drilling site. The eastern cauldron was accessed again in March 2014. The temperature in the upper half of the subglacial lake was then below the temperature of maximum density at ~2.5–3°C, whereas the temperature in the lower half was slightly higher than the temperature of maximum density at ~3.5°C. The different thermal structure in 2014 compared with the earlier measurements indicates a less vigorous geothermal heat-driven convection in the lake and consequently less subglacial melting. This may explain why no jökulhlaup has occurred from this lake since June 2010, making the current interval between jökulhlaups (almost 5 years) the longest since the start of regular monitoring 60 years ago.


The view from the bed: insights into subglacial drainage systems from ice caving

Doug Benn, Jason Gulley, Ken Mankoff, Matthew Covington

Corresponding author: Doug Benn

Corresponding author e-mail: doug.benn@unis.no

During the last decade, we have been able to access and document several subglacial conduits using speleological techniques. Access has been possible in two situations: (1) via moulins to the upper end of conduits; and (2) via portals to shallow, sub-marginal conduits. Although access is limited to conduits <100 m below the surface, our observations shed considerable light on important but poorly constrained aspects of subglacial drainage. Tracer experiments conducted through shallow conduits with seasonally unchanging morphology demonstrate that evolution of due returns does not necessarily imply evolving drainage system configuration. Dispersivity and travel times can both decrease following melt onset due to the changing influence of bed roughness elements under increasing discharges, rather than the enlargement of conduits. Conduit floor mapping using structure from motion software has allowed bed roughness to be quantified in unprecedented detail. We have conducted detailed analyses of the influence of discharge and conduit size on effective roughness, and future work will use detailed 3-D conduit maps as realistic domains for hydrological modeling.


Spatial and temporal patterns of firn temperature evolution at Lomonosovfonna, Spitsbergen, during summer 2012–14

Sergey Marchenko, Veijo Pohjola, Rickard Pettersson, Ward van Pelt

Corresponding author: Sergey Marchenko

Corresponding author e-mail: sergey.marchenko@geo.uu.se

Temperature is one of the basic media parameters. Ability of snow and firn models to accurately describe the processes affecting it is of importance for mass-balance estimations, run-off predictions and studies of glacier dynamics. Due to the high temperature of water fusion snow and firn are extremely inhomogeneous and dynamic (mechanically and thermally) media. Evolution of snow and firn temperature is driven by the thermal gradient and is controlled by the effective thermal conductivity therein; it is also highly dependent on the release of the latent heat from refreezing liquid water. Most models describing vertical water flow in cold subsurface layers assume that it goes in a laterally uniform wetting front. This hypothesis strongly contradicts field and laboratory observations according to which under certain conditions preferential flow of water along vertical channels reaches much deeper than the position of the wetting front. This process has a strong effect on the pattern of temperature evolution. During spring and summer seasons 2012–14 evolution of snow and firn temperatures at 1200 m a.s.l. at Lomonosovfonna (Spitsbergen) was monitored using nine thermistor strings reaching to 12 m below the surface and arranged in a rectangular grid with 3 m spacing between thermistor strings. Components of the surface energy and mass balance were measured by an AWS in 2013 and 2014. In all three field campaigns detailed description of firn stratigraphy and measurements of density at the spot was done using shallow ice cores, borehole camera surveys and GPR profiling. Temperature measurements showed a high level of spatial inhomogenity of the pattern of temperature evolution in all three years. Temperate profile was registered by the end of the ablation season in all three years. However a firn model accounting for conductive heat transfer and influence of water refreezing on firn temperature and density was not able to produce such conditions. The inconsistency is attributed to the role of preferential water flow that delivers energy released as latent heat to deep firn layers. This stresses the importance of implementation of a routine describing the preferential flow of water through porous icy media in firn models widely used for mass, energy balance and climatic simulations in local and regional scales. Rates of water refreezing in pores deciphered from comparison of the measurements and a simple 1-D heat conductivity model are compared with repeated density measurements.


Periodic seismic events and stick–slip glacier motion at Svartisen, northern Norway

Miriam Jackson, Knut Christianson, Peter Moore, Paul Winberry, Sridhar Anandakrishnan, Neal Iverson, Peter Burkett

Corresponding author: Miriam Jackson

Corresponding author e-mail: mja@nve.no

Svartisen Subglacial Laboratory provides direct access to the bed of a glacier. This feature is almost unique in glaciology, but relating results from a small area of the glacier bed (only 200 m2) to those from the glacier surface (about 36 km2) is a major challenge. Further extending the results to other glaciers is even more problematic. However, a fixed seismic array installed in the bedrock tunnel underneath the glacier is able to exploit the subglacial facility, as seismic measurements made at the bed can be directly related to both local subglacial measurements as well as surface observations made over a larger area. We have collected nearly 3 years of continuous seismic and subglacial data, and initial results are presented here. Most previous seismic measurements on glaciers rely on instruments placed at the glacier surface, which are at least hundreds of metres from the glacier–bedrock interface, the region of the glacier that is of most interest. Additionally, other subglacial measurements, such as discharge or subglacial pressure, are generally not available at other glaciers. We highlight two different periods of subglacial hydrological activity. First, in spring 2008, periodic events were observed at the glacier bed, and lasted for up to several days at a time. We interpret these periodic events as stick–slip motion of the glacier. Periodic events were observed only early in the melt season when there is limited water at the glacier bed. As the basal hydrologic system becomes more efficient, the periodicity essentially disappears. There is no clear correlation with the magnitude or gradient of the subglacial discharge. Second, we present seismic and subglacial measurement data from the 2010 spring speed-up event. This begins with extensive surface crevassing associated with high-frequency seismicity followed by rapid drainage of surface meltwater to the glacier bed. After this initial signal, ground tilt derived from ultralong-period seismic signals was recorded, probably associated with bedrock deformation during the pressurization of the subglacial hydrology system. We interpret this signal as evidence of hydraulic jacking as the supply of meltwater is too great to be accommodated by the existing subglacial hydrologic systems, which then transitions into a more efficient channelized hydrology. Our results indicate use of seismic and subglacial records can aid interpretation of subglacial hydrology on other glaciers and ice sheets.


Negative mass balance and high discharge from glaciers in the central Andes of Chile

Andrés Rivera, Thomas Loriaux, Jorge Hernandez, Sebastián Cisternas, Claudio Bravo, Ryan Wilson, Tobias Kohoutek, Francisca Bown

Corresponding author: Andrés Rivera

Corresponding author e-mail: arivera@cecs.cl

The majority of Chilean glaciers have been retreating in recent decades in response to high-altitude atmospheric warming and a general trend of rainfall reductions. Smaller and lower-altitude glaciers have been more drastically affected, in some cases disappearing entirely and in others experiencing high fragmentation and area reductions. Recent dry conditions explained by the more frequent La Niña events since 2007 have reduced water levels in the regional dams to a minimum, stressing irrigation for agriculture activities. These conditions have forced the authorities to implement hydrological emergency plans in several basins, most of them having runoff partially contributed by glacier and snowmelting in spring and summer. Among the most affected areas by these dry conditions are the central Andes of Chile (33–34°S) where glaciers are contributing meltwater to the city of Santiago with more than 6.7 million inhabitants. Santiago is located in the Maipo River where a total area of 388 km2 of ice has been mapped in 2010, representing a total area lost of near 25% since 1955. This shrinkage trend has also been observed in nearby basins, such as Aconcagua (20% area lost since 1955) and Cachapoal (14% area lost since 1955). The main englacied area of the Maipo River is the upper Olivares basin, where Glaciares Olivares Alfa (3.9 km2 of ice in 2014) and Beta (8.8 km2 in 2014) have been monitored in detail since 2012, including mass, energy and hydrological balance. Among the main results of this monitoring program could be mentioned the high negative mass balance (up to –1.13 m w.e.) obtained in the last two hydrological years when a glaciological method (including snow pits, shallows cores and repeated measurements of near 20 stakes), as well as a geodetic method (including several helicopter-borne LiDAR surveys) were applied. We have also measured continuous summer discharge near the glacier outlets detecting very high instant maximum values reaching 7253 L s–1 on 13 February 2014 at the Olivares Alfa front and 16 956 L s–1 on 12 January 2014 at the combined basin of Glaciares Olivares Alfa and Beta. The runoff measured at the foot of Olivares Alfa in January and February has a mean of near 750 L s–1. This presentation will highlight recent results obtained during the implementation of this monitoring program and provide a discussion on the glacier meltwater contribution and the recent glacier changes in the region.


Sensitivities of ice volume changes on various basal sliding conditions

Tatsuru Sato

Corresponding author: Tatsuru Sato

Corresponding author e-mail: tsato@lowtem.hokudai.ac.jp

Basal sliding is an important component in glaicer/ice-sheet dynamics and mass balance. It depends on basal sediment properties, pressure, water pressure and thermal conditions. There are various approachs to implement basal sliding into the glacier/ice-sheet models. Sato and Greve (2012) used a sliding law which depends on pressure, water pressure and temperature. Each ice flow model uses different sliding conditions. It causes difficulties to design/analyze climate sensitivities of ice-sheet evolution using ice flow models. Sensivitiy experiments of each component of sliding, pressure, water pressure, thermal conditions, on basal sliding conditions are conducted. Changing thermal dependancy causes significant effects on ice volume changes, particularly in the Antarctic ice sheet experiments. It changes ice flow speed, ice flux from inland to coast significantly. It causes grounding-line migrations or changes of ice-sheet/ice-shelf area. Consequently, it changes sea-level equivalent ice volume of the ice sheet.


Wide expansion and increased glacial lake hazards during the past decades in central Asia

Xin Wang

Corresponding author: Xin Wang

Corresponding author e-mail: xinwang_hn@163.com

Glacial lake outburst flood (GLOF) is a serious hazard in high mountainous region. In central Asia, catastrophic risks of GLOFs are increasing in recent years since most glaciers experienced remarkable downwasting under a warming climate. In recent years, we conducted some research work on changes and glacial lake hazards in central Asia, including (1) glacial lake inventories were conducted on the basis of remote-sensing data in the Himalaya and Tien Shan, and the results show that glacial lake area was widely expanding at the average rate of 0.8 ± 0.1% a–1 during the past decades. In the Khan Tengri-Tomur mountains, the area of supraglacial lakes is positively correlated to the total precipitation in summer (July to September) but negatively to the mean air temperature over pre-summer (April to June). (2) 142 potentially dangerous moraine-dammed lakes (PDMDLs) and 60 PDMDLs were identified in the Chinese Himalaya and Tien Shan, respectively. 90 lakes in the Himalaya and 12 lakes in Tien Shan are rated as ‘high’ outburst probability status and should be considered for further detailed breach risk assessment. (3) Glacial lake water resource was evaluated using the water depth with the aid of geographical information techniques; an average ≈0.007 ± 0.002 Gt a–1 of glacier meltwater has been temporarily held in lakes over the past two decades, and the increase rate of total lake water volume was ≈0.016 ± 0.002 Gt a–1, representing ≈0.2 ± 0.0% of annual glacier mass lost in the Tien Shan. (4) To model the stabilities of moraine dams and evaluate the probabilities of moraine-dam breach in the Himalaya, a field survey station was set up in 2009 at an altitude of about 5700 m a.s.l. in the Chinese Himalaya. A moraine dam of Longbasaba lake has been investigated in terms of deformation, temperature, water content, heat flux, etc. since 2009. It is found that the annual ablation rate of the active layer of the dam is about 2.1 cm d–1, maximum ablation depths of the active layer of the dam in summer exceed 250 cm, the active layer of the dam is thickening at a rate of 1–2 cm a–1 since 2009, and the stability of the dam is decreasing under the current climate warming setting.


How swampy is the base of Recovery Glacier, Antarctica?

Angelika Humbert, Veit Helm, Thomas Kleiner, Daniel Steinhage

Corresponding author: Angelika Humbert

Corresponding author e-mail: angelika.humbert@awi.de

We use an indirect measurement to infer where the base of Recovery Glacier is wet or swampy, based on radio-echo sounding observations. This requires us to determine the absorption of the radar wave inside the ice as a function of chemical impurities and temperature. The resulting quantity is the basal reflection coefficient, which allows us to distinguish between dry, swampy and wet base. Here we intend to discuss how spatially variable the basal reflection coefficient is and to what extent this is driven by errors in the absorption or real spatial variation. The resulting distribution of wet and swampy base is discussed with respect to the location of the main trunk and a shear margin separating the ice stream from the surrounding area. Using a flux routing approach, we show the distribution of water layer beneath the glacier system and location of sinks, based on CryoSat-2 surface elevation and a new bed topography from our radio-echo sounding campaign.


The Fláajökull glacial landsystem: a geomorphological map of an active temperate glacier

Helga Lucia Bergsdóttir, Ívar Örn Benediktsson, Ólafur Ingólfsson, Sverrir Aðalsteinn Jónsson, Anders Schomacker

Corresponding author: Helga Lucia Bergsdóttir

Corresponding author e-mail: hlb1@hi.is

A new glacial geomorphological map of the Fláajökull forefield in southeast Iceland is presented. The map is based on aerial photographs recorded in 2006 and 1989, LiDAR data recorded in 2010 and a satellite image (Landsat 8) from 2013, as well as ground checking in the field. The landforms were manually registered with the ArcGIS 10 software and afterwards checked in the field. The mapped landforms were of subglacial, ice-marginal, supraglacial and glaciofluvial origin. They depict a typical active temperate glacier landsystem with a high glaciofluvial yield. What sets the Fláajökull forefield apart from most other modern non-surging outlet glaciers in Iceland is a small drumlin field in front of the present glacier snout, which was first exposed in the late 1990’s and has been increasingly exposed during the relatively rapid retreat over the past couple of decades. The landforms furthest away from the glacier were deposited during the glacier’s Little Ice Age maximum position in the late 19th century and the map gives a good overview of the glacier’s modern oscillations and landscape modification since the beginning of the 20th century. The large coverage of glaciofluvial landforms and the poorly preserved subglacial and ice-marginal landforms, especially on the western half of the forefield, along with historical documentation, bear witness to the hard struggle of local farmers to constrain the outwash streams and prevent them from migrating across the foreland, farmland and pastures south of the glacier during the glacial recession since the late 19th century.