Radar-sounding observations of basal water, sediments and geothermal heat flux and their implications for the past and future sea-level contribution of the Amundsen Sea sector of West Antarctica


Corresponding author: Dustin Schroeder

Corresponding author e-mail: dschroed@gmail.com

The basal morphology, lithology and hydrology of ice sheets and glaciers can exert strong even dominating control on their evolution, stability and sea-level contribution. However, the scales at which the physical processes and observable signatures of this control occur are typically smaller than the spatial resolutions achievable using ice-penetrating radar. Further, the strength of radar bed echo returns is a combination of both the material and geometric properties of the ice–bed interface as well as englacial attenuation. This ambiguity makes definitive assessment of basal conditions from echo strengths difficult. To address these challenges in interpreting geometric and material bed properties at glaciologically relevant scales, we apply a new algorithmic approach to measuring the radar scattering function of the ice–bed interface in terms of the relative contribution of angularly narrow specular energy and isotropically scattered diffuse energy. We compliment this specularity analysis with a coupled radar echo strength/subglacial water routing model to constrain the distribution of basal melt. We present the application of these techniques to a radar-sounding survey of Thwaites Glacier. We show that it can be used to assess the extent and geometry of distributed water across the catchment and detect the transition of the water system from distributed canals to concentrated channels. We also constrain the morphology of basal bedforms to infer the distribution of deformable sediments and crystalline bedrock. Finally, we compare our model of radar return strength and subglacial water routing with models of basal melting from ice flow to infer the distribution of geothermal heat flux and interpret its observed heterogeneity in the context of regional volcanism. These observed basal conditions provide new context for the past and potential evolution, stability and sea-level contribution of the rapidly changing Amundsen Sea sector of West Antarctica. We compare the contemporary configuration of Thwaites Glacier with that of the deglaciated Paleo Pine Island ice stream and the sedimentary record of its meltwater intensive retreat. We conclude that a transition in the basal hydrology of Paleo Pine Island was characteristic of its relatively rapid retreat across exposed bedrock on the inner continental shelf and that Thwaites Glacier may be currently configured for a retreat that is similar in character and pacing.


Ice sheets and sea-level projections in the Fifth Assessment Report of the IPCC


Corresponding author: Antony Payne

Corresponding author e-mail: a.j.payne@bristol.ac.uk

This presentation discusses the scientific basis on which the IPCC Fifth Assessment Report based its projections of the contribution of ice sheets to global sea level. The assessment includes both contributions from changing surface mass balance (SMB) and outflow (the flux of ice across the grounding line or leaving the grounded ice sheet as icebergs). Assessment of the former is based largely on the use of process-based models, primarily regional climate models coupled to detailed models of the energy and mass fluxes at the ice sheet’s surface. Process-based projections of changing outflow are still in their infancy so that the assessment relies on a range of literature including process-based modelling, statistical extrapolation and physical intuition. Projections of sea-level rise by 2100 are presented as a likely range, which equates to a likelihood of about two-thirds. The assessed likely range does not allow for a Marine Ice Sheet Instability. Particular attention is paid to this possibility and a further assessment is made of the additional contribution that might arise from it.


The effect of meltwater plumes on the melting of a vertical glacier face


Corresponding author: Satoshi Kimura

Corresponding author e-mail: skimura04@gmail.com

Ice-sheet meltwater is commonly discharged into ocean fjords from the bottom of deep fjord-terminating glaciers. This meltwater forms upwelling plumes in front of the glacier calving face. We simulate the meltwater plumes using a non-hydrostatic ocean model with a mesh that is unstructured in three dimensions and subgrid mixing calibrated by comparison to established plume theory. The presence of an ice face reduces the entrainment of sea water into the meltwater plumes, so the plumes remain attached to the ice front, in contrast to previous simple models. Ice melting increases with height above the discharge, also in contrast to some simple models, and we speculate that this ‘overcutting’ may contribute to the tendency of icebergs to topple inwards toward the ice face upon calving. When two channels are located in close proximity, the meltwater plumes can coalesce and form a single plume. Such merged meltwater plumes ascend faster but occupy a smaller fraction of the ice face, so that the melt rate averaged over the glacier decreases. The overall melt rate is found to increase with discharge flux only up to a critical value, which depends on the channel size, and decrease thereafter. For a given discharge flux, the geometry of the plume source also significantly affects the melting, with higher melt rates obtained for a shallower, wider source. We speculate that the melt rate per unit discharge decreases as the ice-sheet melting season progresses and the subglacial system becomes more channelized. The melt rate is not a simple function of the subglacial discharge flux, as assumed by many previous studies.


Glacier health status classes based on remote-sensing and GIS techniques in Alaknanda sub-basin, Ganga basin, India


Corresponding author: sarvesh palria

Corresponding author e-mail: tak_swati@gmail.com

Glaciers in the Himalaya and Trans Himalaya regions are located in highly rugged and inaccessible terrain. There are 32 392 glaciers spread over the Indus, Ganga and Brahmaputra basins covering an area of 71 182 km2 in the Himalaya and Trans Himalaya regions. The annual mass balance of a glacier is the measure of glacier ice mass at the end of the ablation period, and attributed as positive or negative depending on the change in glacier mass compared with the previous year. The observed globally averaged mass balance of glaciers and ice caps is negative and a matter of great concern. Scientific techniques are very much needed for the systematic parameter-based categorization of glaciers into relatively healthy or fragile glaciers, and thus assessing their health status within a sub-basin. Remote-sensing techniques with sensors operating in different parts of the optical spectrum and having high resolution have completely changed the present scenario for glacier-related studies. The data can be easily interpreted and extrapolated to other areas with much higher confidence levels. Thus, in a minimal time frame, large sub-basin-scale areas can be studied simultaneously. Analysis of IRS-P6 AWiFS and SRTM DEM data along with ancillary data is carried out using image-processing and GIS tools for deriving eight significant glacier parameters for assessing the relative health status of glaciers in the Alaknanda sub-basin, Ganga basin, India. The Alaknanda sub-basin spread over an area of 3552 km2 has 253 glaciers covering 39% percent of the glaciated area. The accumulation area is 787.07 km2, ablation area debris is 349.79 km2 and ablation area ice is 193.07 km2. The estimated total volume of glacier ice in the Alaknanda sub-basin is 183.85 km3. The glacier morphology map and data sheet prepared based on IRS-P6 AWiFS data of 2004–07, SRTM DEM and other ancillary data form the main basis of data for deriving the required parameters for the study. The parameters considered for this study are glacier area, length, orientation, elevation and ice exposed.


Contrasting dynamics and sensitivity of the Amundsen Sea ice streams

Isabel NIAS, Stephen CORNFORD, Tamsin EDWARDS, Tony PAYNE

Corresponding author: Isabel Nias

Corresponding author e-mail: isabel.nias@bristol.ac.uk

Ice loss from Antarctica is centred on an area of West Antarctica known as the Amundsen Sea Embayment (ASE). The stability of this area is a key control on future global sea level. Within the ASE, loss appears to be primarily associated with ice streams draining the area, including Pine Island and Thwaites Glaciers. The majority of research that attempts to understand the mechanisms responsible for this ice loss is based on modelling and satellite studies of Pine Island Glacier (PIG). From these studies a mechanism for accelerated flow and dynamic thinning of PIG has been identified whereby relatively warm Circumpolar Deep Water upwells onto the continental shelf and migrates under the ice shelves, causing increased melt and retreat of the grounding line. By comparison, there has been relatively little model-based research carried out on Thwaites Glacier (TG) and the cause of the thinning observed in the glacier interior is less clear. We seek to understand the differences in sensitivity to various parameters between PIG and TG using an advanced numerical model. BISICLES is a vertically integrated high-order ice flow model with adaptive mesh refinement (AMR). AMR provides a means of accurately modelling grounding-line migration with sub-km resolution, while avoiding the computational demands of a uniformly fine resolution. The position of the grounding line is important to ice-stream dynamics and stability, particularly on upward-sloping bedrock, typical of the ASE. Using BISICLES, we ran a perturbed model ensemble for PIG and TG. Latin hypercube sampling was used to generate sets of parameter values for a range of physical conditions, including ice rheology, basal sliding and bed topography. We present probability density functions of the likelihood of sea-level contributions from PIG and TG under the same oceanic forcing. Initial results suggest that these probability density functions are very different.


On the climate-geometry imbalance of Vadret da Morteratsch (Switzerland) and its response time: insights from numerical 3-D flow simulations


Corresponding author: Harry Zekollari

Corresponding author e-mail: harry.zekollari@vub.ac.be

When a glacier is subject to a change in mass balance it needs a certain amount of time to adjust its length and volume. In the literature the response time is referred to as the time a glacier takes to complete most of its adjustment to a change in mass balance, but such a broad definition is not always satisfactory. A better insight in the response time of glaciers and the main controlling factors is however crucial to improve future projections, as the evolution of glaciers in the coming decades is largely determined by their past evolution and present-day imbalance. Here we investigate the climate-geometry imbalance and the response time of Vadret da Morteratsch (Engadine, Switzerland) by using a 3-D higher-order ice flow model. In earlier studies the glacier flow model was calibrated with observed surface velocities, and used to simulate the glacier evolution since 1865 driven by a 2-D energy surface mass-balance model. This modelling relies on an extensive observational dataset collected on the glacier over the last 13 years. The present-day imbalance between glacier geometry and climate is analysed by performing steady-state simulations. We investigate the retreat under present-day climate and the forcing needed to maintain the present-day glacier length and volume. Subsequently we do this for the previous decades to understand how the climate-geometry imbalance changed over time. In our analysis of response times, we impose idealized step changes in mass balance and quantify the time needed to reach a new equilibrium. We analyse the factors that influence the length and volume response time, such as glacier size, glacier thickness (flow parameters), the magnitude and spatial distribution of the mass-balance forcing and the differences occurring between glacier advance and retreat. Subsequently, we compare our results with simpler analytical approaches from the literature and discuss the applicability of each method to characterize the response time of Vadret da Morteratsch.


Correlating the surface mass balance on Morteratsch glacier (Switzerland) with data from nearby meteorological stations


Corresponding author: Philippe Huybrechts

Corresponding author e-mail: philippe.huybrechts@vub.ac.be

Over the past decades strong negative surface mass balances have lead to a major retreat of glaciers in the European Alps. This trend in mass balance has been widely described in the literature. It has been modelled with simple approaches (such as positive degree-day models) and more complex methods (for instance through 2-D surface energy-balance models). A detailed approach is important to better understand the surface mass balance and its underlying processes, but is often not needed to understand its interannual variation. Here we show this based on an extensive dataset of annual surface mass-balance measurements collected in the ablation zone of Morteratsch glacier (Engadine, Switzerland) over the last 13 years. We correlate these measurements with data from nearby meteorological stations and point out that most of the variability in annual surface mass balance can be explained by only a few meteorological parameters. Subsequently we compare this to results from a simple 2-D surface energy-balance model. Finally we use our simple correlation between meteorological measurements and the surface mass balance to explain the more positive 2012–13 surface mass balance over Morteratsch glacier.


New concept on the influence of glacier runoff to ocean level


Corresponding author: Vladimir Konovalov

Corresponding author e-mail: vladgeo@gmail.com

As shown by the Forth Report of the IPCC, the uncertainty in evaluating the total contribution of land ice to change in ocean level exceeds both any individual contribution and the accuracy of the ocean-level measurements. Many reasons for this situation were revealed recently in Canada, England, the USA and Switzerland after analysis of long-term data of mass-balance measurements on ‘reference’ or representative glaciers. The proposed new concept presents a completely different approach for estimating the contribution of glacier runoff to the change in ocean level, instead of using limited samples of mass-balance data. Estimations of the influence of glacier runoff to the level of the World Ocean were obtained from examples of closed (not drained) river basins of Eurasia and basins connected to the ocean. In the first case, only the seasonal volume of evaporation from the surface of melted ice and old firn could be a potential factor of glacier influence on the level of the World Ocean. In particular, this volume during 1935–94 on an area of 5507 km2 was 2.35% from glacier runoff of 239.6 km3 in the upstream of the Amudarya river basin. In the second case, the extremely small contribution (0.5–1.9%) of glacial water to continental river runoff and the absence of a relationship between the ocean level and annual river runoff within the pan-Arctic and Eurasia show the insignificancy of glacier runoff in the water balance of the World Ocean.


Re-evaluation of the influence of glacier runoff to the ocean level


Corresponding author: Vladimir Konovalov

Corresponding author e-mail: vladgeo-exp@yandex.ru

As shown in the Forth Report of the IPCC, the uncertainty in evaluating the total contribution of land ice to the change in ocean level exceeds both any individual contribution and the accuracy of the ocean-level measurements. Many reasons for this situation were revealed recently in Canada, England, the USA and Switzerland after analysis of long-term data of mass-balance measurements on ‘reference’ or representative glaciers. All continental glaciers, ice caps and ice sheets should be hydrologically differentiated as: (1) marine-terminating, (2) not marine-terminating but connected to the World Ocean through river flow, and (3) not marine-terminating and located in a closed river basin (not drained to the World Ocean). Estimations of the influence of glacier runoff to the level of the World Ocean were obtained from examples of closed (not drained) river basins of Eurasia and basins connected to the ocean. In the first case, only the seasonal volume of evaporation from the surface of melted ice and old firn could be a potential factor of glacier influence on the level of the World Ocean. In particular, this volume during 1935–94 on an area of 5507 km2 was equal to 2.35% from glacier runoff of 239.6 km3 in the upstream of the Amudarya river basin. In the second case, the extremely small contribution (0.5–1.9%) of glacial water to continental river runoff and the absence of a relationship between the ocean level and annual river runoff within the pan-Arctic and Eurasia show the insignificancy of glacier runoff in the water balance of the World Ocean. Quality estimations of calculating volumes of rainfall, evaporation and glacial runoff for the river basin Vakhsh as a whole were obtained by comparing the measured and calculated volume of annual runoff. The difference between the comparable amounts of flow was 5.1%.


Ice storage in Nordenskiöld Land glaciers, Svalbard, estimated using radar data and balance-dynamical models


Corresponding author: Yuri Macheret

Corresponding author e-mail: macheret2011@yandex.ru

Data on ice thickhess, ice volume, area and surface elevation were used to estimate the ice storage in glaciers in Nordenskiöld Land glaciers, Svalbard, using two methods: (1) relationships between glacier area and volume inferred from topographic maps (1936, 1990), satellite Landsat and ASTER images (2006–13) and ground-based radar surveys (1999, 2010–13) at 12 glaciers; and (2) balance-dynamical models for these and other smaller glaciers in this region. We also investigated a the possibility of using these approaches for total ice storage estimations in glaciers in other regions of mountain glaciation in Svalbard and their contribution to sea-level rise under climate warming.


Comparing 60 years of evolution of a debris-covered glacier (Glacier Noir) and a clean-ice glacier (Glacier Blanc) in ‘Les Ecrins’ National Park, France


Corresponding author: Pierre Lardeux

Corresponding author e-mail: pfl4@aber.ac.uk

Debris-covered glaciers respond differently to climate change than clean-ice glaciers. However, we do not clearly understand exactly how climate change impacts the mass balance, hydrology or debris transport of these debris-covered glaciers. To investigate how supraglacial debris modifies glacier behaviour, we compare the glaciological and geomorphological evolution of a debris-covered glacier, Glacier Noir, and an adjacent clean-ice glacier, Glacier Blanc. These two glaciers are located in the same catchment of the St Pierre Creek in Les Ecrins National Park in the French Alps, have a similar orientation and geometry, and as a result probably experience similar climatic conditions. We will use Glacier Blanc as a reference case, and comparing the behaviour of the two glaciers allows us to specifically test the impact of debris on the response of Glacier Noir to climate change during the last century. Using historical aerial photography from the French National Institute of Geographic and Forestry Information (IGN) from 1952 to 2012, we have measured the length and surface of both glaciers and tracked surface features to derive velocity fields from boulders and meltwater ponds on Glacier Noir and crevasses on Glacier Blanc. We have used a digital elevation model (DEM) provided by IGN to analyse the hypsometry of the two glaciers and the snowline altitude. Finally, by comparing DEMs from different dates, we have estimated Glacier Noir’s surface mass-balance change. This remote-sensing comparison of Glacier Noir and Glacier Blanc is the first step in a larger study that will include data collection from the field site (water discharge and bedrock topography) and the application of a 2-D higher-order glacier model with a dynamic debris layer. This modelling will be forced by a reanalysis of climatic data which considers, among other factors, the 1–3°C rise in air temperature in the Alps since 1958. This work will help us better understand the recent evolution (over the last 60 years) of a debris-covered glacier and predict how it will respond to future climate change.


Radiostratigraphy of the Greenland ice sheet


Corresponding author: Joseph MacGregor

Corresponding author e-mail: joemac@ig.utexas.edu

We have produced a dated radiostratigraphy for the whole of the Greenland ice sheet (GrIS) from two decades of airborne radar-sounding surveys performed by the University of Kansas. This radiostratigraphy reveals a wealth of new information regarding this ice sheet’s internal structure and history. For example, disrupted radiostratigraphy is often located near the onset of the largest outlet glaciers, suggesting a strong coupling between the initiation of this faster flow and anomalous basal processes in the ice-sheet interior. Ice-flow modeling constrained by this radiostratigraphy shows that the Holocene-averaged pattern of surface accumulation is similar to the modern pattern, but that during the Holocene surface accumulation was substantially higher in the interior and lower toward the ice-sheet margins, particularly north of Camp Century. The pattern of basal melt is strongly modulated by surface accumulation, suggesting that geothermal flux beneath the GrIS is low except in the vicinity of the Northeast Greenland Ice Stream. The Holocene-averaged flow pattern of the GrIS differs from its present pattern in a manner consistent with ice-core and geologic histories of its deglaciation. This ice-sheet-wide radiostratigraphy is a new and powerful constraint on the dynamics of the GrIS, and it should be used to validate and improve the next generation of ice-sheet models.


Temperature and velocity profiles at the PARCA stakes inferred by transient thermomechanically coupled flowline modeling

Aleah SOMMERS, Harihar RAJARAM, William COLGAN

Corresponding author: Aleah Sommers

Corresponding author e-mail: aleah.sommers@colorado.edu

Most recent changes in the surface mass balance and ice dynamics of the Greenland ice sheet have been restricted to elevations below 2000 m. Substantial computational efficiency can be gained by limiting numerical modeling efforts to this lower-elevation periphery, where changes in ice-sheet form and flow are most pronounced, rather than modeling the entire ice sheet from the main divide to the margin. Accurately modeling the lower elevations with this approach is dependent on prescribing accurate velocity and temperature profiles at the upstream boundary. The Program for Arctic Regional Climate Assessment (PARCA) provides reliable surface velocity data at 161 locations approximately circumscribing the 2000 m elevation contour of the Greenland ice sheet. The Ice2Sea project improved estimates (and constrained uncertainty) of ice thickness at the PARCA stake locations. Without corresponding velocity and temperature profiles, however, these data alone are insufficient to serve as upstream boundary conditions for lower-elevation thermomechanical modeling. Using a fully transient, thermomechanically coupled, two-dimensional flowline model and a range of boundary conditions, we assess vertical velocity and temperature profiles at each of the PARCA stake locations. We employ a forward model selection Monte Carlo approach to match surface velocity and ice thickness observations (within specified uncertainty), and thereby quantify the uncertainty associated with inferred velocity and temperature distributions. This selection exercise provides constraints on the surface forcing and geothermal heat flux required to simultaneously honor velocity and elevation observations. While our specific motivation driving this exercise is to determine upstream boundary conditions for low-elevation flow models, we believe numerous groups with diverse interests may take advantage of such temperature and velocity profiles at the PARCA stakes. For example, robust thermomechanically consistent temperature and velocity profiles that incorporate the temperature-dependence of the flow law parameter will enhance the accuracy of ice flux estimates at each stake location. We anticipate that the spatially widespread velocity and temperature profiles generated through this work will serve as a useful data product for the ice-sheet community.


Antarctic ice sheet mass balance measured by GRACE gravity satellite and the uncertainties

Chunchun GAO, Yang LU, Chuandong ZHU

Corresponding author: Yang Lu

Corresponding author e-mail: luyang@whigg.ac.cn

The Gravity Recovery and Climate Experiment (GRACE) mission opened a new era in gravimetry for estimating the mass balance of the Antarctic ice sheet since 2002. Using Release 5.0 (RL05) GRACE monthly gravity fields for January 2003 through April 2013 from CSR (118 total), temporal and spatial variation of Antarctic ice sheet mass is recovered in two ways: the optimizing averaging kernel method and the two-step filter method. The results reveal that the mass of the ice sheet has decreased significantly for the past 10 years, the changes of –131±55, –97±48 and –43±35 Gt a–1 for three GIA models (GW13, IJ05, W12a), with an acceleration of –12±8 Gt a–2, and most of this mass loss came from the southeast Pacific sector of West Antarctica and the Antarctic Peninsula. In addition, we analyze the uncertainties in GRACE estimates of ice-sheet mass balance with emphasis, indicating that the largest sources of error in Antarctic ice-sheet mass balance are GIA correction. Comparison of the results from the two different methods shows that when the same time span and a consistent set of corrections are used, different GRACE post-processing methods produce consistent ice mass-balance estimates.


Principles of continuum damage mechanics applied to problems in glaciology

Christopher BORSTAD, Eric RIGNOT, Eric LAROUR

Corresponding author: Christopher Borstad

Corresponding author e-mail: cborstad@jpl.nasa.gov

Damage mechanics is a theoretical framework for representing the influence of discrete fractures in a material in a smeared or average sense, thus allowing a continuum representation of the fractured material physics. Damage mechanics has been successfully applied to model the initiation and propagation of fractures in a wide variety of natural and engineering materials, and is applicable for diffuse and heterogeneous microcracking up to macroscopic fracture initiation and propagation. Here, we outline the key assumptions and requirements for formulating a continuum damage model in a glaciological setting, and limitations to the scope over which damage models can be applied. We first classify different fracture processes according to the timescale(s) over which they operate, which dictates whether an elastic or viscous damage model is appropriate. We then proceed with a description of various definitions of damage as a state variable, the physical interpretation of which depends on the choice of linear mapping schemes between the actual and effective material volumes. We then discuss strategies for formulating a nonlocal damage model in order to limit the scale of strain localization in the model. This requires the specification of a nonlocal averaging length scale, and this physical scale varies depending on the type of fracture process of interest, from the kilometer scale for ice-shelf rift propagation to the meter scale for individual crevasses. Depending on the the fracture process of interest, this material length scale defines a limit on the maximum element or grid size in the model in order to satisfy mesh objectivity. Finally, we demonstrate that a properly formulated nonlocal damage model is capable of representing the initiation of a fracture from a smooth boundary without the requirement for a pre-existing flaw or stress concentration.


Subglacial hydrology and the formation of ice streams

Teresa KYRKE-SMITH, Richard KATZ, Andrew FOWLER

Corresponding author: Teresa Kyrke-Smith

Corresponding author e-mail: teresa.kyrke-smith@earth.ox.ac.uk

We present a model of subglacial water flow below ice sheets, and particularly below ice streams. This hydrologic model is coupled to a vertically integrated model for ice flow that describes vertical, lateral and longitudinal stresses. We show that under some conditions, this coupled system gives rise to ice streams by instability of the internal dynamics. The base-level water flow is fed by subglacial melting and is presumed to take the form of a Weertman–Creyts–Schoof rough-bedded film, in which the ice is supported by larger clasts, but there is a millimetric water film that submerges the smaller particles. A model for the film is given by two coupled partial differential equations, representing mass conservation of water and ice closure. We assume there is no sediment transport, and solve for water film depth and effective pressure. If there is a sufficiently small amount of meltwater produced (e.g. if ice flux is low), the distributed film is stable, while for larger amounts of melt, the ice–water system can become unstable, and ice streams form spontaneously as a consequence. We show that this can be explained as a result of a multivalued flux law, which arises from a simplified one-dimensional analysis of the coupled model. We further explore the range of parameters in the model that allow ice-stream formation. By implementing varying bed topography we are also able to provide insight into what levels of topographical constraint result in ice-stream positioning being a direct consequence of topography.


Winter speed-up of ice flow at quiescent surge-type glaciers in Yukon, Canada

Takahiro ABE, Masato FURUYA

Corresponding author: Masato Furuya

Corresponding author e-mail: furuya@mail.sci.hokudai.ac.jp

Water at the base of glaciers and ice sheets plays a primary role in generating short-term ice surface velocity changes. High-rate velocity measurements over wide areas are thus useful to better understand the spatial-temporal distribution of water inside the ice body. Still lacking, however, are winter velocity measurements in the middle to upstream of mountain glaciers. Here we examined spatial-temporal changes in the ice velocity of surge-type glaciers near the border of Alaska and Yukon, where significant contributions of its retreat to the possible sea-level rise are predicted, and ongoing interactions between glacial erosion and landscape evolution are also suggested. We found significant upstream accelerations from fall to winter, regardless of surging episodes. Moreover, whereas the summer speed-up was observed downstream, the winter speed-up propagated from upstream to down-glacier. Given the absence of upstream surface meltwater input in winter combined with an earlier observation of vertical surface motions, we support the hypothesis of englacial water storages that promote basal sliding through increased water pressure as winter approaches. Our findings have implications for more realistic modeling of glacial hydrology and its link to subglacial erosion.


Sensitivity of Austfonna ice cap transient behavior to model physics and basal boundary conditions

Yongmei GONG, Rupert GLADSTONE, Stephen CORNFORD, Martina SCHÄFER, Thomas ZWINGER, John MOORE, Thorben DUNSE, Ruth MOTTRAM

Corresponding author: Yongmei Gong

Corresponding author e-mail: yongmei.gong@ulapland.fi

Austfonna is one of the largest ice caps in the Arctic region. Its future dynamics and ice mass loss concern both the regional isostatic and global eustatic sea-level change. Observations show that one of its outlet glaciers (Basin 3) has accelerated dramatically since 1995. A proper numerical representation of the dynamics of Basin 3 in ice-sheet models is needed for a more reliable future projection. We use two ice-sheet models to carry out the ice dynamic simulation on Austfonna. Elmer/Ice dynamic model implements the finite-element method to solve full Stokes problem across the whole domain, while the BISICLES model implements a vertically integrated approximation. Similar data assimilation techniques are implemented in two models which minimize the mismatch between the magnitudes of modeled and observed velocity in order to infer basal friction coefficient field for basal resistance calculation. The basal friction coefficient is referred in a linear sliding law in both models for inversion. Both models produce a good match between the modeled and observed velocity and similar distribution of the basal friction coefficient. But the differences of the magnitude between the two basal friction coefficient fields can be larger than 2 orders of magnitude in some regions. Sensitive tests were carried out by implementing the Weertman sliding law with two commonly used exponents, m (m = 1 and m = 1/3), as well as driving the models with SMB anomalies from different regional climate models and some idealized SMB anomalies inputs.


Micrometeorological conditions on Chhota Shigri glacier, Himachal Pradesh, Western Himalaya, India, revealed by in situ meteorological data

Arindan MANDAL, Mohd Farooq AZAM, Patrick WAGNON, Alagappan RAMANATHAN, Parmanand SHARMA

Corresponding author: Alagappan Ramanathan

Corresponding author e-mail: alrjnu@gmail.com

Very little uninterrupted reliable meteorological data are available for high mountainous regions, especially over the Himalaya. Here we are presenting data from an automatic weather station on the lateral moraine of Chhota Shigri glacier, Himachal Pradesh, Western Himalaya (benchmark glacier in the western Himalaya, India) over four hydrological years from 1 October 2009 to 30 September 2013. Time series of all the measured parameters are analyzed to characterize meteorological conditions close to the ELA at 4850 m a.s.l. The measured meteorological data include air temperature, wind speed and direction, relative humidity, short- and longwave radiation. The result during the period shows air temperature and relative humidity variations are large enough to characterize seasonal regimes. The coldest month is January with a mean air temperature of –15.8°C, and the hottest month is August with a mean air temperature of 4.3°C. It has been observed that mean air temperature is positive during the monsoonal months only, during this period the glacier experienced maximum ablation. Relative humidity is moderate, defining the circulation system (viz. Indian Summer Monsoon and Mid-Latitude Westerly) operating over Chhota Shigri glacier. Shortwave incoming radiation controls the air temperature as revealed through the analysis. Seasonal variations in shortwave and longwave radiation are quite noticeable, characterizing the clear and overcast sky. Average wind speed is quite high, and two distinct wind directions are controlling the whole wind pattern. Two almost opposite wind directions observed during the winter half years indicate the flow of katabatic winds. Scatter plots between wind speed and air temperature depict more or less absence of katabatic flow of wind in summer half years. This type of in situ meteorological data are sparse, and high-altitude areas of the Lahaul and Spiti region (Chhota Shigri glacier) are quite helpful to understand the regional meteorology and the local climatic behavior.


Calving observations at high resolution from a Greenland tidewater glacier

Christopher WILLIAMS, Adrian LUCKMAN, Doug BENN, Tavi MURRAY

Corresponding author: Christopher Williams

Corresponding author e-mail: c.n.williams@swansea.ac.uk

Calving constitutes a major control on the overall mass balance at an ice-sheet scale and yet a fully comprehensive understanding – a solution to the ‘calving problem’ – remains elusive. Approximately 50% of mass loss from Greenland is due to glacier calving at the margins of tidewater glaciers. The impacts of increasing calving rates are significant for sea-level rise, and the volume and trajectory of icebergs have implications in terms of shipping and marine ecosystems. The calving criterion as established by Benn and others (2007) related crevasse depth directly to calving events, stating that calving occurs when crevasse depth equals ice height above sea level. The presence of water within a crevasse can cause it to further deepen and the ‘Benn criterion’ has become a key parameterization in many calving models. There are however limited field data available to support this mechanism. Here we present results of the analysis of high-resolution survey data from the calving margin of the Kangia Nunata Sermia glacier in West Greenland for which in July 2013 we acquired repeat lidar data suitable for producing a DEM at a spatial resolution of 0.5 m, and hyperspectral data at 3 m resolution. Such high resolution is vital to be able to resolve small calving events and crevasse development between 20 and 22 July. These data provided opportunities for the application of feature tracking, enabling calculation of surface velocities and strain fields, assessment and quantification of crevasse evolution, as well as identification of both presence and absence of water within crevasses. These data show flow speeds at the terminus of Kangia Nunata Sermia were ~25 m d–1 with a calving rate up to 17 m d–1. Crevasses were found to expand up to 2 m and there is clear development of a plume at the margin, which by the end of the second day of observation had an area of ~0.5 km2. The observations and analyses at the tidewater margin of Kangia Nunata Sermia provide novel insights into the calving process and evolution of the margin. The outputs of this research are to be combined with the international CRIOS (Calving Rates and Impact on Sea Level) research project which runs until 2016. This project will provide new, robust and efficient iceberg-calving models that will substantially improve the accuracy of Earth system models used to forecast future environmental change.


Direct observations of evolving subglacial drainage beneath the Greenland ice sheet

Lauren ANDREWS, Ginny CATANIA, Matthew HOFFMAN, Jason GULLEY, Martin LÜTHI, Claudia RYSER, Robert HAWLEY, Thomas NEUMANN

Corresponding author: Lauren Andrews

Corresponding author e-mail: landrews@ig.utexas.edu

Seasonal acceleration of the Greenland ice sheet (GrIS) is linked to increased surface melting, but it remains uncertain whether changes in the duration and magnitude of the melt season will impact GrIS dynamics and its long-term stability. We use the first coincident measurements of moulin and borehole hydraulic head and ice-surface velocity from the western ablation zone of the GrIS to explore spatio-temporal variability in the subglacial hydrological system and to understand how these changes impact ice velocity during the 2011 and 2012 melt seasons. Ice velocity is well correlated with moulin hydraulic head but out of phase with that of nearby boreholes. We infer that these moulins form part of an efficient, channelized component of the subglacial hydrologic system, which exerts the primary control on diurnal and multi-day changes in ice velocity. However, moulin hydraulic head does not display a seasonally decreasing trend thus cannot explain the observed seasonal ice velocity response. Our boreholes monitor an unconnected component of the subglacial system that has little to no connectivity with the moulin-driven channels. Changes in borehole hydraulic head are likely due to passive changes in basal sediment pressurization or cavity volume induced by basal sliding but exhibit characteristics of seasonal change. We hypothesize that capacity and efficiency changes within the non-channelized regions of the subglacial hydrologic system drive observed seasonal trends in ice velocity.


How to build a better bed: assimilating repeat surface observations to reconstruct the basal topography of Helheim Glacier, Greenland

Martin O’LEARY, Tavi MURRAY, Adrian LUCKMAN, Suzanne BEVAN

Corresponding author: Martin O’Leary

Corresponding author e-mail: m.e.w.oleary@gmail.com

Modern glaciology is reliant on the use of numerical models, both for predictive purposes and as a testbed for theories on glacial processes. However, these models are only as good as the input data that are used. In particular, basal topography is often a major unknown. Most observations come from airborne radar, and improvements are usually slow, costly and error-prone. At the same time, model results are highly sensitive to the details of the bed map, and this results in large uncertainties in model forecasts, as well as difficulties in testing theories about calving, sliding and other glacial processes. Flux conservation techniques have attempted to alleviate this problem by combining partial measurements of ice thickness with more easily obtained velocity data. Adjoint-based methods have shown great promise here, and the resulting bed maps are better both in terms of accuracy and the stability of numerical flow models using them. However, existing approaches are still very limited in the amount of data that is used. We demonstrate a new approach which allows for assimilation of heterogeneous observations from multiple time periods. By assimilating observations of velocity and surface elevation over multiple time periods, we can substantially reduce the effects of observational error, and increase the spatial coverage of the resulting bed map. We apply this new method to one of Greenland’s most-studied outlet glaciers, Helheim. Using a series of high-resolution velocity maps and digital elevation models produced from TanDEM-X satellite radar imagery, and incorporating radar flight lines from Operation IceBridge, we produce a new map of the bed beneath Helheim. We show that this approach is considerably more accurate than the most widely used method for interpolating bed topography, kriging, and that it outperforms previous adjoint-based methods in cross-validation tests.


Horizontal stress transfer controls mass transport in the Greenland ablation area

Claudia RYSER, Martin P. LÜTHI, Lauren C. ANDREWS, Ginny CATANIA, Martin FUNK, Robert L. HAWLEY, Matthew J. HOFFMAN, Thomas A. NEUMANN

Corresponding author: Martin P. Lüthi

Corresponding author e-mail: martin.luethi@geo.uzh.ch

We present measurements of surface velocity, ice deformation and subglacial water pressure on the Greenland ice sheet (GrIS) that contradict common notions of the influence of meltwater supply to the base to ice motion. We find 50–70% contribution of basal motion to observed surface velocity in winter, with episodic acceleration in summer. Further, we observe anti-phase behavior of water pressure and surface velocity, and delayed periodic and episodic variations of borehole tilt observed with sensors at different depths throughout the ice column. With the help of a full-Stokes ice flow model, all observations are explained with horizontal stress transfer from slippery to sticky patches of the bed, enhanced by the polythermal structure of the ice sheet. Ice motion in a caterpillar-like fashion during summer is caused by patches of different basal slipperiness, and activation of sticky patches. Variations of basal slipperiness induce characteristic patterns of ice deformation variability that explain the observed behavior. We conclude that ice flow in the ablation zone of the GrIS is controlled by activation of basal patches by varying slipperiness in the course of a melt season, leading to caterpillar-like ice motion superposed on the classical shear deformation. Such behavior illustrates the necessity of using full-Stokes ice flow models for large parts of the GrIS.


Effects of ice-shelf disintegration on tributary glaciers at the Antarctic Peninsula

Thorsten SEEHAUS, Sebastián MARINSEK, Pedro SKVARCA, Daniel STEINHAGE, Matthias BRAUN

Corresponding author: Matthias Braun

Corresponding author e-mail: Matthias.Braun@geographie.uni-erlangen.de

The climatic conditions at the Antarctic Peninsula have shown significant changes during the last century. Numerous ice shelves (e.g. Larsen A and B, Wordie) showed large break-up events or disintegrated completely. This led to an increased flow speed and substantial surface lowering of tributary glaciers due to a reduced buttressing effect on these glaciers. Latest estimates show a positive contribution of Antarctic Peninsula ice to sea-level rise. However, the quantification of ice mass loss along the Antarctic Peninsula is still inadequate due to the insufficient temporal coverage of ice speeds as well as ice thickness and surface mass-balance data. A detailed study on the Dinsmoore–Bombardier–Edgeworth glacier system, which were tributary glaciers of the Larsen A ice shelf prior to its disintegration in 1995, is presented. Changes in ice flow speed are determined by analyzing time series of SAR satellite data (ERS-1/2, Envisat, ALOS, RADARSAT-1, TerraSAR-X) over the last 20 years. The glacier surface velocities are computed by using feature tracking on applicable pairs of SAR images. Results show an increase from 0.9 m d–1 in 1996 up to 3.3 m d–1 in 2007. Afterwards the surface velocity decreased to 1.6 m d–1 in 2010. Since then it is almost constant. Furthermore the calving-front position shifted nearly synchronously with the flow speed. Digital elevation models generated from high-resolution TanDEM-X data (2011–present), previous ASTER and SPOT elevation models (2003, 2006, 2008) as well as spaceborne and airborne (November 2011, November 2013) laser scanning are used to calculate surface elevation changes. The estimates indicate a median surface lowering of ~24 m on Bombardier and Edgeworth Glaciers between 2003 and 2013. Finally, ice discharge estimates are calculated by combining these datasets with airborne ground-penetrating radar data (November 2013, NASA Icebridge November 2011). These results help to understand the reaction of tributary glaciers to ice-shelf disintegration and support imbalance calculations along the Antarctic Peninsula.


Examining the role of submarine melting in the retreat of Greenland’s tidewater glaciers


Corresponding author: Tom Cowton

Corresponding author e-mail: t.cowton@sheffield.ac.uk

Greenland’s tidewater glaciers have undergone dramatic changes in frontal position in recent decades. In many cases there have been regional similarities in the timing of these episodes of advance and retreat, suggesting a common environmental forcing. In comparison, variability in the frontal position of land-terminating glaciers has been modest, illustrating the important role played by the ocean in driving or facilitating rapid changes in ice dynamics. Variability in submarine melt rate at the ice front may be a key factor in the stability of tidewater glaciers. Controls on submarine melt rate have been investigated both theoretically and using numerical ocean models. These studies suggest that submarine melt rate should increase both with ocean temperature and with the discharge of freshwater from the glacier, which stimulates stronger upwelling along the ice front. Testing these theories has however proven problematic due to the inaccessibility of tidewater margins. Here, reanalysis data are used in conjunction with theoretical modelling of ice marginal plume dynamics to derive approximate submarine melt rates at a variety of Greenlandic tidewater glaciers over a 10 year period. To examine the significance of the predicted changes in submarine melt rate as a control on the stability of these glaciers, these values are compared with simultaneous records of glacier front position, collected using MODIS imagery. Through this, we seek to assess whether modelled submarine melting can explain the recent variability in tidewater glacier front positions and, if so, whether these changes are driven primarily by changes in ocean temperature, meltwater runoff or a combination of the two.


Parameterization of basal hydrology near grounding lines in a one-dimensional ice-sheet model


Corresponding author: Gunter Leguy

Corresponding author e-mail: gunter@nmt.edu

Ice sheets and ice shelves are linked by the transition zone, the region where the grounded ice lifts off the bedrock and begins to float. Adequate resolution of the transition zone is necessary for numerically accurate ice-sheet–ice-shelf simulations. The required resolution depends on how the basal physics is parameterized. We propose a new simple parameterization of the basal hydrology in a one-dimensional vertically integrated model. This parameterization accounts for connectivity between the basal hydrological system and the ocean in the transition zone. Our model produces a smooth transition between finite basal friction in the ice sheet and zero basal friction in the ice shelf. Through a set of experiments based on the Marine Ice Sheet Model Intercomparison Project (MISMIP), we show that a smoother basal shear stress, in addition to adding physical realism, significantly improves the numerical accuracy of our fixed-grid model, allowing for reliable grounding-line dynamics at resolutions of about 1 km.


Marine ice formation at the southern McMurdo Ice Shelf, Antarctica


Corresponding author: Inka Koch

Corresponding author e-mail: inka.koch@otago.ac.nz

Marine ice forms at the base of ice shelves from a mixture of sea water and fresh water and can enhance ice-shelf stability and influence its rheology. Although widespread in Antarctic ice shelves, the conditions for marine ice formation remain poorly understood. This study relates variations in marine ice crystal morphology and ice chemistry to the output of a boundary layer freezing model to understand marine ice freezing speed and source water. Marine ice was sampled from the surface of the southern McMurdo Ice Shelf (SMIS), Antarctica, with a Kovacs corer in three shallow ice cores (2.5–9.5 m long); basal ice-shelf layers are locally exposed due to a negative surface mass balance (–0.16 m a–1). SMIS marine ice is on average enriched in isotopes (1.25‰ δ18O, 9.58‰ δD) and weakly saline (0.25 ppt). Banded and granular ice crystals in thin section and a low salinity suggest that the marine ice was generated from fast-forming essentially salt-free frazil ice crystals only. These small disks of freely floating ice form episodically from supercooled water in contrast to ice formed from a slowly advancing freezing front. The boundary layer freezing model suggests a small effective fractionation coefficient for fast forming (10–6 m s–1) frazil ice (α = 1.33 for δ18O and β = 8.77 for δD), hence weakly enriched marine ice source water for most ice samples. Therefore, some marine ice must be recycled to form the enriched marine ice (δ18O > 0.5‰). Only few marine ice samples are relatively depleted in heavy isotopes, pointing toward a mixed water source of sea water and glacial water. A steep co-isotopic slope of 8.4 together with an isotopic range of ~3‰ δ18O confirms that the marine ice was generated from changing water sources rather than a constant water source at varying freezing speeds. Results from this study suggest that frazil marine ice at SMIS is not only formed as a result of a thermohaline circulation in the ice cavity in winter (ice pump mechanism), sourcing glacial water from deeper parts of the ice shelf. It also forms in summer when marine ice surface melt is routed to the ice-shelf base through tide cracks, where the fresher water becomes supercooled upon mixing with colder saline water. In order to predict non-linear ice-shelf behavior in a changing climate, it is therefore necessary to consider oceanic conditions in the ice-shelf cavity as well as ice-shelf surface conditions and internal structure.


Layered model of Antarctic ice-sheet dynamics based on results of borehole directional survey and surface radar profiling

Alexey N. MARKOV, Pavel G. TALALAY

Corresponding author: Alexey N. Markov

Corresponding author e-mail: am100@inbox.ru

For many years four boreholes were logged on the traverse between Vostok and Mirny stations, East Antarctica: at Vostok station to a depth of 1920 m (25 years of monitoring); Vostok-1 and Mirny stations to a depth of 450 m; and Pionerskaya station to a depth of 350 m. Data from the borehole directional survey were compared with data from the surface radar profiling, and as a result a good correlation (0.74) between the change of ice flow velocity with depth and the change of relative amplitude of the reflected radar signal with depth was found. This allows us to assume that the radar profiles identified the layered structure of ice flow, and therefore we can speculate that the entire Antarctic ice sheet also has a layered structure of flow. The changes in the dynamic parameters of the upper snow–firn layer down to 100–120 m depth showed that this layer can be defined as a local dynamic structure on the Antarctic ice sheet. The snow–firn layer’s flow direction greatly differs (by 30…80°) from the flow direction of the underlying layers. Ice sheet under snow–firn cover is divided into three layers, which can be defined according to structural examination of radar cross-section and ice flow. Independently of the ice thickness, the ratio between layer thickness is approximately 1/4 : 2/4 : 1/4. As a result of 25 years of borehole logging at Vostok station, the second layer (in the range 105–800 m) was additionally subdivided into five sub-layers. The flow direction of each sub-layer is changed in the ‘fan’ manner within 17°. Accounting geodetic data of ice-sheet surface, the absolute velocities of ice flow were obtained. The analysis results showed that the middle ‘lens-shaped’ structure of the ice sheet is more rigid and extrudes the softer ‘gelatinous’ lower layer.


Elevation limits to supraglacial lake drainage in western Greenland


Corresponding author: Kristin Poinar

Corresponding author e-mail: kpoinar@u.washington.edu

Moulins formed by hydrofracture under supraglacial lakes have been identified as important conduits for surface meltwater to reach the bed, with implications for ice-sheet sliding velocity and sea-level rise. As the climate warms and lakes form higher on the Greenland ice sheet, drainage of these lakes could shift the basal thermal/hydrological regime. This is because lower on the ice sheet, where lakes annually drain water to an already wet bed, the input of additional water is unlikely to affect long-term sliding velocities, but lake drainages in areas where the bed is frozen could introduce basal sliding there and have a more profound effect on ice flow. It is sometimes assumed that as lakes expand inland onto higher areas of the ice sheet, moulins too will spread inland toward this purported area of basal vulnerability. We explore this assumption: will lakes higher on the ice sheet necessarily drain and, if they do, will they change the basal thermal regime? We use a combination of remotely sensed imagery (RADARSAT, Landsat, ASTER and WorldView) to identify areas where supraglacial lakes are migrating up the ice sheet and to assess whether these high lakes are draining. Because lake drainage requires the nucleation of a crevasse, which requires stress exceeding the tensile strength of ice, we also examine the strain rate field in order to constrain whether drainage cracks can initiate under these lakes. We note that strain rates generally decrease with elevation and ice thickness, so that there may be an elevation on the ice sheet above which lake drainage cannot initiate. Indeed, of the few (five) lakes Selmes and others (2013) identified above 1700 m elevation in western Greenland, none drained during their 5 year study period. We also assess the thermal state of the bed under these high lakes to evaluate the likelihood of lake water reaching previously frozen areas of the bed. Some models of basal temperature predict a wet bed almost all the way to the divide in western Greenland (~3000 m elevation). Our model also finds a wet bed well beyond the zone of lake expansion (~2400 m elevation) there. The spread in model results is likely due to the high sensitivity of models to geothermal flux, which is known relatively poorly. We explore the sensitivity of the area of basal vulnerability to uncertainties in geothermal flux.


Temperature distribution and thermal anomalies along a flowline of the Greenland ice sheet


Corresponding author: Joel Harrington

Corresponding author e-mail: jharrin9@uwyo.edu

Warming of ice in the ablation zone increases glacier motion, both through enhanced internal deformation in the softened ice and basal sliding. The bulk ice temperature increases from the cold interior ice to warmer ice in the ablation zone and finally to temperate ice near the margin. However, this warming has not been documented within the Greenland ice sheet (GIS). Englacial and basal temperature data for the GIS are sparse and mostly limited to ice-core sites deep in the interior and ice streams, providing an incomplete representation of the thermal state of the ice sheet. Here we present 11 temperature profiles along a 34 km E-W transect extending inland from the margin of Isunnguata Sermia, an outlet glacier on the western GIS. Sensor strings were emplaced in multiple boreholes at five drill sites, recording the transect of ice temperatures along a flowline as well as local temperature variations in close proximity boreholes. A temperate basal layer is present in all profiles, increasing in thickness down-glacier, where it expands from ~4% of the total ice height furthest inland to 100% at the margin. This temperate thickness evolution is inconsistent with modeled heat contributions of ice deformation and vertical extension of the temperate layer due to longitudinal ice compression. We suggest that basal crevassing, facilitated by water pressures at or near ice overburden pressure, has allowed for the large temperate thicknesses observed. Some sites exhibit a kink of warmer temperatures at 40–90 m depth, below the influence of the seasonal temperature wave. We propose that this feature is dissipating heat from when the ice occupied a partially water-filled crevasse field up-glacier. These new temperature data provide additional information on the thermal state near the bed of the GIS and highlight the local influence of a variety of meltwater heat sources on ice temperatures in the ablation zone. With increasing meltwater production at the surface as well as a reduced thermal barrier for englacial water propagation, meltwater plays an increasingly critical role in ice temperature control.


A new SPOT5-HRS snapshot of polar ice: the SPIRIT2 project

Etienne BERTHIER, Jérôme KORONA, Marc BERNARD, Frédérique RÉMY, Thomas FLAMENT, Steven HOSFORD, Ted SCAMBOS

Corresponding author: Etienne Berthier

Corresponding author e-mail: etienne.berthier@legos.obs-mip.fr

The SPIRIT (SPOT 5 stereoscopic survey of Polar Ice: Reference Images and Topographies) project took place during IPY (2007/09) and, through >70 publications, contributed to increase our knowledge of the topography of polar ice masses. This project allowed the free delivery of 5 m ortho-images and 40 m digital elevation models. Acknowledging the rapid evolution of glaciers and at the margins of the ice sheets and given that SPOT5 will soon end its life in orbit, CNES, Astrium and LEGOS launched a new acquisition campaign, named SPIRIT2. SPOT5-HRS will acquire new stereo-pairs from November 2013 to April 2014 in Antarctica and in July–October 2014 in the Arctic regions. Here we will briefly recap the main achievements of the SPIRIT project with a focus on selected results, e.g. the depiction of the pattern of elevation changes that occurred in the northern peninsula following ice-shelf collapses and the dramatic surface signature (70 m drawdown) following the drainage of a subglacial lake in Wilkes Land (Antarctica). We will also present the cloud-free coverage achieved during the ongoing Antarctic SPIRIT2 campaign. Finally, we will take this opportunity to clarify how to access previously generated SPIRIT products, as well as how to order new SPIRIT products from archived HRS stereoscopic data.


Glaciers on the northern Antarctic Peninsula are more sensitive to temperature change than precipitation change

Bethan DAVIES, Nicholas GOLLEDGE, Neil GLASSER, Jonathan CARRIVICK, Nicholas BARRAND

Corresponding author: Bethan Davies

Corresponding author e-mail: beth_davies2000@yahoo.com

Glaciers peripheral to the Antarctic ice sheet are projected to make a large contribution to sea level over coming centuries. However, there is high uncertainty in future contributions, because increased precipitation may offset future surface melting from higher air temperatures. To investigate this, we examine glacier–climate relationships during the Holocene by bringing together geomorphology, well-constrained glaciology, a local, highly resolved ice-core record and numerical ice-flow modelling. James Ross Island, NE Antarctic Peninsula, preserves a rare, detailed record of mid-Holocene glacier and ice-shelf interactions in an area of contemporary rapid warming, glacier recession and ice-shelf collapse. During the Holocene, a 10 km readvance of Glacier IJR45 on Ulu Peninsula, James Ross Island, is documented by a large boulder train from the glacier to a prominent moraine, which post-dates 4–5 cal. ka BP. Ice-core records indicate that the period 2–5 ka BP was ~0.5°C warmer. This was also a period of ice-shelf absence. This implies that any readvances at this time must have been driven by increased precipitation, and that the glacier is more sensitive to precipitation than air temperature. A high-resolution numerical flowline model was used to relate these glacier fluctuations to climate. By driving the model with ice-core climate data, we found that glaciers on Ulu Peninsula remained largely stable during the mid-Holocene. From 2 to 5 ka, simultaneous ice-shelf collapse and a small amount of glacier recession occurred during the 0.5°C warming. During a period of rapid cooling starting 2 ka, the ice-shelf reformed and IJR45 advanced to its maximum Holocene position. Despite aggressive scaling of precipitation with temperature, we were unable to force a glacier advance during the warmer mid-Holocene. Following climatic amelioration starting 600 years ago, the glacier receded to its most recent stable position. Like many glaciers around the Antarctic Peninsula, IJR45 shows high sensitivity to atmospheric air temperatures and low sensitivity to precipitation, meaning that advance during a warm period is unlikely. Our study indicates that forecast increases in precipitation are unlikely to significantly offset melt-induced glacier recession. Consequently, the currently observed trends of glacier recession, thinning and acceleration will most likely continue throughout the next century.


Investigating the effect of plume characteristics on ice-front circulation and tidewater glacier submarine melt rates


Corresponding author: Donald Slater

Corresponding author e-mail: d.slater@ed.ac.uk

Recent observations of the mass balance of the Greenland ice sheet have shown significant losses at its coastal margins. Much of this loss has been attributed to the thinning, speed-up and retreat of tidewater glaciers, which occurred in the presence of particularly warm ocean waters. Ocean forcing of tidewater glaciers is thus thought to be key in understanding their dynamics. Estimates of the rate at which warm ocean waters will melt submerged ice at the front of tidewater glaciers have been derived both from observational data and numerical fluid models. Due in large part to the difficulties of directly observing and continuously monitoring the ice–ocean interface, there remain considerable uncertainties in these models and their submarine melt estimates. In numerical modelling, great care must be taken in choosing the domain size and resolution, boundary conditions, numerical parameters and the characterization of ice-sheet runoff (i.e. meltwater volume, portal shape, emerging flow velocity) since all have an effect on the induced circulation and submarine melt rate. In this study, we use a general circulation model (MITgcm) to investigate how varying several of the most poorly constrained variables affects ice-front water circulation and thus submarine melt rates at tidewater glacier termini. More specifically, we investigate the effect of (1) varying the emerging ice-sheet runoff velocity and portal size on plume position relative to the ice and thus ice-front water circulation, and (2) the sensitivity of the results to variation in the model set-up and numerical parameters to understand how these can impact modelled plume dynamics and thus our estimates of submarine melting.


DynEarthSol3D: an unstructured finite-element method to study elasto-viscoplastic ice dynamics


Corresponding author: Liz Logan

Corresponding author e-mail: esl359@gmail.com

We present a new thermomechanical model for dynamic ice flow called DynEarthSol3D (DES, available at http://bitbucket.org/tan2/dynearthsol3d). The model employs an elasto-viscoplastic constitutive law and simulates the self-consistent development of ice yielding using Mohr–Coulomb failure. DynEarthSol3D uses a finite-element method to solve the momentum and heat equations in Lagrangian form on an unstructured tetrahedral mesh. The model adaptively refines the mesh in areas of failure and is explicit in time, which provides stability despite the complicated flow law. While DES has already been benchmarked and presented to the tectonophysics community we present results from a benchmark prognostic experiment for ice flowing down an inclined plane over a basal perturbation. Viscous, Maxwell viscoelastic and elasto-viscoplastic cases are compared to analytic solutions for the Stokes equation for both Newtonian and non-linear fluids and different basal boundary conditions. Based on the closeness of fit between numerical and analytical solutions we suggest DES is a useful tool for exploring ice deformation and yielding patterns on decadal to century timescales.


Internal layering along a traverse from Zhongshan station to DT401, East Antarctica, by ground-based radio-echo sounding

TANG Xueyuan, SUN Bo, WANG Tiantian, CUI Xianbin, ZHANG Dong

Corresponding author: Tang Xueyuan

Corresponding author e-mail: tangxueyuan@pric.gov.cn

During the 21th Chinese National Antarctic Research Expedition (CHINARE 21, 2004/05), a radar dataset was collected by a ground-based radar system, along a 1083 km ice-sheet profile from Zhongshan station to DT401 (130 km from Dome A). Using the new radar data, the ice-sheet structure and subglacial conditions were revealed. The continuous internal layers, disturbed layers and echo-free zones (EFZ) along the profile were identified and classified, and its spatial distribution was presented. Basing on the recent surface ice velocity, we found that several internal layers at a depth of 200–300 m in the upper ice sheet are continuous and smooth, and approximately parallel to the ice surface topography. The thick band of the continuous layers changes little with increasing latitude. Below 300 m, the geometric structure of the internal layers within the ice sheet and the vertical width of the EFZ band were influenced by the surface ice velocity and the bed topography. The relatively high disturbance of the layers and the larger width of the EFZ band correspond directly to higher surface ice velocity and sharper bed topography.


Validating modeled firn densification rates using repeat-track airborne radar data


Corresponding author: Stefan Ligtenberg

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

Density and thickness of the Antarctic firn layer varies both in time and space. These variations are important to consider for ice-sheet mass-balance studies because they contribute to the surface elevation change measured by airborne and satellite altimeters. The firn-layer signal is often modeled using a firn densification model in the absence of direct densification measurements. As a result, these simulated firn densification (FD) rates are difficult to evaluate, as no direct observations exist. For example, remotely sensed and in situ observations of surface elevation change are comprised of several signals including FD, as well as changes due to bedrock movement, ice dynamical changes and surface processes (snowfall, sublimation and melt). Burying measurement devices in firn at various depths and tracking their vertical velocity is another measurement possibility, but their potential contamination of the local FD signal is unknown and this method produces only a few isolated point measurements. Here we use repeat-track radar data over a 3 year interval (2009–2011) from NASA’s IceBridge mission above the Thwaites Glacier region in West Antarctica to directly measure FD rates. We map the distinct radar horizon signature visible in the three survey years and determine the temporal evolution of the layer thicknesses to assess the FD rate. These measured FD rates are compared with a firn model simulation to examine its spatial and temporal variability. Since the radar can see up to 30 annual layers, FD rate is measured at different depths to obtain a vertical profile of FD rate.


Determination of internal and basal conditions of a heterogeneous glacier structure based on 50 MHz radar data

Guisella GACITUA, Jose Andres URIBE, Thomas LORIAUX, Rivera ANDRES, Ryan WILSON

Corresponding author: Guisella Gacitua

Corresponding author e-mail: ggacitua@cecs.cl

A 50 MHz radar was used on a helicopter platform to survey a high-elevation glacier (4400 m a.s.l.) in the central Andes, Chile. We obtained data over a track length of 33 km across the whole glacier surface (4.36 km2). In recent flow velocity estimations, this glacier has shown polythermal conditions, which is a decisive factor on the behavior of the radar signal observed. The surface and internal structure of the glacier show strong heterogeneity and complex signal behavior. The echo amplitudes and signal polarity were used to locate the dominant scattering regions and identify debris-rich basal ice zones. We compute power reflections from interfaces and internal losses to determine dielectric properties of the subglacial material. The signal analysis also provides the bedrock interface roughness estimates, which are then compared with the theoretical reflected energy, which allows for verification of the efficiency regarding methodology used under these heterogeneous conditions of the ice.


A comparison of ice flow relations for ice-sheet modelling

Adam TREVERROW, Roland C. WARNER, William F. BUDD, Tim H. JACKA

Corresponding author: Adam Treverrow

Corresponding author e-mail: adamt0@utas.edu.au

Incorporating a physically accurate description of ice deformation processes into ice-sheet models is a key component in reducing the uncertainty in predictions of ice-sheet contributions to changes in global mean sea level. Recent developments in some ice-sheet models, in particular the capability to determine the three-dimensional distribution of stresses throughout an ice mass, makes it possible to incorporate a description of anisotropic ice flow properties into large-scale dynamic simulations. Despite these model developments the selection of an appropriate numerical relationship between ice strain rates and the stresses driving the flow remains an issue in improving the accuracy of simulations. In this study we use observations from the Dome Summit South (DSS) ice-coring site at Law Dome, East Antarctica, to model the vertical distribution of deviatoric stress components at the DSS borehole site. Data used include ice-core crystal fabric measurements, borehole temperatures and shear strain rates derived from repeat logging of the post-drilling borehole inclination. We compare the predictions four flow relations that have been proposed in the literature. These range from grain-scale flow relations incorporating descriptions of ice microdynamic processes – including nearest-neighbour grain interactions – through to empirical parameterizations derived from laboratory ice deformation experiments. In grain-scale flow relations the effect of polycrystalline anisotropy on enhancing flow rates is based primarily on consideration of individual crystal orientations and their relationship to the applied stress configuration. In combined stress configurations this can lead to lower estimates of flow enhancement than those obtained from scalar flow relations. As a reference we use depth-integrated solutions to the stress equilibrium equations to obtain column-averaged stress estimates and find that of the flow relations considered those where the effects of polycrystalline anisotropy are parameterized by a scalar function provide the most realistic and computationally efficient simulations.


Quantifying mass-balance processes on the Patagonian Icefields


Corresponding author: Marius Schaefer

Corresponding author e-mail: mschaefer@uach.cl

We present surface mass-balance simulations of the Patagonian Icefield that were driven by global climate data (reanalysis/GCM) which were downscaled using the regional climate model Weather Research and Forecasting (WRF) and statistical downscaling methods. The special climatic situation in the region, with sharp climate gradients introduced by the blocking of the westerlies by the high peaks of the icefields, are reproduced by downscaled climatic data. The mass-balance simulations were validated and interpreted using geodetic mass balances, measured point balances and a complete velocity field of the Southern Patagonia Icefield (SPI) from spring 2004. The high measured accumulation of snow as well as the high measured ablation values are reproduced by the model. Subtracting the modeled surface mass balance from the geodetic balances, calving fluxes of major outlet glaciers were inferred. Good agreement with calving fluxes estimated from velocity data was obtained in many cases; however, on several glaciers the inferred calving fluxes seem to overestimate the measured calving fluxes. The measured calving fluxes exhibit large uncertainties due mostly to unknown ice thickness data and evolution of glacier velocities through time. The accumulation of snow and its redistribution due to wind drift present the major uncertainties in the modeled surface mass balance. Assuming no substantial changes in ice flow, the surface mass-balance model driven by ECHAM5 data in the A1B scenario predicts a contribution of the Patagonian Icefields to sea-level rise in the 21st century of 7.3 mm.


Climatic variability in the Davis sea sector (East Antarctica) over the past 250 years based on the 105 km ice-core geochemical data


Corresponding author: Diana Vladimirova

Corresponding author e-mail: divladi_0401@mail.ru

In this study we present the air temperature and snow accumulation rate reconstruction in the Davis sea sector of East Antarctica over the past 250 years based on the geochemical investigations of the ice core from the 105 km borehole (105 km inland from Mirny station) drilled in 1988. The core was dated by counting the annual layers in the stable water isotope content (δD and δ18O) profile and using the absolute date marker (Tambora volcano layer). The accumulation values were deduced from the thickness of the layers multiplied by the core density. The isotope content was transformed into the air temperature by comparing with the instrumental meteorological data from Mirny station. The reconstructed temperature series demonstrates a 0.5°C warming over the last 250 years. At the same time, snow accumulation rate was decreasing at least since the middle of the 19th century. The climatic characteristics demonstrate cyclic variability with periods of 6, 9, 19, 32 and about 120 years. Interestingly, in frames of 19 year cycle the temperature and isotope content are negatively related, which could be explained by a zonal shift of the moisture source area. Based on the data of the sodium concentration and ‘deuterium excess’ values in the ice core, we assumed an increased sea-ice extent in the 19th century comparing with the present day.


Ten years of mass balance from glaciological and geodetic methods for Glaciar Bahía del Diablo, Vega Island, Antarctic Peninsula

Sebastián MARINSEK, Pedro SKVARCA, Evgeniy ERMOLIN

Corresponding author: Sebastián Marinsek

Corresponding author e-mail: smarinsek@dna.gov.ar

Glaciers located on the northeastern Antarctic Peninsula are currently affected by increased atmospheric warming, revealing substantial changes in their dynamic behavior. At present only a few mass-balance programs are being carried out on the Antarctic Peninsula, among them that of ‘Glaciar Bahía del Diablo’, which terminates on land on Vega Island, with 10 years of uninterrupted mass-balance measurements. During this period, the annual mass balances were determined by field observations and measurements utilizing the glaciological method. The cumulated mass loss derived from field surveys resulted in 1.98 ± 0.3 m of water equivalent per unit area. The calibration of the mass-balance results from the glaciological method against geodetic volume change is strongly recommended. We present here the mass change for the same 10 year period derived from high-accuracy surface DEM differencing, obtained by DGPS surveys. Total mass loss obtained by means of the geodetic method yielded 2.3 ± 0.03 m of water equivalent per unit area.


Concentrated englacial shear over rigid basal ice, West Antarctica: implications for modelling and ice-sheet flow


Corresponding author: Neil Ross

Corresponding author e-mail: neil.ross@ncl.ac.uk

Basal freeze-on, deformation and ice crystal fabric reorganization have been invoked to explain thick, massive englacial units observed in the lower ice column of the Antarctic and Greenland ice sheets. Whilst recognized as having very different rheological properties to overlying meteoric ice, studies assessing the impact of basal units on the large-scale flow of an ice sheet have so far been limited. We report the discovery of a previously unknown, extensive (100 km long, >25 km wide and up to 1 km thick) englacial unit of near-basal ice beneath the onset zone of Institute Ice Stream, West Antarctica. Using radio-echo sounding, we describe the form and physical characteristics of this unit, and its impact on the stratigraphy and internal deformation of overlying ice. The lower englacial unit, characterized by a highly deformed to massive structure, is inferred to be rheologically distinct from the overlying ice column. The overlying ice contains a series of englacial ‘whirlwind’ features, which are traceable and exhibit longitudinal continuity between flow-orthogonal radar lines. Whirlwinds are the representation of englacial layer buckling, so provide robust evidence for enhanced ice flow. The interface between the primary ice units is sharp, abrupt and ‘wavy’. Immediately above this interface, whirlwind features are deformed and display evidence for flow-orthogonal horizontal shear, consistent with the deformation of the overlying ice across the basal ice unit. This phenomenon is not a local process, it is observed above the entirety of the basal ice, nor is it dependent on flight orientation, direction of shear is consistent regardless of flight orientation. These findings have clear significance for our understanding and ability to realistically model ice-sheet flow. Our observations suggest that, in parts of the onset zone of Institute Ice Stream, the flow of the ice sheet effectively ignores the basal topography. Instead, enhanced ice flow responds to a pseudo-bed, with internal deformation concentrated and terminating at an englacial rheological interface between the upper ice-sheet column and the massive basal ice. Our results demonstrate that we may need to: (1) adapt numerical models of those parts of the ice sheet with extensive and thick basal ice units; and (2) carefully reconsider existing schematic models of ice flow, to incorporate processes associated with concentrated englacial shear.


Sub-Antarctic dryness causes amongst the largest glacier wastage on Earth

Vincent FAVIER, Deborah VERFAILLIE, Etienne BERTHIER, Jennifer KAY, Vincent JOMELLI, Martin MENEGOZ, Lauren DUCRET, Yoann MALBÉTEAU, Daniel BRUNSTEIN, Hubert GALLÉE, Young-Hyang PARK

Corresponding author: Deborah Verfaillie

Corresponding author e-mail: deborah.verfaillie@ujf-grenoble.fr

Although recent decline in southern sub-polar glaciers has been especially rapid and widespread, the climatic causes of this decline are still rather enigmatic. In the present study, we introduce a mass-balance estimate of Cook Ice Cap, located on Kerguelen Archipelago (49° S, 69° E), which was produced by comparing digital elevation models from 2000 and 2009. This mass-balance estimate is then compared with field topography and stake measurements from the 1970s and 2011. Our mass-balance estimate, –1.51 ± 0.19 m w.e. a–1, is one of the most negative worldwide. Results from a distributed glaciological model show that the ice cap decline is mostly due to a decrease in precipitation, while it was amplified by surface warming. Analysis of historical climate data, climate model outputs and reanalysis data reveals that the inception of fast glacier retreat in the 1960s was concurrent with a short but intense drought, which has continued from the 1970s on.


Extreme SMB gradients on Patagonian icefields revealed by high-resolution regional climate modelling


Corresponding author: Jan Lenaerts

Corresponding author e-mail: j.lenaerts@uu.nl

Gravimetric observations and differential DEMs indicate that the Patagonian icefields and their outlet glaciers are currently thinning and retreating. Little is known, however, about the mechanisms that drive this ice mass loss. In particular, surface mass balance (SMB) of the icefields is poorly constrained, because the wet climate of the southern Andes complicates performing in situ measurements. Instead, regional climate modeling may provide the first independent estimate of the icefield’s SMB. Here we present results from a high-resolution (5.5 km) regional atmospheric climate and multi-layer snow model (RACMO2), forced by ERA-Interim atmospheric and ocean surface fields (1979–2012), which is evaluated using existing weather stations, precipitation gauges and available glacier firn cores. Our results confirm the occurrence of extremely high accumulation on the higher portions of both icefields (10–35 m w.e. a–1 of snow). Precipitation is abundant throughout the year, and driven by persistent atmospheric westerlies, in combination with significant orographic forcing. On the other hand, we find strong ablation on the narrow outlet glaciers tongues, although these are only partly resolved by the model grid. Using modelled spatial SMB gradients, we can downscale the model SMB field to a high-resolution (~100 m) DEM and ice mask. This enables study of SMB at individual glacier catchment scale and, if combined with ice velocity and thickness data, allows for realistic constraints on present-day NPI and SPI mass balance and their contribution to sea-level rise.


Climate and SMB around Antarctic ice rises – a combined modeling and observational approach


Corresponding author: Jan Lenaerts

Corresponding author e-mail: j.lenaerts@uu.nl

Ice rises are topographic features that protrude from Antarctic ice shelves, and form as a result pinning points that connect the floating ice shelves locally to bedrock. The genesis of these features is poorly understood; since local ice flow is generally slow, surface mass balance (SMB) is assumed to control their formation, but very few data exist on the SMB on and around ice rises. Here we focus on Dronning Maud Land (East Antarctica), where many ice rises are found, and combine observations from weather stations and observed SMB from firn cores and GPR with output of the high-resolution (5.5 km) climate model RACMO2 (2001–2012). We find that ice rises strongly control the local climate and SMB. Orographic precipitation is high (> 1 m w.e. a–1) on the windward side of some ice rises, whereas wind-related snow erosion and sublimation, combined with precipitation shadowing, create a local minimum SMB on their lee side. These strong SMB gradients (200–600% per ~50 km) are shown to depend on ice rise width, orientation and maximum elevation. These results suggest that SMB processes and their spatial variability need to be accounted for in future glaciological studies on ice rises.


Assessing the extent of the Last Interglacial Antarctic ice sheet


Corresponding author: Matthew Whipple

Corresponding author e-mail: matt.whipple@bristol.ac.uk

Understanding the response of the Antarctic ice sheet (AIS) to warmer climate, such as during the Last Interglacial (LIG), is of vital importance to understand future changes in sea levels. Kopp and others (2009) gives a 50% chance of LIG sea levels exceeding 8.5 m above current levels, with other studies resulting in a modelled 4.2–5.8 m from Antarctic sources, assuming hemispheric synchronicity. Previous modelling work has not modelled warming to the degree suggested by the ice-core proxy data. Here we attempt to account for part of the Antarctic component of the additional sea level during the LIG through climate sensitivity testing by HadCM3 climate modelling for different ice-sheet scenarios, comparing with precipitation-weighted temperature to stable water isotope data in six East Antarctic ice-core records. We look at the effect of changing ice-core site elevation during the LIG, through Glacial Isostatic Adjustment (GIA) modelling, by adapting the ice-sheet models of Bradley and others (2012, 2013) for use in HadCM3 climate simulations. A collapse of the marine-based West Antarctic ice sheet (WAIS) provides changes in the precipitation regime over the East Antarctic ice sheet (EAIS), meaning that some ice-core temperature reconstructions using precipitation-weighted temperature show warming on the order seen in ice-core isotope data. Other EAIS ice-core isotope data show absolute temperature offsets, which we are unable to resolve within reasonable glaciological limits. However, the suggestion of reduced thickening over the LIG EAIS would come closer to resolving the offset temperature offset, and also to resolving the sea-level budget.


Deglacial ice-sheet meltdown: orbital forcing, CO2 effects, and ocean response as simulated in LOVECLIM-IcIES


Corresponding author: Malte Heinemann

Corresponding author e-mail: malteh@hawaii.edu

One of the most recent massive climate change events in Earth’s history was the last glacial termination 19-9 thousand years before present (ka BP). Northern Hemisphere ice sheets receded quickly, causing global sea level to rise by more than 100 m, meltwater was injected into the North Atlantic halting its deepwater formation, atmospheric CO2 concentrations rose by almost 100 ppmv, and the global surface warmed by about 4°C. It is still unresolved what exactly caused the reconstructed climate change and ice-sheet melting. To address this question, we conduct transient modeling experiments with a coupled 3-D ice-sheet–climate model. The model includes the atmosphere–ocean–sea-ice–vegetation components of the intermediate complexity model LOVECLIM coupled to a Northern Hemisphere setup of the ice-sheet model IcIES via the exchange of surface temperature, precipitation, surface elevation and ice-sheet extent. A successful transient simulation of the reconstructed ice-sheet evolution during the last 80 ka requires temperature bias correction. A series of transient simulations of the last 25 ka of climate and ice-sheet changes supports some elements of the astronomical theory of ice ages: orbital changes have the potential to, and do in fact, initiate the last glacial termination in our model. However, our simulations also demonstrate that rising CO2 concentrations after 18 ka BP largely accelerated the deglaciation. Without this deglacial CO2 increase, wide areas of North America and Scandinavia would be covered by land ice today. Present sea level would be as much as 100 m lower, and still dropping by about 2.5 m per thousand years due to growing ice sheets. When an ice-sheet–land–ocean–freshwater coupling scheme based on present-day catchment basins and river mouths is included, the deglacial meltwater runoff into the Atlantic and Arctic Oceans leads to a complete shutdown of North Atlantic Deep Water (NADW) formation. Our first model results indicate, however, that the NADW recovery critically depends on the opening of the Bering Strait due to sea-level rise, which is not captured explicitly in the ocean model. Furthermore, to realistically capture the deglacial ocean circulation changes, an interactive hydrological model that accounts for regional ice-sheet variations and isostatic adjustments is desirable.


Estimation of the total volume of Svalbard glaciers and their potential contribution to sea-level rise derived from scaling methods


Corresponding author: Francisco Navarro

Corresponding author e-mail: francisco.navarro@upm.es

We estimate the total volume of Svalbard glaciers to be 7504 ± 312 km3, which corresponds to 19 ± 2 mm of sea-level equivalent. Our estimates are based on scaling relationships specifically calibrated for Svalbard glaciers using a sample of 103 volume-area pairs. We assess the sensitivity of the scaling exponents to different glacier characteristics such as glacier size, aspect ratio and average slope. We find the volume of steep-slope and cirque-type glaciers to be less sensitive to changes in glacier area. The most accurate scaling models, in terms of lowest cross-validation errors, are obtained from: (1) a logarithmic least-squares regression of the glacier volume as a function of glacier length; and (2) a multivariate approach using glacier length and elevation range as predictors additional to glacier area, together with an absolute deviation misfit function. Our results indicate that studies based on regional scaling laws provide more consistent volume estimates for Svalbard that those based on global scaling, in spite of being based on a smaller calibration dataset.


Effect of elevation feedback in coupling a climatic mass-balance model to an ice-flow model: Vestfonna ice cap, Nordaustlandet/Svalbard


Corresponding author: Martina Schäfer

Corresponding author e-mail: martina.schafer@fmi.fi

The ice caps Austfonna and Vestfonna on Nordaustlandet, Svalbard, represent one of the largest ice-covered areas in the Eurasian Arctic and might be important contributors to future sea-level changes. We present prognostic simulations into the future of Vestfonna ice cap to investigate the effect of geometry changes (i.e. surface elevation) on the mass balance. In their current configuration, the two ice caps can be treated separately since they are not connected. We use the full-Stokes finite-element code Elmer/Ice (CSC-Finland) and a temperature net radiation index climatic mass-balance model during the period from 2006 to 2100. To conduct prognostic simulations of the ice dynamics it is crucial to understand the basal sliding and possible surge behaviour since Vestfonna ice cap is characterized by fast-flow regions (velocities over several hundreds of meters per year) even though low temperatures and low balance gradients generally result in low flow rates on the glaciers of Svalbard. In lack of an ice-flow model incorporating a physically based model for sliding and surging, we use in this study the spatial pattern of basal drag inferred from satellite-derived surface velocities using a Robin inverse method evaluated for 1995 and 2008 showing clearly different flow regimes and basal drag fields over one of the outlet glaciers. This indicates the importance to understand the temporal evolution of the basal properties in prognostic runs. We present detailed future projections of climatic mass balance by coupling the glacier dynamical model with the mass-balance model, taking into account the topographic feedback between the two models. Resulting ice mass evolution is also derived. The coupling between the ice-flow and mass-balance model is implemented at different coupling intervals (50, 25, 10 years), as well as with different elevation-dependent lapse-rate approaches. The climatic forcing was provided by GCM MIROC-ESM in the form of CMIP5 multi-model output representing the scenarios RCP 2.6, 4.5, 6.0 and 8.5. The climatic mass balance of Vestfonna will become increasingly negative over the 21st century, leading to surface lowering and an overall retreat of the ice cap within the dynamic model. We conclude that coupling of climatic mass-balance models and ice-flow models taking into account the elevation feedback is crucial in prognostic future simulations.


Greenland high-elevation mass balance: reviewing the assumption of reference period equilibrium

William COLGAN

Corresponding author: William Colgan

Corresponding author e-mail: william.terence.colgan@gmail.com

We revisit the input–output mass budget of the high-elevation region (>2000 m) of the Greenland ice sheet initially evaluated by the Program for Arctic Regional Climate Assessment (PARCA). We combine c. 1995 in situ velocities observed at PARCA stakes with surface mass balance from the regional climate model MAR and the ice thickness dataset developed by the Ice2Sea Project. During the 1961–1990 reference period, we assess 335 ± 50 Gt a–1 of net accumulation within, and 301 ± 35 Gt a–1 of outflow from, the PARCA perimeter. Our estimated mass balance of 33 ± 61 Gt a–1 is substantially greater than the 0 ± 21 Gt a–1 assessed by PARCA. This discrepancy is primarily due to differences in assessed mass input. Combining this estimate with 10 previous estimates, each of which has been corrected for surface mass-balance anomaly relative to reference period, yields a consensus reference period high-elevation mass-balance estimate of 13 (–30, +28) Gt a–1. With a 68% likelihood of high-elevation mass gain during the reference period, and a 27% likelihood of mass gain in excess of 20 Gt a–1, the probability distribution of this consensus estimate challenges the assumption that the ice sheet was in equilibrium during the reference period. The recent Greenland mass loss due to either surface mass-balance or ice dynamic processes is equivalent to net contemporary rate minus net reference period rate. While this framework is valid under reference period equilibrium mass balance, it potentially underestimates the recent mass loss due to either process by an amount equivalent to any reference period positive imbalance. Given that reference period accumulation rate appears to be characteristic of the millennial mean accumulation rate, and the apparent slope dependence, rather than surface mass-balance dependence, of limited in situ mass-balance observations, we speculate that long-term ice dynamics are the primary cause of subtle high-elevation thickening. The magnitude of the mass gain we infer is consistent with previously postulated millennial-scale ice thickening associated with the downward advection of the transition between ‘soft’ Wisconsin and ‘hard’ Holocene ice through Reeh (1985). Failure to account for millennial-scale dynamics may therefore underestimate the recent mass loss due to ice dynamics by up to ~9% (or ~13 Gt a–1).


Ice-sheet climates


Corresponding author: Michiel van den Broeke

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

Nowadays, spaceborne gravity, altimetry and feature-tracking observations provide quasi-continuous and ice-sheet-wide estimates of changes in mass, volume and ice velocity. However, satellite time series are still relatively short, sometimes discontinuous and may suffer from a lack in temporal (e.g. laser altimetry) and spatial (e.g. GRACE) resolution. Moreover, the conversion from volume-to-mass change requires a correction for firn processes (laser altimetry) and an additional correction for signal penetration (radar altimetry). Finally, apart from airborne snow radar, no remote-sensing technique exists to date to quantify surface mass fluxes. That is why it remains essential to model the surface mass balance of the ice sheets, which in turn requires detailed in situ climate observations. Here we present recent developments in modelling and in situ observational techniques of the surface climate and mass balance of the ice sheets of Antarctica and Greenland. Continuous improvements in regional climate models have provided reliable (sub-daily) maps and time series of surface mass fluxes (snowfall, sublimation, melt, refreezing and runoff), including a first-order estimate of drifting snow fluxes. Model evaluation benefited greatly from available remote-sensing products, and we review some recent mass-balance results (both Antarctica and Greenland) that are based on a combination of remote-sensing and model data. This includes the modelling of firn depth changes in Antarctica and perennial aquifer formation in Greenland. Finally we discuss recent developments in the field of observation and modelling of ice sheets, e.g. the application of Earth System Models to ice-sheet mass balance and that of very-high-resolution regional models to individual glacier basins and fjord systems and the development of a suitcase automatic weather station, iWS, for easy deployment in remote glaciated regions.


Towards a reconciled estimate of 20th century glacier contribution to sea-level change


Corresponding author: Paul Leclercq

Corresponding author e-mail: p.w.leclercq@uu.nl

Glacier mass loss is a major contributor to the 20th century sea-level rise. Yet, the uncertainty in estimates of the magnitude of the 20th century glacier contribution to sea-level rise is large. Two recent estimates, one based on observed glacier length changes and the other on modelled mass balance, give values of 67 ± 18 and 114 ± 5 mm SLE, respectively. These results differ significantly, despite the fact that both methods use calibration on mass-balance observations. In addition, the calibration on mass-balance observations poses additional uncertainty on the results, as recently concerns on the representation of global glacier mass balance by the available measurements have risen. We look for an explanation of the differences between the two estimates of 20th century glacier mass loss. Comparison of modelled glacier length changes with the observed glacier fluctuations shows that the model in general underestimates glacier length change. This underestimate of glacier retreat could be relevant as the forcing of the mass-balance calculations is dependent on the terminus elevation. Based on worldwide observed glacier length change we will explore the model sensitivity to the volume length scaling included in the model. On the other hand, the model results are used to test the representativeness of the observations to quantify the uncertainty in the estimate based on observed glacier fluctuations due to the sparseness of the glaciological observations. The combination of the benefits of both approaches hopefully brings us a more robust estimate of the glacier contribution to 20th century sea-level rise.


Differing ice dynamics across the Antarctic Peninsula


Corresponding author: Erin Pettit

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

The Antarctic Peninsula (AP) exhibits extreme orographically driven climate gradients because of its geography as a narrow north-south mountain range in the Southern Ocean transverse to the mean atmospheric circulation. The western side of the AP experiences one to two orders of magnitude higher snowfall rates than the eastern side. This spatial pattern of snow accumulation drives ice dynamics and ice–ocean interaction on timescales of decades to centuries, and shapes the subglacial topography and submarine bathymetry on timescales of glacial cycles. In order to balance this snow input, short steep glaciers transfer ice to the fjords on the west side, while longer lower-sloping glaciers on the east side carve deeper fjords. In our study, we calculate the ice flux using a drainage model that incorporates the modern ice surface topography, the RACMO-2 precipitation estimate, and recent estimates of ice thinning. Our results, coupled with observed rates of ice flow speed from InSAR (personal communication from Joughin; and Landsat-derived flow rates), provide an estimate of the fjord depths in grounded-ice areas for the largest outlet glaciers on each side of the peninsula. Despite lower accumulation rates, the large elongate drainage basins on the east side result in a greater ice flux funneled through fewer deeper glaciers. Due to the relation between ice flux and erosion, these east-side glaciers, therefore, have longer and deeper fjords than the west-side glaciers. Using ground-based data from the Bruce Plateau, the ice divide that feeds the Larsen Ice Shelf System to the east and numerous bays from Andvord Bay to Barilari Bay to the west, with a 2-D thermomechanical flowline model we compare ice dynamics between the east and west sides. Our results show that the ice divide is offset to the west from the bedrock divide by several kilometers. Using our measured average annual temperature at the ice divide of –15.1°C, our model suggests that the west-side glaciers have colder beds than those on the east side and that the ice divide position is migrating to the west due to increases in precipitation in recent decades. These distinct differences between the ice dynamics on the west and east side of the AP exert a primary control on the differing ice-shelf and grounding-line retreat, ice–ocean interactions and subglacial erosion rates, and provide context to understand sources of AP mass loss.


Influence of debris cover on glacier surface melting: a case study on Dokriani Glacier, central Himalaya, India


Corresponding author: Bhanu Pratap

Corresponding author e-mail: bhanuglacio@gmail.com

Most of the central Himalayan glaciers have varying layers of debris on their surfaces, which greatly affects the rate of ablation. The amount of debris cover is predicted to increase with ice mass wastage, and its influence on glacier surface melting is important for understanding glacier behavior in response to climate change. Thus an attempt has been made to relate the varying debris cover thickness with glacier surface melting. The repeated coordinates of 30 stakes have been used to calculate ablation for debris-covered and clean ice of Dokriani Glacier during 2010–2013. Dokriani is a valley glacier with 7 km2 of glacierized area formed by two cirque glaciers, having a partially debris-covered ablation area and terminating at 3945 m a.s.l. We found significant altitude wise difference in the rate of clean and debris-covered ice melting. The profile of mean annual mass-balance gradient for clean ice is with the linear fit of R2 ≥ 0.88 during the study period. Conversely, inconsistent ablation was found for debris-covered ice and it varies with variable debris thickness ranges from 1 to 40 cm. Melting was maximum for 1–6 cm of debris and abruptly at ~9 cm of debris it decreased and reached a minimum at 40 cm of debris thickness. Furthermore our study also found that even a small thickness (1–2 cm) of debris cover decreases the ice melting compared with clean ice on an annual basis. Overall the results suggest that 20% of debris-covered ablation area significantly reduces total ablation. However, still a strong phase of downwasting has been observed in the ablation area of Dokriani Glacier during the study period. We found substantial surface melting in the ablation zone by average annual ablation (aa) of –1.82 m w.e. a–1 compared with averaged annual accumulation (ca) of 0.41 m w.e. a–1. This shows that instead of having higher accumulation–area ratio (mean AAR) of 0.67 glacier thinned by –0.31 m w.e. a–1 (mean annual balance) from 2010 to 2013. Our study suggests that thinning ice mass rapidly increases debris cover over the ablation area and insulates ice loss.


Greenland ice sheet melt volume and runoff from satellite and in situ observations

Dirk VAN AS, Jason BOX, Dorthe PETERSEN, Michele CITTERIO, Konrad STEFFEN

Corresponding author: Dirk van As

Corresponding author e-mail: dva@geus.dk

Remote sensing provides surface melt area and regional mass change. In situ automatic weather station (AWS) data provide a relatively precise, but very local surface mass budget. Combining the two methods allows melt quantification for the entire Greenland ice sheet. We use interpolated near-surface air temperature from the GC-Net and PROMICE AWS networks, and remotely sensed MODIS surface albedo to calculate melt with a temperature/albedo-index melt model. The calculations make use of albedo to take into account absorbed shortwave radiation, the dominant melt parameter, and in doing so the darkening due to the melt–albedo feedback is accounted. Calculated ablation is calibrated using AWS data. Assuming that surface albedo is a first-order indicator of the firn’s available pore space and cold content, refreezing is parameterized as a function of it. Meltwater runoff for selected catchments is validated with river discharge data. The product: observation-based daily maps of near-surface air temperature, melt (extent and volume) and runoff for the Greenland ice sheet.


Sea-level contribution from the Greenland ice sheet deglaciation based on a data-constrained modelling methodology

Benoit LECAVALIER, Glenn MILNE, Matthew SIMPSON, Leanne WAKE, Philippe HUYBRECHTS, Lev TARASOV, Kristian KJELDSEN, Svend FUNDER, Antony LONG, Sarah WOODROFFE, Arthur DYKE, Nicolaj LARSEN

Corresponding author: Benoit Lecavalier

Corresponding author e-mail: b.lecavalier@mun.ca

An ice-sheet model was constrained to reconstruct the evolution of the Greenland ice sheet (GrIS) from the Last Glacial Maximum (LGM) to present to improve our understanding of its response to climate change. The study involved tuning a glaciological model in series with a glacial isostatic adjustment and relative sea-level model. The study builds upon the work of Simpson and others (2009) through four main extensions: (1) a larger constraint database consisting of relative sea-level, ice height and extent data; model improvements to the (2) climate and (3) sea-level forcing components; (4) accounting for uncertainties in non-Greenland ice. The research was conducted primarily to address data-model misfits and to quantify inherent model uncertainties with the Earth structure and non-Greenland ice. Our new model (termed Huy3) fits the majority of observations and is characterized by a number of defining features. During the LGM, the ice sheet had an excess of 4.7 m ice-equivalent sea level, which reached a maximum volume of 5.1 m ice-equivalent sea level at 16.5 ka BP. Modelled retreat of ice from the continental shelf progressed at different rates and timings in different sectors. Southwest and southeast Greenland began to retreat from the continental shelf by ~16 to 14 ka BP, thus responding in part to the Bølling-Allerød warm event (14 ka BP); subsequently ice at the southern tip of Greenland readvanced during the Younger Dryas cold event. In northern Greenland the ice retreated rapidly from the continental shelf upon the climatic recovery out of the Younger Dryas to present-day conditions. Upon entering the Holocene (11.7 ka BP), the ice sheet soon became land-based. During the Holocene Thermal Maximum (HTM; 9–5 ka BP), air temperatures across Greenland were marginally higher than those at present and the GrIS margin retreated inland of its present-day southwest position by 40–60 km at 4 ka BP, which produced a deficit volume of 0.16 m ice-equivalent sea level relative to present. In response to the HTM warmth, our optimal model reconstruction lost mass at a maximum centennial rate of 103.4 Gt a–1. Our results suggest that remaining data-model discrepancies are associated with missing physics and subgrid processes of the glaciological model, uncertainties in the climate forcing, lateral Earth structure, and non-Greenland ice (particularly the North American component).


Analysis of front position changes of a tidewater glacier and its control mechanisms


Corresponding author: Jaime Otero

Corresponding author e-mail: jaime.otero@upm.es

A thermomechanical flow model (Elmer) that solves the full Stokes system is used to model the time evolution of Hansbreen Glacier, a grounded tidewater glacier in southern Spitsbergen, Svalbard. The model setup considers climate and oceanic forcing and couples continuum damage mechanics with the equations of viscous deformation of glacier ice. We introduce a scalar damage variable that quantifies the loss of load-bearing surface area due to fractures and that feeds back with ice viscosity to represent fracture-induced softening. The damage variable is inferred by numerical modelling of Hansbreen’s rheology from a complete velocity field derived from remote sensing. Furthermore, the model is forced by atmospheric input fields derived from the European Arctic Reanalysis dataset (TU Berlin) by applying a high-resolution dynamical downscaling using the polar WRF model with updated static geographical fields from airborne and satellite imagery. Finally the model’s ability to reproduce seasonal cycles of advance and retreat of the glacier front position is investigated by examining different control mechanisms, such as back-pressure at the calving front, basal water pressure and sea-ice concentration.


BRITICE-CHRONO: constraining rates and style of marine-influenced ice-sheet decay to provide a data-rich playground for ice-sheet modellers


Corresponding author: Chris Clark

Corresponding author e-mail: c.clark@sheffield.ac.uk

Uncertainty exists regarding the fate of the Antarctic and Greenland ice sheets. If we want to know more about the mechanisms and rate of change of shrinking ice sheets, then why not examine an ice sheet that has fully disappeared and track its retreat through time? If achieved in enough detail such information could become a data-rich playground for improving the next breed of numerical ice-sheet models to be used in ice and sea-level forecasting. We regard that the last British–Irish ice sheet is a good target for this work, on account of its small size, large proportion as marine-based, density of existing information and with its numerous researchers already investigating it. BRITICE-CHRONO is a large (>45 researchers) NERC-funded consortium project (2013–2017) comprising Quaternary scientists and glaciologists who are searching the sea floor around Britain and Ireland and parts of the landmass in order to find and extract samples of sand, rock and organic matter that can be dated (OSL; Cosmogenic; 14C) to reveal the timing and rate of change of the collapsing British–Irish ice sheet. The purpose is to produce a high-resolution dataset on the demise on an ice sheet – from the continental shelf edge and across the marine to terrestrial transition; retreat of the order of 200 km over 10 000 years. Some 800 new date assessments will be added to those that already exist. Data on retreat will be collected by focusing on eight transects running from the continental shelf edge to a short distance (10s km) onshore and acquiring marine and terrestrial samples for geochronometric dating. The project includes funding for 587 radiocarbon, 140 OSL and 158 TCN samples for surface exposure dating; with sampling accomplished by two research cruises and 16 fieldwork campaigns. Results will reveal the timing and rate of change of ice margin recession for each transect, and combined with existing landform and dating databases, will be used to build an ice-sheet-wide empirical reconstruction of retreat incorporating Bayesian analysis to assess uncertainty. BRITICE-CHRONO is advised by its international advisory board of ice-sheet modellers including Flowers, Hubbard, Milne, Pollard, Ritz, Rutt, and Vieli, and led by Hindmarsh. We invite and encourage further ice-sheet modellers to use our data for modelling experiments and in particular to explore the role of bed topography, sea level and tidal regimes in modulating ice retreat.


Development towards a full Bayesian calibration of a 3-D glacial systems model of the Antarctic ice sheet over the last glacial cycle


Corresponding author: Benoit Lecavalier

Corresponding author e-mail: b.lecavalier@mun.ca

How did the Antarctic ice sheet (AIS) evolve over the last glacial cycle? How much do current uncertainties in the thermomechanical present-day state of the ice sheet affect projections of its future evolution? Without the associated quantification of uncertainties, answers to these questions have little value. A developing technique for explicit uncertainty quantification of glacial systems is large-ensemble Bayesian calibration of models against large observational datasets. The foundation for a Bayesian calibration of a 3-D glacial systems model (GSM) for Antarctica has recently been completed. Bayesian calibration thoroughly samples model uncertainties against fits to observational data through Markov Chain Monte Carlo methods using Bayesian artificial neural network emulators of the full GSM. For the first time, this methodology will generate a probability distribution for the AIS deglaciation with explicit and well-defined confidence intervals. Past work has shown the GSM to have likely inadequate range of grounding line migration in certain sectors as well as persistent ice thickness biases in topographically complex regions. To advance towards full calibration, these deficiencies will be addressed through the following. First, basal drag representation will be improved. This will include improved subgrid treatment of the thermomechanical impacts of high basal roughness, examination of the impact of inclusion of fully coupled basal hydrology, and re-evaluation of uncertainties in basal drag parametric representation for regions that are presently marine. Second, an expanded climate forcing using statistical correction of simplified climate models will be added to better capture the uncertainty in past climate. Thirdly, the impact of past changes in ocean temperature on sub-ice-shelf melt will be explicitly incorporated. Finally, the calibration will also incorporate uncertainties in Earth rheology in the context of isostatic adjustment. We will outline the improvements currently being implemented, solicit further data constraints, and outline the remaining steps towards full calibration of the model. This research will improve our understanding of the present-day thermomechanical state of the AIS and its contemporary mass balance through the re-interpretation of geodetic data. By also running the calibrated GSM forward in time, a probability distribution for the future evolution of the AIS will be generated.


Modeling past and future ice retreat in Antarctic subglacial basins


Corresponding author: David Pollard

Corresponding author e-mail: pollard@essc.psu.edu

Geological data indicate that global mean sea level has fluctuated on O(104 to 105 year) timescales during the last ~25 million years. Peak levels are uncertain, but some estimated high stands are ~20 m or more above modern, for instance during the mid-Pliocene. If correct, this implies substantial variations in the size of the East Antarctic ice sheet (EAIS). However, climate and ice-sheet models have not been able to simulate significant EAIS retreat from continental size, given low proxy atmospheric CO2 levels during this time. Here, we use a continental Antarctic ice sheet model with two mechanisms based on previous studies and observations: (1) structural failure of large tidewater ice cliffs, and (2) enhanced ice-shelf calving due to meltwater drainage into crevasses. With climate forcing representing Pliocene warm periods, the two mechanisms accelerate West Antarctic collapse and produce retreat in major East Antarctic basins. Equivalent global mean sea-level rise is ~15 m, in better agreement with past sea-level data. The model is applied to specific past periods and to the long-term (100s to 1000s years) future, in which the ice sheet is found to be considerably more vulnerable to climate warming than previously modeled.


Effect of surface mass-balance forcing errors on uncertainty in predictions of Greenland ice sheet contributions to sea-level rise


Corresponding author: Gail Gutowski

Corresponding author e-mail: gail.gutowski@utexas.edu

Ice-sheet models are used to simulate ice-sheet dynamical and mass-balance contributions to sea-level rise (SLR). Within a coupled-climate model, ice-sheet boundary conditions are subject to biases from other components of the Earth system due to imperfect knowledge of the climate system. The initialization of an ice-sheet model can therefore involve significant errors that affect projections of SLR. Errors in the specification of boundary conditions (bed topography, geothermal fluxes, and surface mass balance and flow physics parameters) may be amplified or compensate for each other in an effort to simulate an ice sheet that matches observations. However, this may result in a ‘realistic’ modeled ice sheet for the wrong physical reasons. We explore how the size of errors in surface mass-balance (SMB) model forcing impacts uncertainty in projections of SLR contributions from the Greenland ice sheet. We are interested in quantifying the relationship between the size of errors in SMB and resulting SLR scatter. We quantify uncertainties associated with imperfect simulation of the pre-industrial surface mass balance using the Community Earth System Model (CESM) that includes an energy-balance surface mass-balance formulation as part of its land surface model. The feedbacks between SMB and sea level are estimated using the Community Ice Sheet Model (CISM), which is based on the shallow-ice approximation. We found significant scatter in projections of SLR resulting from errors in SMB forcing. Errors in SMB are propagated to the initialization of CISM primarily through a surface elevation feedback. In a warming world, we expected a positive feedback between surface elevation and SMB errors in which lower ice sheets melt more rapidly. However we found the opposite result for much of the northern portion of the ice sheet. In these cases, increased SMB corresponded to decreasing elevation, acting to stabilize surface elevation.


Structural and dynamic changes of the Wilkins Ice Shelf, Antarctic Peninsula, derived from SAR satellite data

Melanie RANKL, Matthias BRAUN, Angelika HUMBERT, Ralf MÜLLER, Carolin PLATE, Veit HELM

Corresponding author: Melanie Rankl

Corresponding author e-mail: melanie.rankl@fau.de

The Wilkins Ice Shelf (WIS) has shown considerable ice-front retreat since the 1990s. This retreat includes various break-up events, such as in 2008 (Feb: 425 km2; May: 160 km2; Jul: 1220 km2) and in 2009 (790 km2). The break-up occurred in winter and summer with contrasting surface conditions, which indicates potentially different mechanisms for the various break-up events. The WIS shows quite specific peculiarities like a large number of ice rises, highly variable ice thicknesses across the ice shelf, and only limited nourishing by direct inflow from tributary glaciers. In the present study we investigated dynamic changes leading to disintegration and break-up of the WIS. We analyzed satellite data (especially SAR) to reveal changes in glaciological structures, e.g. fractures and shear margins, the position of the grounding line, changes of frontal positions, ice surface velocities and ice thickness variations. Surface velocities of the ice shelf and its tributary glaciers were derived from different SAR sensors (TerraSAR-X, ALOS PALSAR, ERS, Envisat) and dates (1992–1996, 2007–2010) using intensity tracking. Time periods before and after break-up events were covered in order to study impacts of changes in load resulting from the new ice-front positions. We further exploit data of the TanDEM-X mission (2012–2013) to generate digital elevation models (DEM) of the area, where comprehensive high-resolution elevation data currently are hardly available. The DEMs were linked to different altimeter measurements (ICESat, CryoSat, NASA IceBridge) and ice thickness estimates of the ice shelf were carried out. Subsequently, the remote-sensing products fed into fracture and ice dynamic modeling.


Reducing uncertainties in Antarctic ice sheet mass loss projections


Corresponding author: Frank Pattyn

Corresponding author e-mail: fpattyn@ulb.ac.be

Climate model projections are often aggregated into multi-model averages of all models participating in an intercomparison project, such as CMIP. Several authors have questioned whether this is the best use of the information and whether the community is ready to move beyond the ‘one-model-one-vote’ approach, based on the intrinsic quality of each of the models. Ice-sheet models are not as far developed as climate or ocean models. Many of these models are still struggling over basic thermomechanical issues related to ice deformation, while at the same time disproportionate efforts are made on the interaction with the atmosphere, basal hydrology, sliding, sediment deformation, ice/ocean interaction, calving, grounding-line migration, etc. We can therefore reasonably question whether averaging all model results at equal weight is the best strategy and to which extent coupling of ice-sheet models that are lacking the representation of crucial physical processes, to other components of the climate system could lead to spurious errors. We now have tools available to test parts of the response of marine ice-sheet models to perturbations of climatic and/or oceanic origin. Results show that the type of model as well as the way boundary conditions are implemented greatly affects the response of each ice-sheet system. Based on MISMIP experimental output as well as the experimental response of Antarctic glaciers and drainage basins to ocean perturbations, we provide a guidance for the evaluation of model response to perturbations on century timescales.


Basal temperature calculations of the Greenland ice sheet


Corresponding author: Brice Van Liefferinge

Corresponding author e-mail: bvlieffe@ulb.ac.be

Of all boundary conditions operating on Greenland ice sheet dynamics, the basal conditions are amongst the least well understood. Using a thermomechanical model comparable to the one employed by Pattyn (2010) and Van Liefferinge and Pattyn (2013), we propose an approach to calculate the basal temperature of the Greenland ice sheet using datasets recently made available. Geometrical input parameters are bed topography and surface topography, and ice thickness. Physical input parameters are ice surface temperatures, surface velocities obtained from satellite interferometry further constrained by an ice-sheet model, and accumulation rate. Estimates of geothermal heat flux, probably one of the poorly known constraints, are employed to make ensemble comparisons. In view of the high accumulation rates over the Greenland ice sheet (compared with Antarctica), englacial temperatures are highly influenced by the surface temperature history, especially on glacial-interglacial timescales. Therefore, the method of Pattyn (2010) was adapted to take into account transient temperature change. As done previously, direct englacial temperatre measurements from deep drillings (NEEM, GRIP, NGRIP, GISP, Camp Century and Dye 3) were used to correct modelled estimates through a simple assimilation technique.


Mass loss of the Greenland ice sheet since the Little Ice Age: implications for sea level

Kristian Kjellerup KJELDSEN, Kurt KJÆR, Anders A. BJØRK, Shfaqat Abbas KHAN, Niels Jakup KORSGAARD, Nicolaj Krog LARSEN, Antony J. LONG, Sarah A. WOODROFFE, Glenn A. MILNE, Jonathan BAMBER, Michiel VAN DEN BROEKE

Corresponding author: Kristian Kjellerup Kjeldsen

Corresponding author e-mail: kkjeldsen@snm.ku.dk

Records from globally distributed tide gauges suggest a global mean sea-level rise (SLR) of 15–20 cm during the 20th century. Quantifying the contribution from ice sheets and glaciers to past sea-level change is of great value for understanding sea-level projections into the 21st century. Estimates of the past Greenland ice sheet (GrIS) contribution to SLR have been derived using a number of different approaches, e.g. surface mass-balance (SMB) calculations combined with estimates of ice discharge found by correlating SMB anomalies and calving rates. Here, we adopt a novel geometric approach to determine the post-Little Ice Age (LIA) mass loss of the entire GrIS. We use high quality aerial stereo photogrammetric imagery recorded between 1978 and 1987 to map morphological features such as trimlines (boundary between freshly eroded and non-eroded bedrock) and end moraines marking the ice extent of the LIA, which thereby enables us to obtain vertical point-based differences associated with changes in ice extent. These point measurements are combined with contemporary ice surface differences derived using NASA’s Airborne Topographic Mapper (ATM) from 2002 to 2010, NASA’s Ice, Cloud and land Elevation Satellite (ICESat) from 2003 to 2009, and NASA’s Land, Vegetation and Ice Sensor (LVIS) from 2010, to estimate mass loss throughout the 20th and early 21st century. We present mass-balance estimates of the GrIS since retreat commence from the maximum extent of the LIA to 2010 derived for three intervals, LIAmax (1900)–1978/87, 1978/87–2003 and 2003–2010. Results suggest that despite highly spatially and temporally variable post-LIA mass loss, the total mass loss and thus the contribution from the GrIS to global SLR has accelerated significantly during the 20th century into the 21st century.


Improving the parameterization of changing glacier geometry in surface mass-balance models


Corresponding author: David Parkes

Corresponding author e-mail: david.parkes@uibk.ac.at

Degree-day mass-balance models show considerable potential for recreating past glacier mass changes and predicting the evolution of modern glaciers, but these models show high sensitivity to changes in glacier shape and size over the course of the modelling period. In particular, variations in the terminus elevation of a glacier, combined with the expected temperature lapse rate, result in large changes in the positive degree-day profile over the glacier’s surface and therefore strongly impact ablation rates. Models that convert changes in volume, determined from the surface mass balance and glacier area, into changes in area, length and terminus elevation using a simple scaling formula can introduce errors in the terminus elevation, which drive substantial errors in the overall mass balance for subsequent years. Glacier thickness profiles calculated from ablation and accumulation profiles over the glacier surface and flow rate profiles determined by surface slope offer the opportunity to understand glacier geometry in a much more localized way, at the cost of requiring considerably more information about the glacier surface. Can mass-balance models using this type of thickness profile calculation be used to generate results that simultaneously capture mass-balance time series and glacier geometry evolution with similar accuracy, and can the information from these more complex models be used to inform and parameterize geometry evolution in models with more simple inputs?


Multi-technical monitoring of the Argentière glacier surface velocity variability

Lionel BENOIT, Ha-Thai PHAM, Amaury DEHECQ, Emmanuel TROUVÉ, Flavien VERNIER, Luc MOREAU, Denis LOMBARDI, Olivier MARTIN, Christian THOM, Pierre BRIOLE

Corresponding author: Lionel Benoit

Corresponding author e-mail: lionel.benoit@ign.fr

The Argentière glacier is a 9 km long temperate glacier covering 19 km2 situated in the Mont Blanc area (French Alps). The high temporal and spatial resolution monitoring of its dynamic is a challenging metrological topic and can lead to crucial information about the ice flow. Between mid-September and mid-November 2013, we monitored a small (400 m × 600 m) area of the glacier about 1 km upstream of the Lognan icefall. A multi-technical survey was carried out involving 13 mono-frequency GPS receivers called ‘Geocubes’, four short-period seismometers, two ground-based automatic digital cameras and synthetic aperture radar images from the TerraSAR-X satellite. The processing of GPS data allows us to estimate epoch by epoch the coordinates of each Geocube with a centimetric accuracy. Resulting positions can be used to monitor the variability of the flow with high accuracy and provide ground-control points for the proximal and spaceborne imagery. GPS measurements are available only at 11 points of the glacier, but series of ground-based stereoscopic pictures can be used to map the glacier flow as a whole, including crack areas or icefalls. In addition to the in situ measurements, three TerraSAR-X images acquired at an interval of 11 days each were used to estimate the surface velocity field of the whole glacier. Furthermore the displacement measurements are complemented by seismic measurements providing glacier-flow-induced seismicity. The combination of these multi-source data leads to a high time and high space resolution picture of the ice surface and subsurface. Two aspects of the glacier flow are investigated from this dataset. First the deformation of the ice surface induced by the flow is studied. A higher velocity is recorded at the center of the glacier and leads to internal constraints. The strain tensor analysis shows a good consistency between the main strain axis and the orientation of the cracks on both sides of the glacier. Then the temporal flow variability is investigated and combined with ice-related seismicity. We derived from our dataset an average ice surface velocity at the center of the glacier of 15 cm d–1 with peaks reaching 25 cm d–1 a few hours to 1 day after rainfall periods. During these ice mass accelerations, we are able to monitor the basal displacement near the Lognan icefall within a tunnel dug beneath the glacier. Thus, a simultaneous monitoring of surface and basal velocity variations is available.


Mertz Ice Shelf dynamics in the last 20 years using satellite data

Xianwei WANG, David HOLLAND, Xiao CHENG

Corresponding author: Xianwei Wang

Corresponding author e-mail: wangxianwei0304@163.com

In February 2010, the Mertz Ice Tongue collapsed and generated a giant iceberg, which attracts global focus on it. In this research, ICESat/GLAS data from 2003 to 2009 were used to extract freeboard map and freeboard changing rate of the ice tongue. In 2009, the freeboard varied from 23 to 59 m, with 41 m as an average. Freeboard changing rate from 2003 to 2009 was calculated as about –0.07 m a–1, with 1.06 m as standard deviation error. Ice-shelf boundary was extracted with Landsat TM/ETM+ data from 1989 to 2013. When taking the grounding line as the boundary inside, the Mertz Ice Shelf shows an increased changing rate (36.86 km2a–1 compared with 35.32 km2a–1 before 2008) in area after the calving in 2010. Additionally, the rift on both sides of the Mertz Ice Shelf was extracted. The area of large rift in the right side along the ice shelf advancing showed an increasing trend (4.05 km2 to 19.4 km2) from 1989 to 2003 and a decreasing trend (19.05 km2 to 17.6 km2) from 2003 to 2009. However, the large rift in the left side along ice shelf advancing occurred at about 2002 and the area increased to 11.38 km2 at the end of 2009. Also after 2002, the ice front of the Mertz Ice Shelf changed moving direction by about 40°. Deep crevasses on the surface and expansion of a central large rift had made Mertz Ice Shelf fragile and it disintegrated after collision with an iceberg. By reversing ice bottom data at November 2002 and comparing with sea-floor data from bedmap2, initial results show a grounded ice front and the direction change of the ice front was caused by a hump of the sea floor.


Continuum and discrete models for calving and fracturing in glaciers


Corresponding author: Thomas Zwinger

Corresponding author e-mail: zwinger@csc.fi

Contrary to the long-time behaviour of ice that is governed by continuous flow phenomena, with the ice deformation obeying a Norton-Hoff-type strongly non-linear rheological law, fracturing of ice and (as a consequence of it) calving at the glacier tongue is based on inherent discontinuous processes that project themselves as being non-deterministic on the continuum scale, as they are governed by small-scale flaws in the ice. In order to understand the physics behind fracturing, we need to deploy models that deviate from the traditional continuum approaches in numerical glaciology. To this end we present a discrete particle model that aims to reproduce the short-time viscous but more important elastic and brittle behaviour of ice and its latest implementations to three-dimensional problems, such as tidewater glaciers. Such models, necessary as they are to give insight into the mechanics of fracturing and collapse, have the inherent disadvantage to be limited in their spatial as well as temporal range of computations. As a consequence, their findings have to be (re-)introduced into continuum models. Along simplified but three-dimensional synthetic examples we strive to discuss the possibilities to come after that demand.


Could the non-linear creep behaviour observed at low stresses be associated with the yield stress of ice?


Corresponding author: Fabien Gillet-Chaulet

Corresponding author e-mail: gillet-chaulet@lgge.obs.ujf-grenoble.fr

The formation of isochrone arches beneath divides (Raymond bumps) is the result of strong variations of the vertical velocity profile across ice divides known as the Raymond effect. This effect is expected only with a non-linear flow law. Indeed, a non-linear flow law with a stress exponent n higher than 4 has been deduced from flow modelling and in situ measurements of the Raymond effect at Roosevelt Island, Antarctica, and at Summit, Greenland. However, reliable laboratory experiments support a flow law with n < 2 at relatively low stresses. These contradictory results on the ice viscosity at low stresses deserve discussion since, depending on the value of the stress exponent, the viscosity could vary by several orders of magnitude at low stresses. Newtonian diffusional flow is often evoked in condition of low stresses. From theoretical considerations, we show that the concentration gradient of point defects induced by low applied stresses is not significant in comparison with the free energy formation of these defects. In a first stage we then conclude that Newtonian diffusional flow does not occur in polar ice sheets even for the lowest stresses encountered under ice divides. In a second stage we test the hypothesis of the presence of a threshold stress for dislocation glide, which would induce a high apparent stress exponent in these low stress conditions. This ‘yield’ stress could be the result of a critical stress for dislocation nucleation or a jamming transition induced by long-range dislocation interactions. To do so, new phase-sensitive radar measurements of vertical strain rates, associated with flow modelling at Berkner Island, Antarctica, are analysed in order to show how the existence of this threshold stress for dislocation glide could result in a strongly non-linear behaviour at very low stresses, leading to high values for an ‘apparent’ stress exponent.


A modular software package for integrating multiple observation sources to determine ice mass trends in space and time


Corresponding author: Andrew Zammit-Mangion

Corresponding author e-mail: ggazm@bristol.ac.uk

Combining a dataset with a numerical model output has become routine when assessing an ice sheet’s contribution to sea-level rise. However, there are many datasets and numerical models that can be employed. As a consequence, a wide range of sea-level rise estimates have been produced using markedly different methodologies, data, approximation methods and model assumptions. Current attempts to reconcile these estimates using simple combination methods are unsatisfactory as common sources of errors across different methodologies may not be accurately quantified. A more rigorous way to solve the problem is through a statistical framework able to combine information from all datasets and models concurrently. Data-model fusion has the ability to avoid double counting and common sources of error, and has recently shown to be an effective way for computing reliable sea-level rise contributions from ice sheets. The difficulty, however, is that constructing a data assimilation framework for such diverse data sources is cumbersome, requires a resident expert in computational statistics and is frequently not amenable to new data types and models. To address these issues, we have designed a user-friendly software package specifically designed to estimate regionally differentiated mass trends from ice sheets. The user is solely tasked with loading pre-processed data into objects and linking them up with the physical processes being measured (e.g. GRACE = snow + ice + GIA). The software has the ability to: (1) separate total mass contribution into surface mass balance, GIA and ice dynamics contributions using state-of-the-art statistical methods; (2) provide time-evolving solutions for the individual processes; (3) use multiple datasets (both future and past) concurrently with different observation characteristics (e.g. GRACE, ICESat GPS simultaneously); (4) provide uncertainties on all estimated quantities; and (5) incorporate prior information from models with various levels of confidence (e.g. solely spatial/temporal scales). With this software the user will be able to quickly analyse the effect of a specific dataset, numerical model or assumption on the mass budget estimate, as well as incorporate new observations with ease as they become available. Here, we briefly discuss the underlying methods and show how the software can be used to provide a time-evolving mass budget estimate for the Antarctic ice sheet from multiple, spatially heterogeneous datasets.


Fabric measurement along the NEEM ice core, Greenland, and comparison with GRIP and NGRIP ice cores


Corresponding author: Maurine Montagnat

Corresponding author e-mail: montagnat@lgge.obs.ujf-grenoble.fr

Fabric (distribution of crystallographic orientations) profile along the full NEEM ice core, Greenland, is presented in this work. Data were measured in the field by an Automatic Ice Texture Analyzer every 10 m, from 33 m down to 2461 m depth. The fabric evolves from a slightly anisotropic fabric at the top, toward a strong single maximum at about 2300 m, which is typical of a deformation pattern mostly driven by uniaxial compression and simple shearing. A sharp increase in the fabric strengthening is observed at the Holocene to Wisconsin climatic transition. A similar strengthening, toward an anisotropic single maximum-type fabric, has been observed in several ice cores from Greenland and Antarctica, and can be attributed to a positive feedback between changes in ice viscosity at the climatic transition, and the impact of a shear component of stress. Centimeter-scale abrupt texture (fabric and microstructure) variations are observed in the bottom part of the core. Their positions are in good agreement with the folding hypothesis used for a climatic reconstruction by Dahl-Jensen and others (2013). Comparison is made to two others ice cores drilled along the same ridge: the GRIP ice core drilled at the summit of the ice sheet, and the NorthGRIP ice core, drilled 325 km to the NNW of the summit along the ridge, and 365 km upstream from NEEM. The fabric profile clearly reflects the increase in shear deformation when moving NW along the ridge from GRIP to NorthGRIP and NEEM. The difference in fabric profiles between NEEM and NorthGRIP also evidences a stronger lateral extension associated with a sharper ridge at NorthGRIP.


Nye–Rothlisberger channels with finite ice depth and open channel flow

Geoffrey EVATT

Corresponding author: Geoffrey Evatt

Corresponding author e-mail: geoffrey.evatt@manchester.ac.uk

The study of how water flows through subglacial channels has had many practical applications in glaciology, including helping to understand the dynamics of jokulhlaups and glacier mass balances. The theoretical basis of these subglacial channels relies upon the work of Rothlisberger and Nye, who considered the channel’s dynamics to be governed by a mix between water friction widening the tunnel’s walls and the creep closure of the surrounding ice. Whilst their modelling is evidently well-constructed, there have been two aspects of their work that have gone undeveloped. The first is the consideration of a finite glacier depth, instead of the assumption of a infinite glacier depth. The second is the allowance of a region of open channel flow, so that the water may transition from a region of closed channel flow to one where the water is exposed to the atmosphere. We show how these two extra considerations can be incorporated within contemporary models of subglacial water channels and how they improve our estimates for water discharge levels from underneath glaciers.


Calving rates and controls of Svalbard tidewater glaciers

Adrian LUCKMAN, Doug BENN, Suzanne BEVAN

Corresponding author: Adrian Luckman

Corresponding author e-mail: A.Luckman@Swansea.ac.uk

Controls on iceberg calving remain poorly understood and inadequately represented in prognostic ice-sheet models. This is partly because of the complexity of the issue, which involves ice flow and fracture processes as well as tidewater interactions, but also because of the lack of empirical observations at appropriate temporal and spatial scales. Many have observed occasional sometimes dramatic calving events, but systematic measurements of calving rates are very sparse in both time and space. Here we present calving rates for a number of glaciers in Svalbard based on analysis of time series of TerraSAR-X SAR data of a full seasonal cycle in 2013. The high spatial resolution (~2 m), short orbital cycle (11 days) and repeat reliability (active microwave imaging) of these data allow them to be used both for ice-front monitoring and velocity measurement, which may be combined to give linear calving rate, and volume calving rate where water depth is known. Calving rates are compared between fast-flowing glaciers, surge-type glaciers in active and quiescent phases, and glaciers discharging into different fjord systems. Results demonstrate a highly seasonal calving cycle in sync with ocean temperatures, and a remarkable similarity in calving rates between glaciers of contrasting dynamics. Svalbard calving may not be a good model for large Greenland outlet glaciers, but may be widely representative of calving mechanisms elsewhere.


The Greenland firn layer: explaining the presence of perennial aquifers, and modelling firn response to a warming climate


Corresponding author: Peter Kuipers Munneke

Corresponding author e-mail: p.kuipersmunneke@uu.nl

A significant part of the surface meltwater on the Greenland ice sheet refreezes in the firn rather than running off into the ocean. The firn thus acts as an important buffer between surface melt and sea-level rise. As a spectacular example of this buffer role, recent observations have shown that the firn layer features perennial subsurface bodies of liquid water at a depth of 10–50 m below the surface, mainly in southeast and southwest Greenland. Using a model with basic firn hydrology, thermodynamics and compaction, we are able to explain the formation of these perennial firn aquifers: a high accumulation rate is needed to provide pore space for water storage at sufficient depth to protect it from the winter cold. Low-accumulation sites cannot provide sufficient deep pore space to store liquid water. With the same firn model, we study the transient behaviour of both dry and wet firn as a reaction to a warming climate. We find that the response time of the firn decreases with increasing accumulation rate, and with increasing meltwater production at the surface. Thus, the currently observed changes in the firn layer in the ice-sheet interior are due to snowfall and melt variability of a much longer period in the past than the changes we currently observe in the lower-elevation percolation zones.


Seasonal evolution of basal friction coefficients in Kronebreen, Svalbard, using Elmer/ICE


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 in the archipelago. Recent observations have shown a sudden start of retreat pattern that makes it a fairly interesting study site in terms of sea-level rise and the contribution of glaciers. 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. Subglacial hydrology and sediment deformation might therefore have a non-negligible impact on the dynamics of the glacier. In order to understand the evolution of basal sliding through time, we use the inverse method to infer the spatially and temporally varying basal friction coefficient β(x,y) from surface ice-velocity observations. The glacier is modeled using the finite-element method using an anisotropic mesh refinement. We are 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). The latter are available every 11 days and can be used for the inversion. We are then able to invert for a complete season and analyze the seasonal evolution of these coefficients and relate it to other parameters influencing motion.


Parameter ranges permitting self-sustained grounding line oscillations in a coupled ice-flow–earth deformation model


Corresponding author: C. Rosie Williams

Corresponding author e-mail: chll1@bas.ac.uk

The grounding line, where the grounded ice meets the ice shelf, is known to vary rapidly and dynamically in response to external forcing such as atmospheric or oceanic warming. Additionally, the solid earth beneath the ice sheet can respond to changes in the ice loading caused by the advance and retreat of the ice sheet over time. There have been relatively few investigations of coupling between grounding-line motion and such glacio-isostatic adjustment. Here we investigate the dynamics of an ice-sheet grounding line by coupling a one-dimensional shallow-ice-type model in which the interior ice flux is caused by sliding and the grounding line flux is parameterized with a self-gravitating visco-elastic relaxation model of an axisymmetric spherical earth. On exploring a range of different earth and ice parameters we find that grounding line position is not very sensitive to upper and lower mantle viscosities but is very sensitive to the lithosphere thickness. We show that self-sustained oscillations in grounding line position with cycles of 10–25 thousand years are found when the bed overdeepens for sufficiently high accumulation, or if the initial bed was sufficiently shallow or the lithosphere sufficiently thin. These oscillations are not present when the ice model alone is run, showing that bed deformation is an important aspect of grounding line motion in models that are run over thousands of years. The existence of these oscillations depends strongly on the inclusion of self-gravitation, corresponding to other investigations that show the strong effect of this on marine ice-sheet dynamics.


Evaluating the impact of relative sea-level change on ice-sheet dynamics


Corresponding author: Pippa Whitehouse

Corresponding author e-mail: pippa.whitehouse@durham.ac.uk

The water depth of the surrounding ocean is a key factor in determining the dynamics of a marine-based ice sheet. In this study we outline two key ways in which sea-level changes impact ice-sheet dynamics, and we highlight potential errors that can be made if sea-level changes are not consistently modelled in parallel with the evolution of a marine-based ice sheet. Changes in relative sea level, i.e. water depth, will influence grounding line dynamics, both in terms of the location of the grounding line and the flux of ice across the grounding line. We explore the implications of considering realistic, spatially variable relative sea-level changes – derived using a glacial isostatic adjustment (GIA) model – as opposed to uniform, or ‘eustatic’, sea-level changes when determining the likely configuration of two key Antarctic outlet glaciers during the LGM. In particular, we highlight the different sea-level change experienced by East and West Antarctica due to rotational feedback. Secondly, changes in water depth will determine which portions of the ice sheet are grounded or floating. In the context of a GIA model, this information is needed to determine the magnitude of the ice and ocean load changes that are applied to the solid Earth. We demonstrate that if the evolving topography is incorrectly defined, particularly across ice-shelf regions, errors on the order of 10 mm a–1 can be made when predicting present-day uplift rates due to past ice mass changes. The correct modelling of water depth change is also necessary to study the evolution of ice rises, whose presence will impact the stress regime of an ice shelf and hence the dynamics of the upstream ice sheet, and to determine former ice-sheet thicknesses via the interpretation of iceberg scours.


Simulations of the effect of mass-balance forcing on Midtre Lovénbreen, Svalbard, with a full-Stokes glacier model


Corresponding author: Ilona Välisuo

Corresponding author e-mail: ilona.valisuo@fmi.fi

Midtre Lovénbreen (MLB) (78.53° N, 12.04° S) is a small (~5 km2) Alpine-type valley glacier in northwest Spitsbergen. Mass-balance measurements made since the 1960s show that the glacier mass balance has been nearly always negative, and it has been continuously retreating since about the 1920s. Previously MLB has been modelled with the thermomechanically coupled full-Stokes glacier model Elmer/Ice, with results that agreed well with the observed state of the glacier. In this study we build on the previous modelling work, now investigating the interaction between the atmosphere and the glacier, i.e. the effect of the 20th century mass-balance forcing on the glacier. We seek to (1) reconstruct mass balance from the simulation of different time periods from 1936 to 2005 and (2) compare the reconstructed mass balance to the in situ observations along the center line of the glacier. Surface digital elevation models (DEMs) for various years from 1936 onwards are used to perform steady-state simulations. Mass balance is reconstructed by solving the surface elevation changes and the mean flow velocities between two separate snapshots. This allows us to collect information on the spatial distribution of the surface mass balance. For validation we have the in situ measurements of surface mass balance from 1967 onwards. With better knowledge of the accuracy of simulated mass balance over the past decades, it will be possible to better estimate future mass balance of MLB based on prognostic model simulations.


Beyond back-stress: model experiments investigating the role of gradients in longitudinal stress modulating glacier flow during periods of rapid retreat

Faezeh NICK, Cornelis VAN DER VEEN, Doug BENN, Leigh STEARNS

Corresponding author: Faezeh Nick

Corresponding author e-mail: faezeh.nick@unis.no

Over the last two decades, many Greenland outlet glaciers have experienced rapid change, including dramatic thinning and doubling of glacier speed. Some of these changes appear to have been short-lived, and glacier acceleration was followed by slowdown. The physical processes responsible for these observed changes remain under debate, with one side arguing for terminal forcing initiating up-glacier changes, while others have proposed that glacier speed-up and thinning resulted from processes acting on the grounded lower reaches of these outlet glaciers (such as weakening of the ice in the lateral margins). To identify possible causal mechanisms, two approaches have been followed. In the first approach, the balance of forces acting on outlet glaciers is considered, using measurements of surface velocity and glacier geometry. One such study focused on Jakobshavn Isbræ and concluded that the role of longitudinal stress gradients in glacier flow is small compared with basal and lateral drag. However, that study considered only three time epochs before and during the speed-up of this glacier, and it is not clear whether this conclusion remains valid if the more continuous record of glacier change is considered. The second approach to elucidate processes that may lead to rapid glacier changes is to apply numerical ice-flow models and evaluate which changes in model parameters (back-stress, enhanced sliding, reduction in surface mass balance, increased calving resulting from increased surface ablation, etc.) lead to predicted behavior similar to the observed glacier changes. Several such numerical studies have successfully modeled rapid changes observed on outlet glaciers, but given the multitude of physical processes included, it remains unclear which processes are important. The question we address is quite straightforward: how important are longitudinal stress gradients in modulating glacier flow during periods of rapid change? To address this question, we conducted a series of model experiments comparing results from two flowline models; in the first model the driving stress is balanced by resistive forces associated by gradients in longitudinal stress, lateral drag and basal drag, while in the second model driving stress is balanced entirely by resistance from lateral and basal drags. We applied the model comparison to Jakobshavn Isbræ and other Greenland outlet glaciers for which detailed records of change are available.


Chain-reaction drainage of supraglacial lakes triggered the Larsen B ice shelf break-up: a continuation

Alison BANWELL, Douglas MacAYEAL

Corresponding author: Alison Banwell

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

The collapse of ice shelves of the Antarctic Peninsula during the past few decades has resulted in increased ice mass loss from tributary glaciers due to removal of buttressing ice shelves. Most notably, the explosive disintegration of the Larsen B ice shelf in March 2002 has led to continued acceleration and thinning of tributary glaciers. Using thin elastic plate theory, our previous work showed that Larsen B’s explosive disintegration may have been due to drainage of ~2750 supraglacial lakes through a chain-reaction style process. In addition to demonstrating that the drainage of a single ‘starter’ lake can cause multiple fractures able to hundreds of surrounding lakes, we also showed that this process produced the multitude of sufficiently small ice-shelf fragments characteristic of the explosive ice-shelf break-up. In the present contribution, we build on this previous work by applying a viscoelastic model with linear Maxwell rheology to the Larsen B ice shelf in order to provide a more realistic representation of the ice-shelf break-up through the chain-reaction drainage of lakes. By adding a time dependency to the model, we also show that the majority of these lakes drained within just a few days of one another.


Low post-glacial rebound rates in the Weddell Sea due to Late Holocene ice-sheet readvance


Corresponding author: Richard C.A. Hindmarsh

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

The Holocene deglaciation of West Antarctica resulted in widespread ice surface lowering. While many ice-sheet reconstructions generally assume a monotone Holocene retreat for the West Antarctica ice sheet (WAIS), an increasing number of glaciological observations infer it is readvancing, following retreat behind the present-day margin. We will show that a readvance in the Weddell Sea region can reconcile two outstanding problems: (1) the present-day widespread occurrence of seemingly stable ice streams grounded on beds that deepen inland in apparent contradiction to theory; and (2) the inability of models of glacial isostatic adjustment (GIA) to match present-day uplift rates. Combining a suite of ice loading histories that include a readvance with a model of GIA provides significant improvements to predictions of present-day uplift rates, and we are able to reproduce previously unexplained observations of subsidence in the southern sector of the Weddell Sea. We hypothesize that retreat behind present grounding lines occurred when the bed was lower, and isostatic recovery led to shallowing, ice-sheet re-grounding and readvance. We will conclude that some sections of the current WAIS grounding line that are theoretically unstable may be advancing and that the volume change of the WAIS may have been more complex in the Late Holocene than previously posited. This revised Holocene ice-loading history would have important implications for the GIA correction applied to Gravity Recovery and Climate Experiment (GRACE) data, likely resulting in a reduction in the GIA correction and a smaller estimate of present-day ice mass loss within the Weddell Sea region of the WAIS.


Rapid grounding line migration induced by internal ice-stream variability

Alexander ROBEL, Christian SCHOOF, Eli TZIPERMAN

Corresponding author: Alexander Robel

Corresponding author e-mail: robel@fas.harvard.edu

Significant variability in the velocity of ice streams is a prominent feature of satellite and geomorphological observations from West Antarctica. These observations also indicate that grounding line position is strongly influenced by ice-stream variability, producing significant grounding line migration in the past and the modern. We analyze ice-stream grounding line variability using a stretch-coordinate flowline model. This model is based on that described in Schoof (2007), with a mesh refined near the grounding line to ensure accurate resolution of the mechanical transition from sheet to shelf flow. Here we have added an integrated lateral shear stress and a modified undrained plastic bed. We find that thermally induced internal ice-stream variability can cause rapid grounding line migration even in the absence of retrograde bed slopes or external forcing. Shock-like activation waves propagate along the ice stream length via a coupling of membrane stresses and frictional heating. These activation waves then trigger grounding line migration through a mass flux mechanism, which relates to important observables, such as mass balance and grounding line flux. Furthermore, we show how internally induced grounding line migration interacts with overdeepenings in the bed, and revisit some of the canonical aspects of the marine ice-sheet instability.


Mass-balance and runoff sensitivities of Norwegian glacier catchments to climate perturbations


Corresponding author: Markus Engelhardt

Corresponding author e-mail: Markus.Engelhardt@geo.uio.no

The climate sensitivity of a glacier catchment can be expressed in changes of mean annual mass balance (mass-balance sensitivity) or mean annual runoff (runoff sensitivity) to climate forcing. This study evaluates mass-balance and runoff sensitivities to annual changes in air temperature and precipitation for four highly glacierized catchments: Engabreen in northern Norway, and Ålfotbreen, Nigardsbreen and Storbreen which are aligned along a west-east profile in southern Norway. In addition to mean annual changes, mass-balance and runoff sensitivities are also calculated with respect to monthly perturbations in temperature and precipitation. The mass-balance sensitivities to annual air temperature range from 1.74 m w.e. °C–1 for Ålfotbreen, the most maritime glacier catchment, to 0.55 m w.e.°C–1 for Storbreen, the most continental glacier catchment in this study. In spite of these large differences, the increase in annual precipitation necessary to compensate for the mass loss due to a warming of 1°C is similar for all catchments, with 38% for Ålfotbreen and 33% for Storbreen. The runoff sensitivities show that a 20% increase in runoff is expected per degree temperature increase and a 10–20% runoff increase is expected for a 30% precipitation increase. Mass-balance and runoff sensitivities to monthly perturbations in temperature and precipitation give a much more structured description of the climate–mass-balance and climate–runoff relation. At Storbreen, both mass balance and runoff are sensitive to temperature changes only during summer months, whereas at Ålfotbreen, due to the snow/rain transition, temperature changes in autumn show a similar impact as in summer. Whereas mass balance is only sensitive to precipitation during the respective accumulation periods, runoff is mainly sensitive to precipitation changes in summer for Storbreen and mainly sensitive for both summer and autumn precipitation for Ålfotbreen, due to the precipitation maximum in autumn. Although the Norwegian mainland has a large north-south extend, the increasing continentality from west to east in southern Norway yields larger differences in mass-balance and runoff sensitivities between the catchments of Ålfotbreen and Storbreen to changes in temperature or precipitation than between Engabreen in northern Norway and the three catchments in southern Norway.


Drumlin formation at Múlajökull, Iceland


Corresponding author: Reba McCracken

Corresponding author e-mail: mccracrg@iastate.edu

The forefield of Múlajökull, Iceland, comprises the only known field of drumlins shaped by a modern glacier and provides an opportunity to study past patterns of sediment deformation and effective stress in and around these drumlins. Their geometric characteristics fall within ranges for Pleistocene drumlins, but glaciological conditions during drumlin formation are better known than for Pleistocene drumlin fields. Studying these drumlins can provide the basis for both better models of drumlin formation and better parameterizations of basal boundary conditions in glacier flow models. Initial measurements of the anisotropy of magnetic susceptibility (AMS) in one drumlin, from samples collected in vertical profiles at both its head and tail, show that simple-shear fabrics like those obtained experimentally through shear of the Múlajökull till are found throughout the drumlin. The mean deformation azimuth (weighted by fabric strength) of these fabrics differs from the drumlin long axis by 39.1 ± 29.9°. Fabrics from an inter-drumlin area adjacent to this drumlin are also offset from the drumlin long axis, by 35.5 ± 13.8°. No AMS fabrics indicative of pure shear – like those expected due to longitudinal compression or extension of till – were measured at any location in the drumlin or inter-drumlin area. These results, combined with observations of truncated beds at the sides of the drumlins, suggest that sediment deformation may have preceded drumlinization at Múlajökull, and that sediment deformation may not have been the primary mechanism of sediment transport that caused drumlin formation. Future measurements from till samples collected from different depths in other drumlins and inter-drumlin areas will yield preconsolidation pressures, revealing patterns of past effective stress and the possible role of groundwater flow in drumlin formation. This work motivates consideration of other mechanisms of sediment transport, in addition to subglacial sediment deformation (i.e. sediment erosion by subglacial water and by ice). Further work will include development of a three-dimensional model of drumlin formation. It will incorporate channelized flow of water around drumlins, groundwater flow through them, dynamic ice pressure distributions on drumlin surfaces, and associated effective stress distributions, and relate these factors to bed deformation and sediment transport by water and ice.


Experimental testing of glacier sliding rules


Corresponding author: Neal Iverson

Corresponding author e-mail: niverson@iastate.edu

Glacier sliding rules exist in various forms and are applied in modeling of glacier dynamics. Sliding rules have been, in most cases, theoretically derived but not experimentally tested. Under certain conditions, ice sliding over a rigid bed will generate cavities in the lees of bedrock bumps. These cavities will redistribute shear stress to regions of the bed that are in contact with ice. Sliding rules that incorporate cavity formation relate drag to the maximum adverse slope of the region of ice–bed contact. A sinusoidal bed is predicted to have a double-valued drag response as a function of sliding velocity. We have conducted an experimental study of sliding rules using a ring shear apparatus that slides ice over a rigid bed. The device rotates a ring of ice that is 20 cm wide, 20 cm tall, with an outer diameter of 90 cm. The sliding speed at the ice ring’s center line was incrementally stepped between 0.8 and 324 m a–1, and a vertical stress of 500 kPa was applied to the ice ring. The ice consisted initially of randomly oriented crystals that with sliding quickly developed a fabric like those observed in ice near glacier beds. The temperature of the ice is held at the pressure melting point and is regulated to ~0.01°C by a bath of circulating fluid that surrounds the sample chamber. Experiments have been conducted on a sinusoidal bed with a wavelength of 183.3 mm and an amplitude of 15.3 mm. Water was allowed to drain from cavities, so effective stress at the bed was equal to the total vertical stress. In addition to a rigid bed, experiments were conducted where ice was slid over a 7 cm thick deformable bed constructed from Horicon Till. The results from the sinusoidal bed demonstrate a double-valued drag response where shear stress initially increases then decreases. The decrease in shear stress for the sinusoidal bed is smaller than theory would predict over the same velocity range. These results provide the first experimental targets for models of sliding that attempt to assess effects of ice–bed separation.


Low-relief transverse moraines of the Des Moines Lobe of the Laurentide ice sheet as indicators of surging

Neal IVERSON, Suzanne ANKERSTJERNE, Mitchell CLINE, Sarah DAY, Lucas ZOET

Corresponding author: Neal Iverson

Corresponding author e-mail: niverson@iastate.edu

Low-relief transverse moraines, depending upon their origin, can yield information about flow or retreat kinematics of either terrestrial or marine-terminating glaciers. Some submarine transverse ridges adjacent to West Antarctic ice streams have been interpreted – based purely on ridge morphologies – to have been formed at the groundling line by daily tidal action, such that ridge spacing indicates the rate of grounding-line retreat. These submarine ridges are very similar geometrically to those formed by the fully terrestrial Des Moines Lobe (DML), the largest lobe (~200 km wide, 900 km long) along the southern margin of the Laurentide ice sheet and sometimes described as a surge-type glacier. The spacing of these DML ridges, called washboard, minor or corrugated moraines, was used 60 years ago to calculate the retreat rate of the lobe. Unlike submarine moraines, however, both the morphologies and internal characteristics of DML moraines can be readily studied. To determine the origin of these moraines, we used 1 m lidar data of washboard moraines in Iowa to analyze their geometric and spatial attributes, consolidation tests on tills of the ridges to estimate past basal effective stresses, magnetic fabric analyses calibrated to results of ring-shear experiments to assess till deformation kinematics, and traditional sedimentological analyses. Moraine ridges tend to be spaced with statistically significant periodicity (91–110 m as indicated by Fourier analysis) and are, on average, symmetric in longitudinal cross sections. Ridge trends deflect up-glacier coincident with outwash trains, which may have supported low basal pore-water pressures and anomalously slow basal slip. The moraines consist of overconsolidated basal till and so were beneath the glacier. Magnetic fabrics indicate flow-parallel simple shear on proximal sides of ridges and upward and transverse extension of till in pure shear close to ridge crests. Considered collectively these data indicate that the moraines, rather than being ice marginal features, formed subglacially as crevasse-squeeze ridges. As such, they provide good evidence of surge-like motion but provide no means of calculating retreat rates, since ridge periodicity reflects crevasse spacing rather than margin positions. These results highlight potential pitfalls of interpreting origins of transverse moraines based only on their morphology.


Modeling of hydrologically induced glacier velocity variations: toward a better understanding of the interaction between ice and subglacial water


Corresponding author: Basile de Fleurian

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

The interaction between subglacial water and basal sliding appears to be determinant to achieve a better understanding of ice dynamics. This statement is particularly true on the Greenland ice sheet margins where some large glaciers seem to undertake hydrologically induced velocity variation. These phenomenon have been observed on Russell Glacier on the southwest coast of Greenland where a number of field campaigns show a clear relation between water melt availability and glacier velocity. The spring speed-up observed on this glacier and on a number of other coastal or mountain glaciers is now accepted as hydrologically induced. The winter slowdown of these glaciers is still misunderstood even if an agreement seems to be reached on the role of the spreading of the subglacial drainage system efficient component. Models that are designed to account for inefficient and efficient drainage systems at the ice base are crucial to gain a better understanding of the complex feedback that exists between subglacial water and sliding velocities. Our approach, based on a double continuum method, in which water pressure is computed using Darcy’s equation for two different media, allows for a computation of water pressure on large glaciers during pluriannual simulations. A sediment layer with small hydraulic conductivity represents the inefficient drainage system and another layer with a much larger conductivity plays the role of an efficient drainage system. The dataset acquired on Russell Glacier allows for a realistic parameterization of the hydrological model and a way of validating its results. Once a proper parameterization of the model is achieved, application of idealized forcings to the hydrological model is a good means by which to investigate in a more detailed way the interactions between the subglacial water system and glacier dynamics.


Sensitivity of response of the Greenland ice sheet to global warming on surface mass balance and initialization methods


Corresponding author: Fuyuki SAITO

Corresponding author e-mail: saitofuyuki@jamstec.go.jp

We present a series of numerical experiments on the Greenland ice sheet under global warming conditions using the Ice sheet model for Integrated Earth system Studies (IcIES). This study focuses on the influence on the simulation from the difference in the method to compute the surface mass balance. Typically, the ice-sheet simulation is driven by a reference anomaly method, in which the surface temperature and/or the accumulation are decomposed into the reference terms (e.g. observation), the anomaly (e.g. climate scenario from climate models). Then the surface melting is computed using a parameterization such as the positive degree-day (PDD) method with the temperature. These decomposed terms have their own uncertainties, which may influence the ice-sheet simulation. In this study, the impact of these properties on the present-day control case, as well as the response under uniform warming conditions, are discussed, which is thought be useful and basic information on the property/sensitivity of the Greenland ice sheet. In addition, several initialization methods (free spin-up, fixed-topography spin-up, etc.) are applied to the IcIES in order to evaluate the influence of the error in the present-day simulated topography on the short-term response of the Greenland ice sheet.


Mass balance of the Sor Rondane outlet glacier system, East Antarctica


Corresponding author: Denis Callens

Corresponding author e-mail: dcallens@ulb.ac.be

The Antarctic mass balance is primarily governed by the difference between mass accumulated through accumulation and outflow through large outlet glaciers into the ocean. Therefore, outlet glacier dynamics play a crucial role in determining this balance. While West Antarctic glaciers are currently well monitored, knowledge of East Antarctic outlet glacier dynamics remains poorly understood and their dynamics are complicated by the outflow through and around coastal mountains. A significant part of ice flowing from the Dome Fuji ice divide flows around the Sor Rondane Mountains, establishing a system of four major outlets. Based on a recent airborne radar survey, we estimated their mass budget using the input–output method. To calculate the inflow, we determine the surface extent of each of the drainage basins from flowline backtracing. The total accumulation was then integrated within each basin. The outflux is calculated from the gate size and the velocity field. According to different assumptions on the vertical profile of the horizontal flow field, a range of possible outflow estimates has been obtained. The regional mass balance in this area is therefore estimated between 2.74 ± 1.77 and 10.13 ± 1.76 Gt a–1. This is slightly positive, and the wide range depends on the different assumptions made.


The length of the glaciers in the world – a straightforward method for automated calculation of glacier centre lines

Horst MACHGUTH, Matthias HUSS

Corresponding author: Horst Machguth

Corresponding author e-mail: homac@byg.dtu.dk

Glacier length and glacier length changes are probably the most widely used parameters in communicating glacier changes to a broader public. The now available worldwide glacier inventories and digital elevation models provide the basis for building a worldwide dataset of glacier length. Due to the large number of ~200 000 individual glaciers, automated approaches are needed. We present a straightforward approach for the calculation of so-called glacier centre lines. The method relies on two basic criteria: (1) hydrological flow and (2) maximal distance to the glacier margins. A series of trade-off functions flexibly controls the weight of the two basic conditions. For input, glacier polygons and a digital elevation model are required. The output is a set of centre lines comprising every individual branch of a glacier. The approach is fully automated and no glacier-, glacier-type or region-specific adjustments are made. The calculated centre lines are used as a measure of glacier length for large glacier samples. In a first step the method was validated for a region with numerous local glaciers in eastern Greenland (i.e. outside the ice sheet). For 100 randomly selected glaciers (0.01 km2 to 4200 km2 in area) the length of the longest centre line was compared with manually measured glacier length. The two length measurements agree very well (R2 = 0.99). By average automatically derived glacier length is 0.5% above the manually measured value. Deviations are small for most size classes except for glaciers < 0.5 km2 where the bias reaches 5%. Subsequently, the length of all glaciers represented by the Randolph Glacier Inventory 3.2 was calculated. For the first time we quantify glacier length characteristics for all glacierized regions of the world. It is shown, for instance, that the longest glacier in the world is located in Alaska/Canada (Bagley ice field, Bering Glacier, 187 km), while maximum length for low-latitude glaciers is around 7 km. Furthermore we present regional area–length scaling laws and global statistics on glacier surface slope along the centre lines. Automatically calculating glacier centre lines opens up new possibilities for analysing changes in large glacier samples: automated measurement of glacier length changes or detection of glacier surges might be of particular interest.


Evolution of subglacial meltwater channels near ocean termini


Corresponding author: Michael Dallaston

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

Melting at the ice–ocean interface exerts a critical control on ice-sheet response to climate change. The temperature and salinity of the ocean water play a role in this melting, but it is also strongly influenced by discharge of meltwater from beneath the ice, which initiates a buoyant plume that rises up the ice face entraining heat from the ocean to melt the ice. This raises the interesting question of how meltwater emerges from beneath the ice sheet – how spatially and temporally localized is it? We examine models of the subglacial drainage system to determine the possible modes of meltwater delivery across the grounding line. Rothlisberger channels could provide a localized source, although the lack of confining stress at a floating margin means that such channels become very large close to the margin. Moreover, the models suggest that stable well-defined channels may be unfavourable beneath rapidly moving outlet glaciers, where distributed flow through sediments or canals becomes more likely. We discuss the likely distribution of meltwater delivery across the grounding line, and the implications for the frontal melting rate and calving rate.


Towards an improved glacier representation in a global hydrological model


Corresponding author: Rianne Giesen

Corresponding author e-mail: r.h.giesen@uu.nl

Studies of the global glacier contribution to sea-level rise assume that glacier meltwater directly ends up in the world’s oceans. They do not account for any delay in the transfer of meltwater from the glaciated mountains to the ocean. Global hydrological models consider terrestrial fresh water in detail to assess water availability but generally simulate glacier processes in a simple way. We aim to improve the representation of glaciers in the global hydrological model PCR-GLOBWB, in order to obtain better estimates of downstream water availability for the past, present and future. PCR-GLOBWB calculates glacier melt with a positive degree-day model, applied using a lapse rate to a hypsometrical curve that is binned into a fixed number of classes on the basis of cumulative frequency. As a first step, we replaced this melt model by a simplified surface energy balance model which includes a separate term for the contribution of net solar radiation to glacier melt. This for instance allowed us to incorporate the effects of topographic exposure on surface melt. The model was run for the 20th century, using air temperature, reference potential evapotranspiration and precipitation from the Climate Research Unit TS3.21 dataset at a 0.5 arc degrees resolution. To validate the performance of the mass-balance model, we present modelled mass balances for a selection of glaciers with long observational mass-balance records. Results show a clear dependence of the mass balances on the coarse climate input. Therefore the next phase of our study will concentrate on the downscaling of the meteorological variables to better represent the regional and altitudinal variations in remote and rugged terrain.


Consistent modelling of subglacial drainage and sliding


Corresponding author: Ian Hewitt

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

Models of ice flow require a basal boundary condition that relates sliding speed and basal shear stress. In many operational models this is accomplished through the use of a basal drag coefficient that is treated as a spatially varying fitting parameter. This has the drawback that temporal changes in the drag coefficient are not accounted for. In other models, the boundary condition is assumed to depend on the effective pressure, with the effective pressure related in some way to the basal hydrology, potentially allowing for temporal changes. Theoretical reasons for the effective pressure dependence include pressure-induced cavitation and weakening of deformable till. There is clear field evidence of such effects, although the precise nature of the dependence is much less apparent. I discuss efforts to couple computational models of basal hydrology and ice dynamics, and highlight some of the potential issues involved. The basal boundary condition itself is only appropriate on a macroscopic scale, so any effective pressure dependence must represent a locally averaged effect. This raises questions about the assumed connectivity of the drainage system and how it may change in time, and it also makes comparison with borehole measurements a delicate task. Moreover, rapid temporal changes in the structure of the drainage system may invalidate much of the theoretical underpinning of current sliding laws. I show results of computations that couple variable basal water flow, evolution of the subglacial drainage system, and variable ice flow. These demonstrate the capacity to explain many observed meltwater-induced ice velocity changes. They also demonstrate that consistent model descriptions of basal hydrology and sliding require intricate coupling.


RFID technology applied to the glacial environment: MALATRA electronic system design and experimental data

Claudio LUCIANAZ, Marco ALLEGRETTI, Salvatore BARONE, Silvano BERTOLDO, Michèle CURTAZ, Giampaolo GRECO, Elena MOTTA, Andrea ROASIO, Oscar RORATO, Eliana VITTAZ, Giovanni PERONA

Corresponding author: Andrea Roasio

Corresponding author e-mail: aroasio@fondms.org

The higher mountains of the Alps focus in the western part of Europe and favor a high concentration of glaciers in this area. The Aosta Valley region is surrounded by mountains, more than the 50% of its territory lying above 2000 m a.s.l. In the summer, most of the water supply of the region relies on the contribution given by snowmelt and, partially, by ice melt. Study of glacial processes is thus very important in this region. In this context the MALATRÀ project (led by Fondazione Montagna Sicura and Envisens Technologies) is created to develop a low-cost instrumentation capable of measuring with continuity the physical parameters of snow and ice. The instrumentation consists of a miniaturized electronic device (tag) equipped with sensors and placed inside an ovoidal small-dimension (48 mm diameter and 180 mm length) plastic capsule. Moreover, the implementation of radio frequency identification technology (RFID) allows remote communication from the surface with the tags placed deep into the glacier, thus saving time, effort and cost in collecting data. Tags allow communication at long distance working at 315 MHz frequency. At this step, the goal is to use such devices during the annual glaciological campaigns to measure the weight of the snowpack above the tag (with a pressure sensor), in order to derive the snow water equivalent (SWE) and temperature inside the ice. As a first step, the capsules will be coupled with ablation stakes installed in the ice, placed at the bottom of boreholes. Each capsule is uniquely identified by a code and can be located in a 3-D local system via radio using a localization algorithm under development. It is then, during the installation, georeferenced absolutely using a GNSS receiver. This functionality also allows for the glacier displacement measurements. Once the device has been identified, all the data stored in the internal memory can be remotely downloaded from the reader. At the current development stage the board is equipped with a precise thermometer (PT1000) and a pressure sensor to catch ice data, a magnetometer and a tri-axial accelerometer sensor to study the movement of the capsule within the ice. The performance of the system has been tested in the glacial environment with excellent results.


Future projections of the Greenland ice sheet (GrIS) mass balance over 2006–2100 using the regional climate MAR model coupled with the ice sheet GRISLI model

Xavier FETTWEIS, Charlotte LANG, Catherine RITZ, Hubert GALLÉE

Corresponding author: Xavier Fettweis

Corresponding author e-mail: xavier.fettweis@ulg.ac.be

We present here future projections of the Greenland ice sheet (GrIS) mass balance over 2006–2100 simulated by the regional climate MAR model fully coupled with the ice sheet GRISLI model. The scenarios RCP6.0 and RCP8.5 from the general circulation MIROC5 model (known to perform well over Greenland) are used as forcing. MAR runs at a resolution of 30 km while GRISLI runs at 5 km. During the coupled simulation, a surface mass balance (SMB) local gradient based interpolation is first used to downscale the 30 km MAR-based SMB to the 5 km topography with the aim of annually forcing GRISLI. The topography and ice sheet mask changes simulated by GRISLI at 5 km are then interpolated to the 30 km grid to be taken into account in the MAR simulation through the coupling. For the first time, the melt–elevation feedback (known to enhance the surface melt as a significant thinning of the ice sheet should induce an additional warming) is explicitly taken into account in GrIS mass-balance future projections. The same future MAR projections but using a fixed ice-sheet topography will be used to evaluate the importance of this feedback to both SMB and ice dynamics changes in the coupled simulation.


Volume and frequency of ice avalanches from the Taconnaz hanging glacier (French Alps)

Christian VINCENT, Emmanuel THIBERT, Maxime HARTER, Soruco ALVARO, Adrien GILBERT

Corresponding author: Christian Vincent

Corresponding author e-mail: christian.vincent@ujf-grenoble.fr

In densely populated mountainous areas, breaking off from hanging glaciers can cause catastrophic damage to life and property. The hanging glacier of Taconnaz (Mont Blanc area) threatens the downward inhabited areas in the valley of Chamonix due to ice-falling events during the winter, particularly under unstable snow-cover conditions where they may trigger large snow avalanches. Regular terrestrial photogrammetric surveys have thus been carried out between April 2010 and August 2012 in order to assess the volume and the frequency of the largest collapses. We pointed out several major collapses which systematically occurred once the cliff edge reached a threshold distance. We found a single pseudo-period of about 180 days is associated with the largest ice-calving events. From the comparison of longitudinal cross sections, we assessed the volume of ice discharged by calving. We conclude that a critical geometry is a necessary condition, but not sufficient to lead to a large breaking-off. Indeed, in some cases, although the cliff edge reached the threshold length, the seracs disintegrated into small ice blocks without leading to a large breaking-off. It results that a precursory sign which can be used for the instability of this hanging glacier could be the geometry changes of the ice cliff, but it is probably not sufficient.


Improved location of the grounding line of an Antarctic outlet glacier from combined hydrostatic and kinematic methods


Corresponding author: Emmanuel Le Meur

Corresponding author e-mail: lemeur@lgge.obs.ujf-grenoble.fr

There are several methods for mapping the grounding line of an outlet glacier or ice stream. One of the most commonly ones used so far consists of detecting mostly from satellites surface typical features (for instance the so-called break in slope) supposed to represent the surface expression of the transition between grounded and floating ice. Here we propose the combiantion of two alternative approaches, namely the hydrostatic one and the kinematic one. The hydrostatic one relies on the measurements of both the ice thickness and corresponding surface elevation which, once treated with a floatation criterion, give areas where the ice is floating (or close to be so) thereby providing a preliminary outlining. The kinematic approach aims at detecting areas that undergo vertical displacements in response to the tidal cycle. The method requires high-accuracy GPS measurements of the ice uppper surface. Results from both methods are then compared with the restriction that deviations have to be expected given the rigid stresses mainly occuring under the short-term tidal displacements. A 2-D elastic ice slab is run in order to assess what can be the magnitude of the discrepancy between the two methods. Results show that this difference remains small and make the two methods consistent especially given the uncertainties associated to the hydrostatic approach. This leads us to propose a refined grounding line for Astrolabe Glacier, which in some places significantly deviates from the previous spaceborne ones.


Statistical study of the subglacial connectivity as revealed by pressure records


Corresponding author: Pierre-Marie Lefeuvre

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

At the Svartisen Subglacial Laboratory, pressure cells installed at the ice–rock interface under ~200 m of glacier ice have been recording basal pressure for 20 years. Four sensors in particular are studied, two pairs of sensors located 0.45 and 1 m apart that generally show correlation within the pair in the sensor response to changes in the subglacial hydrology. By studying the nature of this correlation as well as the correlation (or anti-correlation) between sensors in different pairs, it is possible to investigate evolution of the degree of connectivity of the hydrological system at the glacier bed. We compare the response of the first pair of sensors installed on an exposed gentle bump to a second pair, which are installed at the base and top (facing down) of an overhanging cliff. The sensitivity of the exposed pair is more pronounced than the latter, indicating the importance of their location and giving us vital information on the extent of the area affected by these events. A statistical overview of the nature of the response shows that certain signals are seen repeatedly in the load cell records and reflect the importance of the different kinds of perturbations in the hydrological system. It is also a firm confirmation that the connectivity of the subglacial hydrological system is highly dependent on the season, with the connectivity showing a monotonic increase in connectivity from early spring to early autumn (September). Significant changes in connectivity appear to be triggered by an increase in the rate of subglacial discharge measured nearby, and may be related to cavity opening at the glacier bed.


Thermal structures in proglacial lakes in the Southern Patagonia Icefield


Corresponding author: Shin Sugiyama

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

Mass loss of the Patagonian Icefields is one of the greatest unknowns in the contribution of mountain glaciers to sea-level change. The eastern part of the icefield is characterized by a number of outlet glaciers calving into freshwater lakes. Many of these calving glaciers are retreating, but rates are significantly different in each glacier. For instance, Glaciar Upsala retreated by 2.9 km over the period 2008–2011, being the most rapidly receding glacier in the region. On the other hand, Glaciar Perito Moreno has shown no significant change in the terminus position for the last several decades. Water properties and bathymetry of the lakes are suspected to control the inhomogeneous glacier retreat, but studies on freshwater calving glaciers are very scarce. To investigate the role of proglacial lakes in glacier changes, we measured temperature and other properties of lakewater in front of calving glaciers in the Southern Patagonia Icefield. Lake measurements were carried out at three large glaciers in the region: Glaciar Upsala, Glaciar Viedma and Glaciar Perito Moreno. Water in front of Glaciar Upsala was relatively cold (2–4°C), probably because of the large amounts of meltwater and ice discharge from the glacier. Turbid and cold water (< 1°C) was found at the deepest part of the lake (> 500 m below the lake surface), suggesting subglacial discharge of meltwater. In the lake of Glaciar Viedma, a sharp temperature transition exists at the depth of 100 m, from warm surface water (~6°C) to a very cold layer (< –0.5°C) in the deeper region. This stratification was due to a sill located at 2 km from the glacier front. A 300 m deep overdeepened basin between the glacier and the sill was entirely filled with the dense cold water. Contrasting to these two glaciers, cold deep water was missing in front of Glaciar Perito Moreno. Water temperature was relatively uniform at about 6°C, showing no clear stratification observed in the other two lakes. The lake is relatively shallow and the bed flat, which may have influences on the observed temperature field. Presented data indicate different thermal structures in front of the freshwater calving glaciers in Patagonia. These characteristic thermal regimes should play crucial roles in the melting of the calving face below the lake surface.


Glacier variations of SE Vatnajökull, Iceland, 1890–2010


Corresponding author: Hrafnhildur Hannesdóttir

Corresponding author e-mail: hrafnha@hi.is

The outlet glaciers draining towards the SE margin of Vatnajökull ice cap are located near the coast, in the warmest and wettest area in Iceland; they are non-surging and sensitive to climate change. The Little Ice Age (LIA) maximum extent of the glaciers (~1870–1890) is outlined by well-preserved glacial geomorphological features. A reconstruction of the LIA glacier surface was made, based on geomorphological and historical data, and information from the oldest reliable topographic maps of 1904. A lidar (laser scanning) digital elevation model (2010) was used as a topographic basemap for the reconstruction. Various archives on glacier extent and geometry since 1890 have been used to derive glacier changes of the last 120 years, including maps, aerial images, GPS measurements and the lidar survey. From this dataset, areal and volume changes are deduced, and the average geodetic specific mass balance estimated for several time periods. The ratio between volume and area changes is similar as reported in other areas of the world. The bedrock topography of SE Vatnajökull is known from radio-echo soundings, allowing estimation of relative volume changes. In the period 1890–2010 the outlet glaciers have been lowered by 150–270 m near the terminus and they have collectively lost 57 km3 of ice. These outlets constituting about 13% of the 1890 area of Vatnajökull, contribute about 15% of the post-LIA ice loss of Vatnajökull. The volume loss of individual outlet glaciers is in the range 15–50%, equivalent to mean geodetic mass balances of –0.2 to –0.8 m w.e. a–1. The most negative balance is observed between 2002 and 2010, when the glaciers lost –1.0 to –1.5 m w.e. a–1 on average. A coupled ice flow and positive degree-day model is used to simulate the 20th century evolution of the larger outlet glaciers, constrained with the history of volume change. The degree-day model is calibrated with annual mass-balance observations from SE Vatnajökull since 1996, and uses a 1×1 km grid of daily precipitation, and temperature from a meteorological station outside the glacier as input. The model results are in good agreement with glacier volume changes of the last 120 years.


Simulating the climatic response of Hardangerjøkulen ice cap since the Little Ice Age with ISSM


Corresponding author: Henning Åkesson

Corresponding author e-mail: henning.akesson@student.uib.no

Glaciers and small ice caps respond considerably faster to climate change than the large ice sheets Greenland and Antarctica. Evidently, half of the current sea-level contribution from the cryosphere comes from glaciers and small ice caps. Here we use the Ice Sheet System Model (ISSM) to model the dynamics and evolution of the maritime-continental Hardangerjøkulen ice cap (73 km2, 60.55°N, 7.43°E) in southern Norway from the Little Ice Age (LIA) until today. ISSM is a finite-element model with anisotropic mesh capabilities (resolution can be refined in regions of interest) and includes different approximations for the dynamics of ice flow, including the shallow-ice approximation (SIA) and full-Stokes. The SIA neglects important stresses when topography is complex; however, it has proved accurate in representing glacier volume fluctuations on decadal and longer timescales. As Hardangerjøkulen has relatively gentle slopes and lacks areas of very fast flow, we choose to use the SIA to study this ice cap on climatic timescales. As initial forcing for the ice flow model, we use a dynamically calibrated mass-balance history corresponding to moraine evidence from the Little Ice Age maximum in 1750 AD, as well as later outlet glacier front positions from moraines, direct measurements and aerial photographs. For the 1900s, we use surface mass balance from a spatially distributed energy-balance model using data from meteorological stations as forcing. Glaciological mass-balance records and front positions for the two main outlet glaciers, along with surface DEMs, are used for calibration. We investigate total ice volume and outlet glacier responses since the LIA. The sensitivity to surface mass balance as well as the applicability of the SIA to small ice caps is also discussed. Finally, our findings are compared and contrasted with previous model results for Hardangerjøkulen.


A multilayer ice flow model generalizing the shallow-shelf approximation

Guillaume JOUVET

Corresponding author: Guillaume Jouvet

Corresponding author e-mail: guillaume.jouvet@fu-berlin.de

A new hybrid model for the dynamics of glaciers, ice sheets and ice shelves is introduced. Unlike the traditional ice flow models that are simplified by neglecting higher-order terms in the aspect ratio, here the simplifications rely on an empirical approximation of the streamlines. In this model ‘multilayer’ the domain of ice consists of a pile of thin layers, which are aligned with the streamlines and which can spread out, tighten and slide over each other. Consequently, the two most relevant types of stresses are accounted for: the membrane ones and the vertical shear ones. Assuming the velocity field to be vertically piecewise-constant on each layer, the model derives from local depth-integrations of the hydrostatic approximation of the Stokes equations. These integrations give rise to interlayer tangential stresses, which are redefined by keeping the vertical shear components of the stress in the local frame of the interface. The final model consists of a tridiagonal system of two-dimensional non-linear elliptic equations, the size of this system equal to the number of layers. By construction, the model is a multilayer generalization of the shallow-shelf approximation (SSA). Like the SSA, the multilayer model can be advantageously reformulated as a minimization problem such that numerical techniques developed for the SSA can be easily extended. When running the model for the prognostic ISMIP-HOM experiments, the multilayer solutions show very good agreements with the Stokes solutions if no severe depression occurs in the bedrock. The multilayer approach, which is of mathematical 2-D complexity, offers to glacier and ice-sheet modellers a mechanically exhaustive and computationally efficient alternative to 3-D models when the streamlines can be empirically estimated as it is the case in many practical applications.


Simulation of sub-ice-shelf melt rates in a general circulation model: velocity-dependent transfer and the role of friction

Véronique DANSEREAU, Patrick HEIMBACH, Martin LOSCH

Corresponding author: Véronique Dansereau

Corresponding author e-mail: veronique.dansereau@lgge.obs.ujf-grenoble.fr

Interactions between the ocean circulation in sub-ice-shelf cavities and the overlying ice shelf have received considerable attention in the context of observed changes in flow speeds of marine ice sheets around Antarctica. Modeling these interactions requires parameterizing the turbulent boundary layer processes to infer melt rates from the oceanic state at the ice–ocean interface. Here we explore two parameterizations of heat and freshwater fluxes in terms of their impact on melt rates and on the wider sub-ice-shelf circulation. One form assumes transfer coefficients that are independent of the velocity of ocean currents at the ice-shelf base. An augmented form accounts for frictional turbulence via transfer coefficients that depend on boundary layer current velocities via a drag law. Both parameterizations are implemented in the MIT ocean general circulation model (MITgcm), and explored in simulations of the cavity circulation under Pine Island Ice Shelf (PIIS), West Antarctica. Significant differences in melt rate patterns between the velocity-independent and -dependent formulations are found. Whereas patterns are strongly correlated to those of thermal forcing for velocity-independent transfer coefficients, melting in the case of velocity-dependent coefficients is collocated with regions of high boundary layer currents, in particular where rapid plume outflow occurs. Both positive and negative feedbacks between melt rates, boundary layer temperature, velocities and buoyancy fluxes are identified. Melt rates are found to increase with increasing drag coefficient, in agreement with plume model simulations, but optimal coefficient values inferred from plume models are not easily transferable. Simulations with realistic tidal forcing suggest that the relatively weak tidal currents under PIIS neither increase the area-averaged melt rates significantly nor support extensive tidal mixing. Nevertheless, strong along-ice-edge barotropic currents arise from tidal rectification at the shelf front and significantly intensify velocity-dependent melt rates there, resulting in a positive feedback between strengthened currents and accelerated melting. This feedback is not captured by the velocity-independent melt rate formulation, with potential implications for the simulated cross-ice-edge heat exchanges and the sensitivity of basal melting under PIIS to offshore ocean warming.


Modelling the trajectory of the corpses of mountaineers who disappeared in 1926 on Aletsch glacier

Guillaume JOUVET, Martin FUNK

Corresponding author: Guillaume Jouvet

Corresponding author e-mail: guillaume.jouvet@fu-berlin.de

In this work we reconstruct the space–time trajectory beneath the surface of Aletsch glacier, Switzerland, of the corpses of three mountaineers who disappeared in March 1926 and reappeared at the glacier surface in June 2012. Our method consists of integrating the time-dependent velocity field of an existing full-Stokes glacier model, starting at the point where the corpses were found at the glacier surface. As a main result, the immersion location where the brothers presumably died on the glacier could be localized. As a second result, the upstream end point of the computed trajectory emerges very close to the glacier surface in 1926, giving a new and global validation of the glacier model in space and time. Testing the sensitivity of the obtained immersion location with respect to the model and other uncertainties indicates an area of 0.6% of the entire glacier area where the accident could have occurred. Our result refutes the theory of death caused by an avalanche or a fall into a crevasse, and posits that the mountaineers became disoriented in prolonged severe weather conditions, and froze to death.


Observation and modelling of Greenland ice sheet albedo and consequences on the projections of the ice-sheet surface mass balance

Eric BRUN, Marie DUMONT, Ghislain PICARD

Corresponding author: Eric Brun

Corresponding author e-mail: Eric.Brun@meteo.fr

Recent studies have shown that most of the decline in the Greenland ice sheet (GrIS) mass balance can be attributed to the decrease in the surface mass balance. The time and space evolution of surface albedo is a key factor to understand the dynamics of surface energy and mass budget of the ice sheet. Here we investigate the recent change in the GrIS albedo using MODIS retrievals. The observations are compared with a detailed modelling of snow and ice superficial characteristics. Simulated and observed by passive microwave melt events were also compared, showing that the model realistically simulates the interannual variability of both the GrIS surface mass balance (SMB) and the extension and duration of surface melting events. Building on the recent changes detected on the GrIS albedo, we performed a detailed analysis of how CMIP5 climate models simulate past and future snow albedo of the GrIS. Most of the climate models exhibit obvious weaknesses, which limit their capacity to accurately predict the future contribution of GrIS SMB to sea-level rise.


Long-term survey of Antarctica with repeat altimetry: ERS1/2, Envisat, SARAL

Frédérique RÉMY, Michel AURÉLIE, Denis BLUMSTEIN, Thomas FLAMENT

Corresponding author: Frédérique Rémy

Corresponding author e-mail: frederique.remy@legos.obs-mip.fr

Thanks to the launch of the CNES/ISRO mission SARAL in February 2013, the historical 35-d orbit initiated in 1991 by ERS-1, followed by ERS-2 and Envisat is extended and provided a unique dataset for ice-sheet mass balance survey for almost 20 years. The along-track processing allows to retrieve height variability and trend with a good space and time resolution for the objectives of ice-sheet survey. It also allows to capture acceleration of losses or gains and some short-scale events, such as emptying and refilling subglacial lakes. The space and time variability of the height but also of the backscatter and waveform shape parameters will be shown over Antarctica with the help of this high resolution. We will focus on the last observations provided by Altika on SARAL, exactly on the historical tracks. We will show the strong acceleration of mass loss of few glaciers on the West Antarctica ice sheet (especially Pine Island Glacier but also Thwaites and Smith Glaciers). Moreover, AltiKa operates in Ka band (35.75 GHz), a higher frequency than the classical Ku band (13.6 GHz), leading to important modifications and potential improvements in the penetration of the radar wave within the snowpack and a new kind of observation. In particular, we hope to catch accumulation rate and derive snowpack properties such as snow densification. First results will be presented.


Flow and seasonal calving dynamics in a 2-D model of Store Glacier, West Greenland


Corresponding author: Joe Todd

Corresponding author e-mail: jat71@cam.ac.uk

The rapid retreat and speed-up of many marine-terminating outlet glaciers explains about half of the 200 km3 net annual ice loss from Greenland, with the other half attributed to surface melting and runoff. Whereas the increase in surface melting and runoff is caused by atmospheric warming, the dynamic loss from outlet glaciers stems from their interaction with the ocean as well as the atmosphere, making future predictions of their stability and contribution to global sea-level rise difficult. Several mechanisms have been proposed to explain recent changes on calving glaciers. These include: undercutting of the terminus by submarine melting; seasonal formation of ice melange, a rigid mixture of icebergs and sea ice, which provides a buttressing stress on the terminus; and changes in internal dynamics, particularly basal sliding which affects the stress field near the terminus. A need to investigate the relative importance of these processes has prompted the development of increasingly sophisticated numerical calving models. A 2-D flowline model of Store Glacier in West Greenland was developed, using the model Elmer/Ice, with the aim of improving understanding of the processes of calving. In the model, calving occurs when the stress distribution of the ice allows surface crevasses to extend and meet basal crevasses, leading to iceberg detachment. We begin by modelling Store Glacier without imposing the effects of seasonal dynamics, ice melange or submarine melting, and then add these in turn, to investigate their relation to calving. We find that variability in basal slip has only a small effect on the calving rate. The effect of submarine melting is also limited, suggesting that fjord geometry exerts a strong control on the summer terminus location. However, when the buttressing effect of ice melange is applied, the modelled calving terminus advances and retreats in response to changing calving rate. This implies that ice melange is responsible for observed seasonal advance and retreat. Using the seasonally evolving yet stable glacier simulation as a starting point, we perform a suite of perturbation experiments in order to investigate the response of Store Glacier to potential future changes in these environmental forcings. We find that more than double observed submarine melt rate is required to initiate long-term retreat. However, we also find that, once the terminus loses its current stable position, the glacier retreats suddenly and by up to 30 km.


Measuring changes in the vicinity of the Seal Nunataks Ice Shelf remnant from imagery and altimetry

Christopher SHUMAN, Etienne BERTHIER, Ted SCAMBOS

Corresponding author: Christopher Shuman

Corresponding author e-mail: christopher.a.shuman@nasa.gov

Acquisition and analysis of a combination of repeated imagery and ICESat-1 altimetry has enabled the ongoing losses from the northern Larsen B ice shelf remnant to be visualized and measured. The northern remnant, the Seal Nunataks Ice Shelf (SNIS), has four ICESat tracks that cross it, as well as a track that obliquely crosses adjacent Robertson Island (RI). Two ICESat tracks cross the remaining tributary glacier system still contributing ice to the SNIS. The available altimetry data from ICESat (2003–2009) shows that elevation losses increase from west to east across the SNIS. In addition, the estimated ice shelf thickness losses adjacent to Robertson Island are almost 1.5× the average elevation losses measured for the grounded ice on RI suggesting a significant amount of basal shelf melting. Imagery analysis shows that ice area losses continued on both margins of SNIS, with the largest losses occurring on the north side in late 2004 into 2005. New rifts that form on SNIS do not tend to fill with meltwater and then commonly form icebergs. The tributary glacier system has experienced elevation losses as well as increased flow through an outlet discharging into Vaughan Inlet. In contrast to SNIS, RI has experienced relatively minor area losses suggesting that most of its ice is now grounded and is less impacted by ocean interactions as compared to the shelf remnant. This analysis is supported by DEM-differencing results (Berthier and others, 2012) and is compatible with the analyses of Pritchard and others (2012) and Rignot and others (2013). The combined datasets provide some insights about ongoing ice loss processes.


Spatio-temporal modelling of Antarctic mass balance from multi-satellite observations

Jonathan BAMBER, Nana SCHOEN, Andrew ZAMMIT-MANGION, Jonathan ROUGIER, Scott LUTHCKE, Thomas FLAMENT, Frédérique REMY, Liz PETRIE

Corresponding author: Jonathan Bamber

Corresponding author e-mail: j.bamber@bristol.ac.uk

Constraining past ice mass changes, identifying their cause(s) and determining rigorous error estimates, is important for closing the sea-level budget and as an input for, and test of, numerical models. For the Antarctic ice sheet, considerable uncertainty remains between different methods and groups. Estimates obtained from altimetry, gravimetry and mass-budget methods can yield conflicting results with error estimates that do not always overlap, while the commonly adopted use of different forward models to isolate and remove the effects of glacio-isostatic adjustment (GIA) and surface mass-balance (SMB) processes introduces another source of uncertainty which is hard to quantify. To address both these issues, we present a statistical modelling approach to the problem. We combine the observational data, including satellite altimetry, GRACE, GPS and InSAR, and use the different degrees of spatial and temporal smoothness to constrain the underlying geophysical processes. This is achieved via a spatio-temporal Bayesian hierarchical model, employing dimensionality reduction methods to allow the solution to remain tractable in the presence of the large number (> 106) of observations involved. The resulting trend estimates are only dependent on length and smoothness properties obtained from numerical models, but are otherwise entirely data-driven. The statistical methods are presented in a separate submission. Here, we present the annually resolved spatial fields for (1) dynamic ice loss, (2) SMB anomaly, (3) firn compaction and (4) GIA, using a combination of GRACE, ICESat, Envisat and GPS vertical uplift rates, for 2003–2009. The elastic flexure of the crust is also determined simultaneously. We estimate that, between 2003 and 2009, there has been an acceleration in dynamic ice loss, from close to balance in 2003/2004 to a rate of –200 Gt a–1 by 2009. This was predominantly driven by losses in West Antarctica and the Antarctic Peninsula. These dynamic losses have been partially compensated by an overall positive trend in SMB over the whole continent. We conclude that there was no statistically significant net imbalance in the 7 year period. Other data will be included to allow extention back to 1995 and forward to the present day using, for example, CryoSat 2, ice-core records and accumulation radar data.


High-resolution bed topography beneath the trunk and tributaries of Pine Island Glacier from ice-penetrating radar

Damon DAVIES, Robert G. BINGHAM, Stephen L. CORNFORD, Jan DE RYDT, Edward, C. KING, Andrew M. SMITH, Matteo SPAGNOLO, David G. VAUGHAN

Corresponding author: Damon Davies

Corresponding author e-mail: D.Davies@ed.ac.uk

Pine Island Glacier (PIG) in West Antarctica is currently losing ice at a rate equivalent to ~7% of current sea-level rise, and predicting its future is therefore an important scientific goal. Though the glacier has now been the focus of several modelling studies, the different models disagree on the likely future pace of loss and its spread inland, and few consider the potentially significant issue of subglacial erosion with time. A critical limitation has been a dearth of information both at spatial scales of sufficient resolution to be useful for ice-sheet model calibration, and a lack of information concerning temporal changes at the bed. In 2013/2014, as part of the UK NERC iSTAR (Ice-Sheet Stability and Response) traverse across PIG, we deployed the British Antarctic Survey’s DEep-LOoking Radio Echo Sounder (DELORES). To aid in calibrating ice-sheet models with specific reference to the issue of basal drag, we sounded ‘high-resolution’ patches of the bed in six 10 × 15 km locations across PIG. We also repeat-surveyed along several radar tracks previously surveyed in 2007/08 and 2010/11 (6 and 3 years earlier, respectively). This presentation presents some of our preliminary results.


Whole-continent Antarctica simulations using the BISICLES AMR ice-sheet model coupled with the POP2x ocean model

Daniel MARTIN, Xylar ASAY-DAVIS, Stephen PRICE, Stephen CORNFORD

Corresponding author: Daniel Martin

Corresponding author e-mail: dfmartin@lbl.gov

We present initial results from fully resolved (sub-kilometer resolution at grounding lines) simulations of the Antarctic ice sheet and its response to realistic forcing from sub-shelf incursions of warm water, obtained by coupling with the POP2X ocean model. The BISICLES model (Cornford and others, 2013) uses adaptive mesh refinement (AMR) to fully resolve dynamically important regions like grounding lines, while using much coarser resolution where the dynamics operate on slower and coarser time and length scales. We use a version of the vertically integrated formulation of Schoof and Hindmarsh (2010), which has compared well with models which solve the full-Stokes system (Pattyn and others, 2013). We demonstrate the importance of adequate spatial resolution in correctly resolving the dynamics of the ice sheet and its interaction with the ocean. BISICLES has been coupled with the POP2x ocean model – uncoupled and fully coupled results will be presented, including results from an idealized test case due to Goldberg and others (2012) and preliminary results of large-scale coupled simulation comprising the full-continent Antarctic ice sheet coupled to the full Southern Ocean.


Towards a minimal approach for assessing the dynamic behaviour and mass loss of marine-terminating glaciers

Andreas VIELI

Corresponding author: Andreas Vieli

Corresponding author e-mail: andreas.vieli@geo.uzh.ch

Glaciers and ice-sheet outlets that terminate in the ocean are exposed to both atmospheric and oceanic forcing. Their response, however, is often highly non-linear due to the sensitivity of their dynamics to glacier geometry (water depth and trough width) as indicated in recent rapid changes of tidewater glaciers in Alaska and Greenland. This makes interpretation of terminus changes and future predictions challenging, and current large-scale numerical models are still struggling to reproduce such behaviour. Here a simple approach is proposed to quantitatively assess the dynamic changes of tidewater glacier termini and to get a first-order estimate of potential retreat rates and mass loss. The water depth dependent flux-boundary condition for grounded calving termini of Schoof (2007), with an extension for variations in glacier width, is used for accounting the effects of fjord geometry on ice flux. Atmospheric and oceanic forcing are included through a simple parameterization in the boundary condition at the calving front. We explore the potential use of such a minimal approach on current and past examples of changing tidewater glaciers from the Greenland ice sheet and Alaska. Although this approach is highly reduced, it provides a physically based tool for assisting interpretation and assessment of past, present and near-future behaviour in tidewater glacier retreat. Crucially, this approach relies on only minimal data (bed topography and approximate surface mass balance) and is therefore widely applicable.


Modelling the subglacial hydrology of Russel/Leverett Glacier and implications for ice dynamics


Corresponding author: Mauro Werder

Corresponding author e-mail: m.a.werder@bristol.ac.uk

We apply the glacier drainage system model GlaDS to the catchment of Russel/Leverett Glacier, a land-terminating outlet glacier of the Greenland ice sheet. GlaDS is a 2-D subglacial hydrology model which incorporates both distributed and channelized drainage. There is a wealth of data available for Russel/Leverett Glacier which we use for model setup and forcings (bed and surface topography, surface mass balance), model tuning (proglacial discharge) and validation (tracer experiments, surface flow speed and uplift). We run the model for the ordinary melt year 2009 and for the extreme melt year 2012. For both years we show comparisons of summer flow speed versus modelled effective pressure and try to arrive at an empirical relation between flow speed and meltwater forcing.


Using data assimilation methods to explore the role of longitudinal stress gradients in Greenland outlet glacier flow


Corresponding author: Jesse Johnson

Corresponding author e-mail: jesse.v.johnson@gmail.com

Recent observations suggest that increases in outlet glacier speed are caused by geometric changes at the margins of the ice sheet. In many cases, warmer ocean water is believed to be promoting both melting at the grounding line and thinning or breaking up of ice shelves, altering the ice geometry. The proposed mechanism relating changes in geometry to increases in velocity is increases in longitudinal stress gradients. This component of the force balance, while small, is believed to play a large role in observed changes in ice dynamics. In this paper, we investigate the relative strength of the longitudinal stress gradients in a number of glaciers. Our strategy is different from what has been done previously because our analysis is carried out using vertically averaged velocities coming from a numerical model that has assimilated data. This provides the following advantages over force budgets that take only surface velocity and thickness into account: (1) the surface velocity field is smoothed by the momentum balance enforced by the model; (2) the coverage of the surface velocity does not have any gaps; (3) the steady-state temperature field is modeled so that temperature-dependent changes in viscosity are accounted for; and (4) the basal traction is computed directly from the inversion process, rather than as a residual of the force balance calculation of lateral drag and longitudinal stresses. Our investigation is carried out using VarGlaS (Variational Glacier Simulator), a 3-D, fully parallel, open source ice-sheet model that uses variationally derived forms for: (1) full-Stokes or first-order momentum balances, (2) enthalpy or internal energy, (3) free-surface evolution. These forms facilitate automatic differentiation and construction of adjoint operators for rapid assimilation of data. In this work, VarGlaS is used to invert surface velocity measurements for basal traction. Once modeled surface velocity is consistent with observation, the stress tensor is projected along flow to determine longitudinal stress, and across flow to determine lateral drag. Basal drag is taken directly from model inversion, hence a complete force balance is determined everywhere on the Greenland ice sheet. Using our force balance, we investigate the role of longitudinal stresses in the dynamics of outlet glaciers, quantifying the importance of each of the stresses in observed flow.


Modelling the role of oceanic and glacial forcings on the submarine melting of an East Greenland outlet glacier

Roberta SCIASCIA, Claudia CENEDESE, Fiammetta STRANEO, Patrick HEIMBACH

Corresponding author: Roberta Sciascia

Corresponding author e-mail: sciascia@mit.edu

Increasing evidence indicates that changes at the marine margin of Greenland’s tidewater glaciers may have triggered their recent acceleration and retreat and sizably increased Greenland’s contribution to global sea-level rise. One of the proposed mechanisms involves changes in submarine melting at the ice–ocean interface. Yet the parameters and processes controlling the submarine melt rate are largely unclear. Here we investigate the influence of the glacial forcing (from submarine melting and subglacial discharge) and the forcing from the continental shelf (e.g. wind events) on fjord dynamics and submarine melting of Helheim Glacier, a large outlet glacier of the Greenland ice sheet. We use a numerical non-hydrostatic ocean model (the MITgcm) with an ice-shelf parameterization initialized with data collected from Sermilik Fjord, where Helheim discharges. Our results show that the glacier experiences significant seasonal variability both in the vertical distribution and magnitude of submarine melting, largely due to discharge from the glacier. In summer, subglacial discharge increases submarine melting by an order of magnitude compared with winter. Maximum melting occurs near the glacier’s grounding line during summer, and near the interface between the two water masses characterizing the fjord stratification during winter. However, in winter, when subglacial discharge is small, the continental shelf variability regulates the submarine melting by varying the water properties at the glacier head. Our results compare favorably with recent findings based on fjord data and idealized laboratory experiments.


Modelling the dynamics of ice-stream shear margins

Marianne HASELOFF, Christian SCHOOF, Olivier GAGLIARDINI

Corresponding author: Marianne Haseloff

Corresponding author e-mail: mhaseloff@eos.ubc.ca

The Siple Coast ice streams are long, narrow bands of ice which are moving significantly faster than the surrounding areas. At the stream margins, fast-flowing ice slows down rapidly over a short distance. Observations suggests that these margins can migrate over time, leading to ice-stream widening and thinning. While previous research has shown that this migration results from an interplay of heating through lateral shearing and cooling through advection of cold ice from the sides of the stream, the relative influence of the different effects on the migration speed has not yet been quantified. Here we present results from a new boundary layer model that accounts for the mechanical and thermodynamic properties of ice-stream margins. The numerical solution of this model enables us to calculate the margin migration speed as a function of large-scale ice-stream properties such as ice-stream width, ice thickness gradient between ice stream and adjacent ice ridge, and geothermal heat flux. The influence of different basal boundary conditions at the ice-stream/ice-ridge transition on the margin migration velocity is investigated. Our results are verified with semi-analytical solutions for high heat production rates and high advection velocities, a limit that is likely to appear in real shear margins.


Bed topography under the Greenland ice sheet based on mass conservation


Corresponding author: Mathieu Morlighem

Corresponding author e-mail: mathieu.morlighem@uci.edu

Bed topography, together with ice thickness, is an essential characteristic of glaciers and ice sheets for many glaciological applications. Despite significant technical advances, it remains challenging to measure remotely. Here, we employ a mass conservation optimization approach that combines radar sounder collected by NASA’s Operation IceBridge since 2009, complemented by data acquired by NASA in 2001–2008, with high-resolution ice motion data from interferometric SAR (ALOS PALSAR, RADARSAT-1 and Envisat ASAR) to reconstruct bed topography beneath the Greenland ice sheet at an unprecedented level of detail. The results reveal overdeepening in the glacier fjords that are not apparent in current maps, and deep subglacial valleys that channelize ice flow to the coast. These features, mapped for the first time around Greenland using a combination of OIB and InSAR data, have vast implications for the modeling of the evolution of the Greenland ice sheet in a warming climate and suggest that the ice sheet will be more vulnerable to rapid retreat in the coming century than previously thought. We evaluate here the impact of this new description of the bed topography on ice-sheet models, and provide guidelines for future deployments. This work was performed at the University of California Irvine and the California Institute of Technology’s Jet Propulsion Laboratory under a contract with the National Aeronautics and Space Administration, Cryospheric Sciences Program, grant NNX12AB86G.


Sustained increase in ice discharge from the Amundsen Sea Embayment, West Antarctica, from 1973 to 2013


Corresponding author: Jeremie Mouginot

Corresponding author e-mail: jmougino@uci.edu

The glaciers draining into the Amundsen Sea Embayment (ASE) are known to be major contributors to sea-level rise from Antarctica, with a total mass flux comparable to the entire Greenland ice sheet. Since first revealed with satellite radar interferometry in the 1990s, this sector has been significantly out of balance due to glacier speed-up. Here, we combine measurements of ice velocity, and ice thickness from existing compilations to document 41 years of change in mass flux from the ASE. We derive ice-surface velocity from Landsat satellites between 1973 and 1989, ERS-1 for the winters of 1992 and 1994, ERS-1/2 for the winter of 1995, RADARSAT for the six winters between 2000 and 2005, ALOS PALSAR for the five consecutive winters between 2006 and 2010, RADARSAT-2 during fall 2011 and spring 2013, and TANDEM-X for winter 2012 and summer 2013. We also present the evolution of the grounding lines of the ASE glaciers between 1992 and 2011 using differential synthetic aperture radar interferometry (dinsar) data from the ERS-1/2 satellites. We estimate here that the total ice discharge has increased by 77% since 1973, with half of the increase occurring between 2003 and 2009. Grounding-line flow speeds at Pine Island Glacier stabilized between 2009 and 2013, following a decade of rapid acceleration and ungrounding of its ice plain, but acceleration reached far inland and occurred at a rate faster than predicted by advective processes. Ice flow speeds across Thwaites Glacier increased rapidly beginning in 2006, following a decade of near stability, leading to a 33% increase in ice flux between 2006 and 2013. Haynes, Smith, Pope and Kohler glaciers all accelerated during the entire study period, undergoing rapid ungrounding of ice plains or losing floating ice extensions. These results and satellite measurements give a good overview of the ice dynamic of the ASE during the last four decades , which is of great importance for understanding the evolution of a major part of West Antarctica.


Winter speed-up of Matanuska Glacier, Alaska

Masato FURUYA, Takahiro ABE

Corresponding author: Masato Furuya

Corresponding author e-mail: furuya@mail.sci.hokudai.ac.jp

Seasonal glacier velocity changes are attributed to subglacial slip associated with water pressure changes because of the seasonal variability of meltwater input. However, the dynamics of basal water are still uncertain particularly in the middle to upstream of mountain glaciers and ice sheets, limiting our ability to better simulate future ice dynamics and their possible impacts on sea-level rise. Abe and Furuya (2014) reported winter speed-up signals and their downglacier propagation at a number of glaciers at Yukon. Here we perform similar analyses at the Chugach mountain range of south central Alaska, and report the spatial-temporal evolution of Matanuska Glacier. Matanuska Glacier is the largest accessible glacier in Alaska, with its nearly 40 km length and 5 km width near the terminus. Although many field studies have been performed near the terminus, there are no reports of temporal surface velocity evolutions. We examined the surface velocity profiles at the lower 20 km from 2007 to 2011, using ALOS/PALSAR radar imageries. Comparing the winter velocity images in 2007, 2008 and 2010, those in 2010 were about 1.5–2.0 times faster than those during the previous 2 years. These data may indicate ‘mini-surge’ signals. In addition, comparing the fall and winter velocities, winter velocities were apparently faster at every 2007–2008, 2009–2010 and 2010–2011 season. These data indicate winter speed-up or mini-surge signals even at a temperate and non-surge-type Matanuska Glacier. Winter speed-up may not be uncommon at Alaskan/Yukon glaciers. Lingle and Fatland (2003) detected faster speed in winter than in fall at non-surging Seward Glacier in the St Elias Mountains; this is the only published report of winter speed-up, to our knowledge. Moreover, they also detected bull’s-eye-like localized signals at both surging and non-surging glaciers that represented vertical motions of the glacier surface. Combined with earlier glacier hydrological studies, Lingle and Fatland proposed englacial water storage and gravity-driven water flow toward the bed in winter regardless of whether a given glacier is surge-type or not. Basal crevasse observed at Bench Glacier, Alaska, by Harper and others (2010) is a preferred form of englacial water storage. Because it has no direct route to the surface but can store a significant volume of water near the bed, basal crevasse may generate high water pressure when it becomes constricted due to creep closure in winter.


Computation of mass change of Antarctic coastal areas from coastline products

Yixiang TIAN, Xiangfeng LIU, Xiaohua TONG, Rongxing LI

Corresponding author: Xiaohua Tong

Corresponding author e-mail: xhtong@tongji.edu.cn

Measurement of changes in area and mass of the Antarctic ice sheet is critically important and has been made by using different remote-sensing technologies and ground exploration data. For estimation of contribution to sea-level change, one of the methods of mass loss estimation is based on glacier ice flows as mass passing grounding lines. This way of computation requires the mass balance consideration of the grounded part of the ice sheet. It should be noted that changes on ice shelves have little or no direct effect on sea-level change contribution. On the other hand, areas other than ice shelves may include ice-sheet boundary on land, or ice walls where the ice sheet directly interacts with water. They can be generally described by coastlines. Mass change in these areas should contribute to sea-level change. Therefore, the analysis of the coastline changes can be used to estimate this part of the Antarctic mass change. Coastlines extracted from the MODIS Mosaic of Antarctica (MOA) and RADARSAT-1 image mosaic are used in this paper to compute the area of changes. The MOA grounding lines and other sources were used to exclude ice shelves. Ice thickness associated with the changed coastal areas is derived from the Bedmap 2, and then used to estimate the mass change. The result shows that the mass loss from the coastal areas is relatively small, a few percent of mass loss through glacier ice flows. However, this is one part of the Antarctic mass changes that should be precisely characterized.


Southeast Greenland net snow accumulation derived from airborne IceBridge accumulation radar, ground-based radar and firn core

Clément MIÈGE, Richard FORSTER, Jason BOX, Lora KOENIG, Joseph McCONNELL, Evan BURGESS

Corresponding author: Clément Miège

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

Net snow accumulation over the Greenland ice sheet needs to be constrained in a recent warming climate because it is the only component that could offset the dramatic mass loss observed for the last decades in Greenland. The southeast sector of the ice sheet witnesses the highest accumulation (up to 5 m w.e.) near the coast, contributing to one-third of the total ice sheet annual accumulation. However, this region still challenges regional climate models, and averaged accumulation differences are found up to 1.5 m w.e. between climate models. Here, we derive accumulation rates by combining three different datasets. (1) 400 MHz ground-based radar data were collected in spring 2010 and 2011. Radar profiles were obtained following two east-west radar transects near the Arctic Circle, and imaging internal horizons up to 50 m depth in the firn, with an annual resolution restricted to the dry firn zone. (2) NASA’s Operation IceBridge Accumulation Radar (750 MHz) has been flown in conjunction with ground radar collection. The airborne accumulation radar profile displays up to 150 m of internal isochrones in the dry zone but only few horizons are detected in the percolation zone. (3) A total of five shallow firn cores (50–60 m) provide the depth age scale at point location to date the internal radar layers and test their isochronal behavior. We found good agreement between ground and airborne accumulation products. The ground radar gives more detail in the percolation zone, but the airborne radar has a greater penetration depth. Surface undulations (< 10 km) observed in the southeast, show differential accumulation with an accumulation increase up to 10% in topographic depressions, not yet taken into account by the coarser-resolution (> 10 km) regional climate models. From the interior toward the margin of SE Greenland we note an increase of both accumulation rates and interannual variability magnitude, in agreement with regional climate models. Over the last decades, despite the climate warming, there is no significant increase in accumulation. This expected increasing trend might be masked by high interannual variability and the relatively short time span of this study. The surface melt increase, both in spatial extent and duration, is happening at gradually higher elevations over the ice sheet, and allows meltwater to penetrate in the snow/firn. The meltwater refrozen subsurface features challenge internal layer detection and tracking over long distances.


The role of grounding-zone processes in sea-level rise


Corresponding author: Knut Christianson

Corresponding author e-mail: knut@nyu.edu

Much of the threshold behavior of marine ice sheets is thought to result from processes occurring in the grounding zone, where the ice sheet transitions into the ice shelf. At short timescales (decades to centuries), grounding-zone behavior is likely to be influenced by processes not included in the current generation of ice-sheet models. Here we report on several such processes including: the flow of subglacial water beneath the ice sheet, and the associated transport and deposition of sediment at the grounding line, the penetration of ocean water inland of grounding due to ice flexural effects, and variation in ocean-driven basal melt rates at the grounding line. We present data from ground-based geophysical studies across the grounding zones of two major West Antarctic ice stream (Whillans Ice Stream (WIS) and Pine Island Glacier (PIG)). At WIS, we image a subglacial estuary, which consists of a hydropotential low upstream of the grounding zone, which is linked to the ocean by a hydropotential trough and a large subglacial channel. Pressure differences along the trough axis are within a range that can be overcome by tidally induced processes, making the interaction of subglacial and ocean water likely. At PIG, we use GPS data to estimate seasonal variation in basal melt rate. We assess the impact of grounding-zone processes on ice-sheet mass balance using a high-order flowband model and a shallow-ice/shallow-shelf approximation three-dimensional ice-sheet model. Our results indicate that grounding line stability is sensitive to basal sliding coefficient in the grounding zone and the basal rheology assumed by the model. Basal rheology is especially important as long-term ice-stream stabilization or rapid retreat can result from assuming either a linear-viscous or effectively plastic bed.


Precision assessmentof remote-sensing coastline products for boundary change analysis of the Amery Ice Shelf, East Antarctica

Tiantian FENG, Chenxi WU, Xiangfeng LIU, Xiaohua TONG

Corresponding author: Tiantian Feng

Corresponding author e-mail: fengtiantian@tongji.edu.cn

Investigating the changes of the Amery Ice Shelf is essential to the understanding of the entire Antarctic ice sheet. As one of the important indicators of such changes, the dynamics of the ice-sheet boundary can portray the growth or retreat of the ice shelf. As a result, a lot of coastline extraction methods based on different remote-sensing missions have been developed. There are a number of Antarctic coastline products derived based on remote-sensing data of different missions. These coastline products include ice-shelf boundaries and should be useful for ice-sheet change study. However, there have been few systematic investigations on accuracy and cross assessment of these products. This paper will focus on the precision assessment of three coastline products of the Amery Ice Shelf based on a cross-validation method. The earliest product, DISP coastline, was extracted based on the Declassified Intelligence Satellite Photographs (DISP) by using a procedure of automatic extraction and manual refinement. The DISP images were acquired on 29 August 1963, 29 October 1963 and 3 November 1963 respectively. The second product, RADARSAT coastline, was generated by using a RADARSAT-1 image mosaic. An automated coastline extraction algorithm was applied to extract the coastline. The data acquisition time is from 9 September to 20 October 1997. Finally, the third product, MODIS coastline, was generated by using a MODIS Mosaic of Antarctica (MOA). The dataset was acquired over the austral summer season between 20 November 2003 and 29 February 2004. This coastline product was digitized based on visual interpretation. Four evaluation indexes are proposed to assess the precision of the above three coastline products in the Amery Ice Shelf. They are Cumulative Radio Curve (CRC), Fractal Dimension, Deviation Distance, and Change Areas. CRC is a buffering-based index, which can be regarded as a distribution curve for the distance between two coastline products. Fractal Dimension is a statistical index of complexity that reflects how detailed the coastline is. Discrete Frechet Distance is adopted here as Deviation Distance, which is used to calculate the maximal distance between two coastline products. The last index is used to describe the change areas between two coastline products. The quantitative assessment results will be given in this paper, and the impact of several factors, such as ice calving, tide and ice velocity, will be discussed to analyze the uncertainty of these coastline products.


Modelling runoff glacierized drainage basin using a coupled glacio-hydrological model in a changing climate


Corresponding author: Kjetil Melvold

Corresponding author e-mail: kjme@nve.no

The future runoff from two highly glacierized catchments in western Norway is assessed for the period 1961–2100 using a glacio-hydrological model. The hydrological model is a distributed HBV model, simulating mass-balance runoff response as a function of altitude based on temperature and precipitation. The HBV model is coupled to flow models computing the length and volume change of the glaciers with time. The model is calibrated using observed discharge, or mass-balance data and elevation-change data. The model is run with transient simulations spanning a range of possible future changes. We apply gridded scenarios for daily temperature and precipitation data derived from regional climate models. Three selected climate projections from the ENSEMBLES project with the A1B emission scenario have been used. Before 2005, the model is forced with gridded temperature and precipitation values from seNorge.no based on data from the Norwegian meteorological station network. Model simulations reveal that during the first half of the 21st century, substantial future variations in runoff superimposed on a rising trend and a reduction in ice volume. The runoff changes from the glaciated part fluctuate close to ~2.8 m a–1 for the first decades after 2005, rising to ~3.8 m a–1 near the middle of the century and continue to increase to ~4–5 m a–1 at the end of the century. If we look at the total runoff from the catchment the picture is different in the later part of the century. The runoff then drops down to present level.


Using a simple ice flow model to investigate the large-scale spatial variability of the subglacial system of northern Greenland


Corresponding author: Nanna Karlsson

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

It is well known that subglacial water has the potential to modify and even control the flow of ice. Even so the inclusion of basal processes in ice flow models remains a challenge, while at the same time observational evidence increasingly underscores the important role liquid water plays in ice flow dynamics. In this study we use a simple, vertically integrated, uncoupled ice flow model to investigate elevation changes in northeast and northwest Greenland. Assuming that the pathways of the subglacial water are controlled by the hydropotential (usually a valid assumption on large spatial scales), we study the effects of surface elevation changes on the subglacial system. Our model indicates that even variations in surface elevation of the order of 1 m a–1 or less can cause changes in subglacial water outflux at the margin over timescales varying from decades to millennia. Our results emphasize the need for a coupling between ice flow models and subglacial processes to allow for simultaneous changes in ice flow mode and basal properties.


Experimental results of grounding line change analysis of the Antarctic ice shelves using MODIS, ICESat and InSAR data

Yang XU, Huan XIE, Saisai LU, Marco SCAIONI, Xiaohua TONG

Corresponding author: Huan Xie

Corresponding author e-mail: Xiehuantj@gmail.com

The aim of this research is to investigate grounding line changes in the Antarctic ice shelves by comparing three existing grounding line products derived from satellite images, satellite altimetry data, and satellite interferometric synthetic aperture radar (InSAR) covering the period 1992 2009, which are available at National Snow and Ice Data Center. These products used data collected by MODIS image, Earth Remote-Sensing Satellites 1-2 (ERS-1/2), RADARSAT-1 and 2, the Advanced Land Observing System (ALOS) PALSAR, and ICESat/GLAS laser altimetry. The MODIS grounding line (MOA) is represented by polylines, and the InSAR and ICESat grounding lines are represented by different feature points of the grounding lines. Three methods are used to compute the grounding line changes. They are: (1) passing through rate of grounding zone; (2) distance analysis between different grounding line products; (3) buffer analysis between different grounding line products. The errors in the original data sources, the actual grounding line changes, and the errors in the generation of grounding line products are discussed. The objectives of this research were to analyze errors, mitigate systematic effects when possible, and discuss limitations of using grounding line products at local and continental scales. Finally, a grounding line difference map of the Antarctic ice shelves was generated.


Inferring basal conditions at Engabreen from surface measurements


Corresponding author: Anne Solgaard

Corresponding author e-mail: solgaard@gfy.ku.dk

In this study, we combine a range of surface measurements and inverse modelling to infer basal conditions at Engabreen in northern Norway. We investigate the distribution of the basal friction coefficient and how much of the total velocity component can be attributed to sliding and deformation, respectively. The open-source finite-element code Elmer/Ice, which solves the full-Stokes problem, is used. Inputs to the inversion scheme are surface velocities from SAR and time-lapse imagery, radar measurements of the ice thickness as well as a high-resolution surface DEM from recent laser-scan measurements. Due to the coarse resolution of the radar data, we improve the ice thickness distribution in the lower part of the glacier by combining ice flux considerations and information from the surrounding ice-free topography.


Flow modeling of Juneau Icefield

Florian ZIEMEN, Constantine KHROULEV, Regine HOCK

Corresponding author: Florian Ziemen

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

Juneau Icefield is located in the Coast Mountains on the border between Alaska and British Columbia. Its main outlet glacier, Taku Glacier, has shown intriguing retreat and advance behavior in the last century. There have been several measurement campaigns on Juneau Icefield (including the well-known Juneau Icefield Research Project), so the data coverage is comparatively good. It includes transient snowlines and local measurements of ice thickness, flow velocities and mass balance. We model the mass balance and flow behavior of Juneau Icefield with the Parallel Ice Sheet Model (PISM). PISM is a hybrid shallow-ice/shallow-shelf ice-sheet model that can model the interior of the icefield as well as sliding glaciers. It handles temperate ice by modeling the ice enthalpy instead of temperature. Due to its parallel nature, it can be used at the high resolution necessary to resolve the intricate structure of Juneau icefield and its several outlet glaciers while still providing good performance. We calibrate the model on the past behavior of the icefield and study its future evolution in CMIP5 climate change scenarios. Preliminary results under pre-industrial forcing show recurring speed-ups of Taku Glacier related to an internal thermal instability mechanism.


Origin of englacial diffraction features in radio-echo sounding data from the ablation area of the Greenland ice sheet


Corresponding author: Katrin Lindbäck

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

Drainage of surface meltwater to the bed of the Greenland ice sheet is believed to affect the ice dynamics. An understanding of the geometry and evolution of the englacial drainage pathways is important to assess the effects of melt-induced dynamic response. In this work we use radio-echo sounding to investigate the characteristics of englacial features that is commonly observed in radar data from the ablation area on western Greenland. These radar features are attributed to moulins or cracks intersecting internal layers causing point reflections forming vertically stacked hyperbola patterns in the radar data, although their origin is not fully understood. In this study we investigate the origin of englacial features that is commonly observed in radar data from the ablation area on western Greenland. We find a close spatial agreement between filled surface lakes observed in remote-sensing data and the presence of englacial diffraction features in ~1500 km of radar data covering an area of 50 × 50 km in the upper ablation area. Using a 2-D electromagnetic wave-propagation model we simulate a set of geometries of englacial features including moulins, supraglacial lakes and englacial conduits. We find that the simplest model domain that reproduces the observed features is water-filled surface lakes which support the observed spatial correlation between the englacial diffraction features and filled supraglacial lakes. This contrasts with earlier studies that indicate that these features originate most likely from moulins that reach the bed, which may be due to limited field observations.


Investigating conditions for freeze-on at the ice-sheet base and its effect on the ice sheet

Gwendolyn Leysinger VIELI, Carlos MARTIN, Richard HINDMARSH

Corresponding author: Gwendolyn Leysinger Vieli

Corresponding author e-mail: gwendolyn.leysinger@geo.uzh.ch

Large englacial features have recently been observed in radio-echo sounding data in both Antarctica and Greenland. These features resemble the shape of plumes, which can, in terms of their geometric structure, be reproduced by modelling internal isochrone layers including basal ice accretion. However, to produce such large freeze-on features sufficiently high basal accretion rates are needed. This may be possible through the process of Clausius-Clapeyron cooling of basal water, which in turn requires a high water flux and a steep rise in basal topography (inverse to the surface slope). Interestingly, such features are found in North Greenland, a region which at the base is influenced by a high geothermal heat flux allowing for large basal melt rates, thus high availability of basal water, and a region of onset of fast-flowing ice streams. Further, in Antarctica the englacial plumes so far observed are near Dome A, East Antarctica, a slow-flowing region, where the alpine-like relief of the Gamburtsev subglacial mountains allows for basal melt in its deeply incised valleys, due to increased pressure melt points, and for freezing on its steep ascending slopes. Here we are investigating the conditions required in terms of water flux and bed slope to allow sufficiently high basal accretion rates, and the distribution of the plumes in relation to ice flow and topography, both by using the 3-D numerical model BASISM to calculate temperature, melt and water flux. As the ice rheology might be affected by basal freeze-on, we also use the Elmer/Ice model to include the effect of ice fabric evolution and internal temperature variation.


Evolution of Antarctic Peninsula outlet glaciers after disintegration of the northern Larsen Ice Shelf


Corresponding author: Helmut Rott

Corresponding author e-mail: helmut.rott@uibk.ac.at

For estimating ice-sheet contributions to sea-level rise under climate warming scenarios it is crucial to understand the response of grounded ice masses to disintegration of buttressing ice shelves. Comprehensive satellite datasets, acquired over the years following the ice-shelf disintegration events on the northern Antarctic Peninsula (API), offer excellent opportunities to study the response of outlet glaciers that previously drained into the ice shelves. We studied the temporal evolution of outlet glaciers of the API to Prince Gustav Channel and Larsen A over the period since final disintegration of the ice shelves in January 1995, and the behavior of outlet glaciers to the Larsen B embayment after ice-shelf disintegration in March 2002. Detailed maps of ice motion, covering the whole region, were derived from imaging radar (SAR) data of the ESA satellites ERS-1, ERS-2, Envisat, and the German TerraSAR-X and TanDEM-X satellites. Topographic change over 2011–2013 was mapped using high-resolution digital elevation models (DEMs) generated from bistatic interferometric SAR data of the TanDEM-X/TerraSAR-X satellite configuration. Precise data on glacier surface elevation and elevation change of previous years are available along ground tracks of NASA’s ICESat and of the airborne ATM lidar sensor of the IceBridge campaigns. These datasets were used to analyse the temporal behaviour of velocity on glacier tongues and calving fronts, to derive calving fluxes in different years, and to estimate mass depletion at the scale of individual glaciers. The flow of the majority of the glaciers showed similar temporal behaviour, with rapid acceleration soon after ice-shelf collapse. The acceleration, triggered by stress perturbation at the glacier front, propagated rapidly upstream through dynamic coupling. The maximum frontal velocities were reached a few years after the ice shelf disappeared. On several of the main glaciers the velocities started to decrease significantly a few years later. Slowing down of calving velocities and decreasing calving cross sections due to dynamic thinning result in reduced calving fluxes. The trend of decreasing ice export is confirmed by the glacier volume change measured by TanDEM-X DEM differencing. These observations indicate gradual dynamic adjustment of the outlet glaciers and are of relevance for estimating the sea-level contribution from downwasting of Antarctic Peninsula glaciers.


Recrystallization diagram for polar ice

Ilka WEIKUSAT, Daniela JANSEN, Nobuhiko AZUMA, Sérgio H. FARIA

Corresponding author: Ilka Weikusat

Corresponding author e-mail: ilka.weikusat@awi.de

With the warming climate, ice masses on Earth are expected to increasingly contribute to a rising sea level. As for any material, the ice bodies’ temperature is a key variable to change the material’s properties, especially the rheology. In the case of ice in natural environments on Earth, temperature is always close to the material’s melting point. Therefore ice can be regarded as a ‘hot material’ (homologous temperatures T/T_m ca. 0.7 to 0.9). This means that recrystallization plays a decisive role in governing the state and thus the behaviour of the material, as it continuously resets the mechanical properties. Recrystallization as a set of control mechanisms has been recognized and interpreted in many ice cores in the last decades, and certain recrystallization regimes have been assigned to special ice-sheet depth ranges. This assignment was based on microstructure observations (mainly grain size) and estimated boundary conditions (temperature and stress/strain amounts) which change systematically with depth. To generalize the use of recrystallization regimes we decouple their occurrence from the ice-sheet depth information and connect them directly to the activators and causes: strain rate and temperature.


Towards a dynamic Greenland ice sheet in an Earth system model

Michiel HELSEN, Roderik VAN DE WAL, Richard BINTANJA

Corresponding author: Michiel Helsen

Corresponding author e-mail: m.m.helsen@uu.nl

The Earth system model EC-Earth is a valuable tool to calculate the effects of greenhouse gas (GHG) scenarios on various components of the climate system. Coupling a thermomechanical ice-sheet model to EC-Earth offers the possibility to not only quantify the effects of GHG scenarios on ice mass changes, but also to take into account the effects of geometrical changes of the ice sheet on the climate system. One of the mayor issues related to coupling an ice-sheet model with a climate model is related to the calculation of the surface mass balance (SMB) over ice sheets, which is vital to obtain a realistic climate forcing. In this study we use a surface energy-balance approach to calculate SMB within EC-Earth, and compare our results with SMB from the regional atmospheric climate model RACMO2. The ice-sheet integrated SMB from EC-Earth for the present-day climate is too low, which is predominantly the effect of an underestimation of the precipitation on southwest Greenland. Ice-sheet model simulations for the present-day climate, driven by both EC-Earth and RACMO2 climate forcings, show that simulations are most sensitive for albedo changes.


Thermodynamically coupled grounding line migration

Martin RÜCKAMP, Angelika HUMBERT

Corresponding author: Martin Rückamp

Corresponding author e-mail: martin.rueckamp@awi.de

Until recently ice sheets were considered to change substantially only on timescales of centuries to millennia. This view is changing now as geophysical and satellite observations indicate a significant response of ice sheets to climatic change on decadal timescales. The cause is believed to be a dynamic response of ice streams draining the ice sheet. Crucial findings about grounding line dynamics are (1) to consider all stress components in this transition zone to model the flow adequately and (2) the strong dependency on the horizontal grid size at the grounding line, as the dynamical processes need to be resolved on a horizontal scale in the order of the local ice thickness and below. For glacier ice specifically, viscosity depends both on temperature and liquid water fraction (moisture), leading to a thermomechanically coupled and polythermal flow problem. We apply the thermodynamically coupled full-Stokes finite-element model (COMice) to an idealized ice-sheet/ice-shelf system. Temperature and moisture are solved using the recently proposed enthalpy method according to Aschwanden and others (2012). We investigate the effect of horizontal grid size and thermal diffusivity in temperate ice on groundling line migration and the corresponding mass flux across the grounding line.


Investigating the impact of melange on glacier dynamics


Corresponding author: Jean Krug

Corresponding author e-mail: jean.krug@ujf-grenoble.fr

Ice melange is a heterogeneous mixture of sea ice, wind-blown snow, fragments of marine ice and calved icebergs which can be found in the fjords of outlet glaciers in Greenland, or in Antarctica, inside the rifts of larges ice shelves. Its behaviour is highly dependant on the season: in winter, freezing sea ice rigidifies the melange by binding together its component; in summer, melting of ice weakens the melange, allowing each of its components to move independently from the others. In the Greenlandic fjords, observations have shown that the seasonal cycles of advance and retreat of the outlets glaciers are correlated with the state of the melange, with an advancing front in winter preceding a rapid retreat of the glacier in the late spring, when the melange weakens. Previous studies suggest that the back force applied by the rigid melange layer in winter may prevent calved icebergs from rotating away from the glacier front. As a response, the glacier front advances and slows down. On the contrary, the abrupt disintegration of melange in spring releases the back pressure, allowing icebergs to detach from the glacier, leading to a front retreat and an increase of ice velocity. Here, we study this process using a new calving law based on both continuum damage mechanics and fracture mechanics. This framework is implemented in the Elmer/Ice full-Stokes finite-element model, and thus allows for a reliable representation of processes occurring at the front. Several experiments are carried out, investigating the response of a synthetic outlet glacier to different parameters, such as the melange thickness, the applied back force and the glacier size. At last, the impact of a seasonal variability of the strength of the melange layer on the behaviour of the glacier over several years is investigated.


Surface mass-balance pattern over the Antarctic plateau from ground-penetrating radar and shallow cores

Emmanuel LE MEUR, Olivier MAGAND, C. LACHAUD, Laurent ARNAUD, Michel FILY, Joseph ERBLAND, Gregory TESTE, Eric LEFEBVRE

Corresponding author: Emmanuel Le Meur

Corresponding author e-mail: lemeur@lgge.obs.ujf-grenoble.fr

During the 2011–12 scientific traverse between the bases of Dome C and Vostok, 630 km of ground-penetrating radar profiles have been measured along portions of the route. Horizons are clearly visible down to 150 m and feature layers of the same age, the so-called isochrones. Along these profiles, several shallow cores have been drilled and locally provide a depth–age relationship from the identification of atmospheric thermonuclear deposits (1950s to 1980s) and/or volcanic horizons. At the intersections with ice-core sites, horizons can be locally dated and because of their isochronous nature, the age obtained can be spatially extrapolated all along the isochrones as long as they remain continuous. After a post-processsing phase in order to improve the signal/noise ratio, major isochrones have been picked on the radargrams, thereby providing a continuous record of accumulated firn/ice over the period corresponding to the age of the isochrone. Last, from the density profiles along the ice cores, these firn/ice thicknesses can be expressed in terms of water equivalent accumulation rates. We first present a description of the data (cores and radargrams), the way they are post-processed and interpreted before synthetizing all the results in term of spatially varying mass balance over different periods.


Large-scale reconstruction of Holocene accumulation rates in northern Greenland

Christine S. HVIDBERG, Lisbeht T. NIELSEN, Nanna B. KARLSSON

Corresponding author: Lisbeht T. Nielsen

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

Internal layers mapped in polar ice sheets by means of radio-echo sounding contain imprints of past accumulation patterns and ice flow. Thus, the layers can potentially provide estimates of past accumulation rates, which are of interest for investigations of the past evolution of the ice sheets. In this study, the average accumulation rates for the last 14.7 ka in North and central Greenland are recovered by inversion from the mapped depth of a pronounced isochrone using a simple ice flow model accounting for both vertical and horizontal flow combined with formal inversion techniques. The pronounced isochrone marks the transition into the Bølling–Allerød interstadial at 14.7 ka and was previously mapped from radio-echo sounding data. The strength of the method applied in this study is the 2-D inverse approach that takes upstream variations of accumulation into account. The simple ice flow model is based on balance velocities in the horizontal direction and a simple laminar profile in the vertical, although additional calculations using other simple profiles were also performed. The isochrone is a deep layer, and the depth of the layer is expected to be influenced by conditions near the bed, which is not included in the simple ice flow model. The inferred accumulation rates are found to agree well with Holocene estimates from deep and intermediate ice cores in the interior of the ice sheet. Across the central ice divide, the inferred accumulation pattern shows a pronounced gradient in accumulation rates, similar to the present-day pattern, suggesting that the patterns have been consistent throughout the Holocene. It is concluded that the 2-D approach is an improvement over 1-D methods, despite the simple ice flow model. The agreement at ice-core drilling sites is partly due to the drill sites being located near the divide, where the simple flow model is a better approximation to the vertical velocity profile compared to elsewhere, which is in agreement with previous studies. In areas of large horizontal ice flow, the inversion is found to be only weakly constrained by the layer depth, limiting the applicability of the analysis based on the deep isochrone to the interior regions of the ice sheet in the absence of further data constraints. Later analysis could include additional shallow radar layers and thereby extend the region with well-constrained accumulation rates.


Dynamics of glacier calving at Helheim Glacier, southeast Greenland


Corresponding author: Tavi Murray

Corresponding author e-mail: t.murray@swansea.ac.uk

By bringing together expertise in glaciology, GNSS (Global Navigation Satellite System) technology and processing, and wireless networks we have designed, installed and operated a wireless network of GNSS sensors very close to the margin of the heavily crevassed and fast-flowing Helheim Glacier in southeast Greenland. The network has low energy consumption and a novel base-station topology providing diversity and redundancy. Data collection occurred at rates up to every 7 s. The network is robust to the loss of nodes as the glacier calves, and data are collected right to the point of sensor loss. Such a network would also be suitable for data collection in a number of harsh environmental settings such as earthquake, landslide or volcano monitoring. In 2012, we undertook field trials installing three GNSS sensors on the glacier’s flowline, and observed the dynamic effects of a major calving event. In 2013, a 19-node wireless network was installed together with five oblique cameras, instrumenting an area ~16 km2 of the glacier margin. The network ran for 55 days during July through to early September, and many sensors survived in locations right at the glacier front to the time of iceberg calving. The sensor network provides velocity and elevation data of unprecedented resolution in time and space for the key marginal area of the glacier, where recent changes in glacier dynamics appear to have initiated. The observation period included a number of significant calving events, and over the observation period the glacier retreated ~1.5 km. Auxiliary data include airborne lidar measurement of surface topography, crevasse spacing and calving rates, oblique photogrammetry, and DEMs and velocity fields from TanDEM-X satellite imagery. We present data showing the glacier’s dynamic and topographic response to iceberg calving events. The data provide evidence of the overall mechanism of iceberg calving, which appears to be controlled by the formation of basal crevasses, as well as the instantaneous response to calving which provides evidence of the physical processes that generate glacial earthquakes. The data thus provide rich opportunities for testing calving models and for improving understanding of the controls on the contribution of these tidewater glaciers to sea-level rise.


Glacier instabilities: processes and early warning perspectives

Jerome FAILLETTAZ, Martin FUNK, Christian VINCENT

Corresponding author: Jerome Faillettaz

Corresponding author e-mail: jerome.faillettaz@geo.uzh.ch

Glacier instabilities are gravity-driven rupture phenomena. Three different types of instabilities can be identified depending on the thermal properties of the ice/bed interface. If cold (1), the maturation of the rupture is shown to be associated with a typical time evolution of surface velocities and seismic activity generated by the glacier. A prediction of the final break-off is possible using these precursory signs. For the other types of instabilities, water plays a key role in the initiation and the development of the instability. If the ice/bed interface is partly temperate (2), the presence of meltwater at the interface reduces its basal resistance, which promotes the instability. No clear and easily detectable precursory signs can be evidenced in this case, and the only way to infer any potential instability initiation is to monitor the time and spatial evolution of the thermal regime at the interface. The last type of instability (3) concerns steep temperate glacier tongues switching for a couple of days/weeks during the melting season into a so-called ‘active phase’ followed in rare cases by a major break-off event. Although the prediction of such events is still far from being achievable, critical conditions promoting the final instability are identified with a newly developed numerical model including water flow in a subglacial drainage network.


The application of active remote-sensing techniques to the Roi Baudouin Ice Shelf (East Antarctica) in order to investigate the interplay between locally grounded features and the surrounding ice-shelf dynamics

Sophie BERGER, Reinhard DREWS, Frank PATTYN

Corresponding author: Sophie Berger

Corresponding author e-mail: sberger@ulb.ac.be

The Roi Baudouin Ice Shelf in Dronning Maud Land ( East Antarctica) is laterally confined by two ice rises and additionally buttressed by a comparatively small pinning point on its front. Both ice rises and the pinning point deviate and decelerate the ice flow of Western Ragnhild Glacier, one of the biggest outlet glaciers of the Dronning Maud Land coast. Large parts of the ice-shelf surface exhibit along-flow topographic depressions, which correspond to sub-ice-shelf channels penetrating up to 60 % of the ice thickness. Our aim is to better understand the interplay between the ice-stream/ice-shelf system with the regrounded features located seawards of the grounding line. This study applies multiple active satellite remote-sensing techniques (SAR interferometry, intensity tracking, coherence tracking) with various sensors (ERS, ALOS, Envisat) to retrieve the surface velocities. The data are supplemented with on-site ground-penetrating radar and GNSS data collected in two consecutive field seasons in 2012 and 2013. We find that, despite its small extent (a few kilometers), the pinning point plays a fundamental role in buttressing the ice shelf. This becomes evident in a zone of horizontal shearing which originates from the pinning point and extends upstream almost up to the grounding line. Based on the horizontal strain band, we find that the pinning point exhibits the bulk part of the western buttressing force. The much bigger ice rise on the western side plays only a minor role. We also find a second zone of horizontal shearing which corresponds to the longitudinal channeling; however, this must be treated with care as the topography of the surface depressions is not sufficiently resolved. In order to clarify whether or not these channels are potential weak spots in the ice shelf, we combine the different remote-sensing techniques with the available ground-truth data. We hypothesize that the sub-ice-shelf channels are an important indicator for spatially changing ice-shelf properties which may play an important role in defining the ice shelf’s buttressing strength.


Altimetry waveforms inversion over Antarctica

Denis BLUMSTEIN, Fernando NINO, Frédérique REMY, Etienne BERTHIER

Corresponding author: Denis Blumstein

Corresponding author e-mail: Denis.Blumstein@legos.obs-mip.fr

Measurements provided by spaceborne radar altimeters are much richer than the few parameters traditionnally used in the applications (mainly ground altitude and backscatter). Indeed, the whole history of the radar return is available; this is called radar waveforms. By a careful analysis of sequences of consecutive waveforms, it is possible to retrieve crucial information about the nature of the soil backscatter as well as details about the topography at a resolution much better than the footprint of the altimeter. In particular the shape of the waveforms allows us to discriminate the power return by the surface from the return by the subsurface. These parameters can then be used to provide information about geophysical characteristics of the terrain (snow grain size, etc.) and its temporal evolution through the analysis of the penetration of the radar wave in the snow. This poster will describe the technics we have developped to perform waveform inversions through the use of an accurate waveform simulation model that is able to handle the Envisat mission (Ku band, 13.6 GHz) as well as a new mission AltiKa from CNES/ISRO that provides measurements in Ka band (35.75 GHz) on the same orbit. We will also show how we can use good high-resolution DEM, e.g. from the Spirit projet (CNES/SPOT IMAGE/IGN), in order to improve the retrievals in regions that are notoriously difficult for radar altimetry (near the coast). Finally we will show results obtained on a few places of the Antarctica ice sheet.


Changing surface geometric effects on mountain glacier mass balance

Christopher WILLIAMS, Jonathan CARRIVICK, Andrew EVANS

Corresponding author: Christopher Williams

Corresponding author e-mail: c.n.williams@swansea.ac.uk

Glacier mass balance depends upon the dynamic change of glacier geometry. Despite this effect being recognized as important, few modelling studies have addressed and quantified it specifically. This study presents a 99 year reconstruction and mass-balance response analysis of Kårsaglaciären, a small (0.89 km2) mountain glacier located in Arctic Sweden, using a number of techniques in order to overcome this limitation. A geodetic approach was used to assess changing mass balance over time. Data were derived from topographic maps (1909–1991) and contemporary field surveys (2008–2011). These data were used to interpolate a number of DEMs and full 3-D reconstructions were derived, providing information on spatial change for the period 1909–2010. A long-term trend of negative mass balance was identified. The glacier retreated 1292 m, thinned by 0.35 m w.e. a–1, decreased in area by 0.04 km2a–1 and reduced in volume by 1.33 km3a–1. The 3-D reconstructions provided the input for a user-friendly, simple distributed surface energy-balance model, aimed at facilitating the assessment of the effect of geometry change on mass balance – designed specifically for this study and made available to other researchers online (https://github.com/Chris35Wills/SEB_model_java_files). The model is open source and users are encouraged to further develop the existing model structure. Using the reference balance approach, it was possible to assess change in mass balance over time both with and without dynamic surface adjustment, allowing disentanglement of these effects with climate. Geometry change on an annual basis had little effect on glacier mass-balance response to climate but has a significant dampening effect for the period 1926–1943, reducing maximum surface mass-balance change by 1.15 m w.e. a–1. These results provide evidence of Kårsaglaciären showing a strong pattern of retreat throughout the 20th and early 21st century. From these analyses it is apparent that the effects of glacier geometry on mass-balance response are not simply linked by time. Future mass-balance studies should consider changes in glacier geometry for accurate assessments of glacier response to climate.


Sensitivity study of the Antarctic surface mass balance to snow erosion by the wind

Hubert GALLÉE, Charles AMORY, Cécile AGOSTA, Alexandre TROUVILLIEZ, Xavier FETTWEIS

Corresponding author: Hubert Gallée

Corresponding author e-mail: gallee@lgge.obs.ujf-grenoble.fr

A new version of the regional climate model MAR has been developed. New numerical schemes allow us to set-up the model over Adélie Land, Antarctica, with a fine horizontal resolution (5 km) and an improved vertical resolution near the surface (lowest level is now situated 1 m above the surface). The domain of the model covers the steepest slopes of Adélie Land, on an area of 500 times 500 km2. Simulations last the two summer months (December and January), allowing an optimal interaction between observers and modellers. The influence of model improvments on the wind speed, the horizontal blowing-snow flux and the surface mass balance simulated by the model is discussed.


Velocity variations and calving flux of Kronebreen, NW Svalbard, from SAR offset tracking


Corresponding author: Thomas Schellenberger

Corresponding author e-mail: thomas.schellenberger@geo.uio.no

Kronebreen, a tidewater glacier in NW Svalbard, is among the fastest-flowing glaciers in Svalbard, and an important contributor to mass loss from the archipelago. Here we present a time series of area-wide surface velocity fields based on offset tracking of repeat high-resolution RADARSAT-2 data from April 2012 to May 2013. Maximum velocities of 3.1 m d–1 are measured close to the calving front in summer 2012. Additional velocity snapshots since December 2007 from different SAR sensors, and continuous GPS measurements since September 2008, reveal different patterns in seasonal and interannual velocity variations. We observe high variability due to variations in amount and timing of meltwater input and rainfall, both of which have a strong influence on the efficiency of the drainage system. Due to its fast flow, the calving flux Q is an important factor in the mass balance of Kronebreen. Beside the continuous mass loss through ice movement, front position changes represent a second important component of Q. During the observation period, Kronebreen retreated up to 850 m, for a total area loss of 2.11 km2. Total ice flux from Kronebreen, obtained by combining calving and ice loss through frontal retreat, is 0.19 ± 0.04 Gt a–1 for the period 8 May 2012 to 3 May 2013.


Modelling the behavior of Jakobshavn glacier in the last century

Ioana S. MURESAN, Constantine KHROULEV, Shfaqat A. KHAN, Kurt H. KJÆR, Jason E. BOX

Corresponding author: Ioana S. Muresan

Corresponding author e-mail: iomur@space.dtu.dk

Current model estimates of the Greenland ice sheet (GrIS) are almost entirely based on coarse grids (> 10 km) and constrained by climate models that span from 60s to present. To improve the projection of future sea-level rise, a long-term data record that reveals the mass balance beyond decadal timescale is required. Here, we use a continuous 171 year reconstruction (since the end of the Little Ice Age) by J.E. Box of the GrIS climatic surface mass balance and its sub-components to study the interaction between climate and the cryosphere originating in changes in the surface mass balance and dynamics of the GrIS over the last 111 years. Throughout our study, we use the Parallel Ice Sheet Model (PISM) capabilities. The initialization of the ice sheet is performed on a 5 km grid using paleo-climatic forcing (–125 ka to present) based on a positive degree-day (PDD) model. For a better overview and for the purpose of increasing the resolution to 2 km, our study focuses only on Jakobshavn glacier. In order to determine the locations of the flow for the regional model, a drainage basin mask was extracted from the surface elevation data based on the gradient flow. While inside the basin mask the full PISM model is applied, outside the basin mask the boundary conditions are taken as captured by the whole Greenland initialization. Considering the surface mass balanced reconstruction where the monthly accumulation rates are assumed to be 1/12 of the annual accumulation, a yearly 1900–2011 climatic forcing is applied in the regional run.


Modelling of the future climate and surface mass balance of Svalbard with the regional climate model MAR

Charlotte LANG, Xavier FETTWEIS, Michel ERPICUM

Corresponding author: Charlotte Lang

Corresponding author e-mail: Charlotte.Lang@ulg.ac.be

We have modelled the climate and surface mass balance of Svalbard over 2010–2099 at a resolution of 10 km forcing the regional climate model MAR with MIROC5 based on the RCP8.5 scenario. MAR predicts negative values of the SMB over the whole of Svalbard for the period 2070–2099, with the biggest losses located in the southern parts of Spitsbergen. It also predicts a similar evolution of the SMB over every region of Svalbard up to 2055 with a stable SMB or a slight acceleration of the loss. After 2055, the ice loss accelerates suddenly but the acceleration is stronger in the south than in the north of Spitsbergen and on Nordaustlandet. This difference is caused by the larger amount of shortwave radiation absorbed by the surface of the snowpack (SWDnet = SWD*(1-a)) in the south. The beginning of the melt season is not expected to occur much sooner in the next 50 years (0.05 days sooner per year until the 2060s) but the trend is going to accelerate afterwards (0.65 days per year) whereas the trend of the date of the end of the melt season should be higher but steadier. Because the glaciers are no longer covered with freshly fallen snow at the time of the maximum of melt as soon as the 2030s, there will not be any delay between the maximum of melt and the maximum of runoff as the ice or the packed snow covering it can not store water anymore. Finally, we have investigated the influence of the energy-balance components on the sensitivity of melt with regard to the temperature rise. It appears that the evolution of the shortwave radiation is the dominant factor through the effect of the decreasing albedo on the increasing amount of absorbed radiation (36% of the net energy flux over 2080–2099). It is followed by the sensible and latent heat fluxes (24 and 20%) and finally by the net longwave radiation flux (19% of the net energy flux).


Beyond back-stress: investigating the importance of small changes in effective basal pressure in causing rapid accelerations of Greenland outlet glaciers


Corresponding author: Leigh Stearns

Corresponding author e-mail: stearns@ku.edu

The Greenland ice sheet’s contribution to sea level has doubled in the last decade, partly as a consequence of increased ice discharge through outlet glaciers. Recent studies suggest that iceberg calving or melange weakening cause acceleration and thinning through a reduction of the longitudinal stress gradients (‘back-stress’). However, satellite remote-sensing data reveal that while acceleration and retreat patterns are often sudden and short-lived, these events are preceded by several years of thinning. Thinning usually persists throughout the lower ~50 km of the trunk, with rates that range from 5 to 150 m a–1. The timing of thinning, acceleration and retreat patterns can be examined using the force-balance technique. We use coincident ice velocity and surface elevation measurements to investigate the relative importance of driving and resistive stresses at different epochs, on several Greenland glaciers. Results from the force-balance model allow us to explore the back-stress hypothesis by calculating the magnitude of longitudinal stress gradients at different epochs. Results using this traditional force-balance method are compared with results from flowline and fully coupled models. Preliminary results suggest that resistive stresses, including longitudinal stress gradients, remain fairly constant despite large flow perturbations. Small changes in effective basal pressure, however, can drive many of the observed changes in flow behavior. We explore the implications of these results on ice-sheet mass-balance projections.


Reducing the uncertainty in projections of future ice-shelf basal melting


Corresponding author: Ralph Timmermann

Corresponding author e-mail: Ralph.Timmermann@awi.de

Simulations of ice-shelf basal melting in future climate scenarios from the IPCC’s Fourth Assessment Report (AR4) have revealed a large uncertainty and the potential of a rapidly increasing basal mass loss particularly for the large cold-water ice shelves in the Ross and Weddell Seas. The large spread in model results was traced back to uncertainties in the freshwater budget on the continental shelf, which is governed by sea-ice formation. Differences in sea-ice formation, in turn, follow the regional differences between the atmospheric heat fluxes imprinted by the climate models. A more recent suite of BRIOS and FESOM model experiments was performed with output from two members of the newer generation of climate models engaged in the IPCC’s Fifth Assessment Report (AR5). Comparing simulations forced with output from the AR5/CMIP5 models HadGem2 and MPI-ESM, we find that uncertainties arising from inter-model differences in high latitudes have reduced considerably. Projected heat fluxes and thus sea-ice formation over the Southern Ocean continental shelves have converged to an ensemble with a much smaller spread than between the AR4 experiments. For most of the ten larger ice shelves in Antarctica, a gradual (but accelerating) increase of basal melt rates during the 21st century is a robust feature throughout the various realizations. Both with HadGem2 and with MPI-ESM forcing, basal melt rates for the Filchner–Ronne Ice Shelf in FESOM increase by a factor of two by the end of the 21st century in the RCP85 scenario. For the smaller, warm-water ice shelves, inter-model differences in ice-shelf basal mass loss projections are still slightly larger than differences between the scenarios RCP45 and RCP85; compared with AR4 projections, however, the model-dependent spread has been strongly reduced.


Greenland ice flow controlled by soft bed response to surface meltwater drainage


Corresponding author: Marion Bougamont

Corresponding author e-mail: mb627@cam.ac.uk

Flow variation has been observed on a large number of glaciers in Greenland, on timescales varying from hours to years. In particular, the sudden deliveries of surface meltwater to the bed subsequent to supraglacial lake drainages cause rapid and high-magnitude perturbations to the basal environment and drive short-lived but pronounced flow accelerations. The overall effect of surface meltwater on the annual ice budget, however, remains unclear, some studies suggesting that more melt will increase ice flow, while others suggesting the opposite. Theoretically, the evolution of the subglacial drainage system from low to high efficiency is currently the principle mechanism through which ice flow seems able to self-regulate over the season. The process is however complex and its representation in numerical models is challenging. Additionally, by focusing on the character of the hydrology system, the potential implications of a different bed substrate, known to be important in other glaciated regions, has been overlooked. Here, we quantify the large-scale, seasonal effect of surface meltwater drainage on the ice flow of Russell Glacier, a land-terminating glacier in West Greenland. Recent geophysical surveys have revealed the presence of a weak sediment layer beneath the glacier, with a typical porosity of 30–35%. We integrate these observations in a numerical model, and study the hydrological impacts on flow over a soft bed. To do so, we implemented into CISM (Community Ice Sheet Model) a detailed subglacial sediment model coupled to a regional hydrology model. Using the 2010 surface melt record to force the model, our results show that the weakening and strengthening of the subglacial sediment associated with spatio-temporal intake/outflow of water routed along the bed modulates the ice flow consistently with observations. We conclude that the sediment control on ice flow is a viable alternative to existing hydrology models. Furthermore, sensitivity experiments indicate that the annual ice motion over a sedimentary bed may increase together with surface melt, evoking a potential key role of subglacial sediment in defining the future stability of the ice sheet.


The influence of the Greenland ice sheet on Arctic sea ice in a warm scenario

Marianne Sloth MADSEN, Shuting YANG, Guðfinna AÐALGEIRSDÓTTIR, Synne Høyer SVENDSEN, Christian RODEHACKE

Corresponding author: Marianne Sloth Madsen

Corresponding author e-mail: msm@dmi.dk

Fully coupled models including a dynamically interactive Greenland ice sheet are a very useful tool to explore the interactions of ice sheets with the atmosphere, ocean and sea ice. Here we use the global climate model EC-Earth coupled to the Parallel Ice Sheet Model (PISM) to investigate the processes and possible feedbacks important for Arctic sea-ice evolution in a 4×CO2 scenario. The coupling includes a modified surface physics parameterization in EC-Earth along with the calculation of surface mass balance (SMB) which accounts for precipitation, evaporation and surface melting (snow and ice). The SMB and temperature are used to force the ice-sheet model, which returns the simulated ice discharge, basal melt, ice extent and ice thickness to be used as boundary conditions for EC-Earth. The coupling time step is 1 year. We consider two 350 year climate change experiments: (1) abruptly increasing the CO2 concentration to four times the pre-industrial value and (2) increasing the CO2 concentration 1% per year until 4×CO2 is reached. These scenario runs are compared with the corresponding simulations performed with the EC-Earth CMIP5 version without the interactive ice sheet. In the coupled simulations, we find increased freshwater fluxes from the Greenland ice sheet into the Arctic and North Atlantic basins and a slower recovery of the reduced North Atlantic Meriodinal Overturning Circulation as compared with the CMIP5 runs. We show that the disappearance of Arctic sea ice in summer is occurring later in the coupled simulations, emphasizing the needs for detailed process-based studies in order to quantify climate change in the Arctic.


A long-term history of Antarctic grounding line change


Corresponding author: Hamish Pritchard

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

Rapid grounding line retreat is underway along several sections of the Antarctic ice sheet margin, forced by the arrival of warm water at the coast and leading to the continent’s significant ongoing contribution to sea-level rise. Even an ice sheet stable to grounding line perturbations will continue to lose mass if this forcing is sustained, but the forcing is poorly understood. Is it unusual? Will it continue or increase? We can start to answer these questions by looking for evidence of coastal dynamic change stored within the ice itself. Fortunately, a record of flow dynamics along the ice-sheet margins – the glacial response to grounding line change – is preserved within ice divides. ‘Raymond bumps’ in the internal ice stratigraphy form over a long (and calculable) time under an unmoving divide, hence their presence indicates a sustained, unchanging flow regime. A present-day divide that is offset from its underlying Raymond bump indicates a long spell of unchanged flow perturbed at some time in the more recent past. I present here a map, derived from both ground and satellite radar remote sensing, of this Antarctic flow history spanning centuries to millennia that serves as a proxy for changes at the grounding line. This reveals a surprising picture of prolonged stasis where we now see rapid and recent change, implying a new and abnormal shift in the coastal regime.


Three distinct meltwater retention regimes in the West Greenland percolation zone revealed by comprehensive field studies

Horst MACHGUTH, Michael MacFERRIN, Dirk VAN AS, Charalampos CHARALAMPIDIS, William T. COLGAN, Robert S. FAUSTO, Andreas B. MICHELSEN, Jason E. BOX, Carl E. BØGGILD

Corresponding author: Horst Machguth

Corresponding author e-mail: homac@byg.dtu.dk

The retention and refreezing of meltwater in firn is presently one of the largest uncertainties in assessing the mass balance of the Greenland ice sheet. Two comprehensive field campaigns were carried out in spring 2012 and spring 2013 in West Greenland to (1) record the current state of the percolation zone and (2) infer recent meltwater percolation depth by quantifying changes in firn structure relative to legacy data. Twelve firn cores of 11–19 m length and a number of more shallow cores were drilled along an elevation transect at 67°N. The transect spans from 1690 to 2360 m a.s.l. and is 140 km in length. A continuous GPR profile and 2012/13 IceBridge flight line link all cores, AWS and thermistor string installations are operated along the profile. Analysing 2013 firn structure in the context of both 2012 and legacy observations reveals three distinct meltwater retention regimes. (1) At lower elevations (1750–1900 m a.s.l.) a massive (~8 m thick) near-surface ice layer overlies less dense firn. Comparison of the spring 2012 and 2013 cores shows that this massive perched ice layer prevented percolation of the extraordinary 2012 melt into the underlying firn. This elevation band experiences intensive surface runoff during extreme melt events. (2) At medium elevations (1900–2100 m a.s.l.) the firn contains numerous ice lenses homogenously distributed with depth. Comparison of 2013 and 1998 cores shows a uniform increase in firn density down to 15 m depth. We interpret this as evidence of very deep meltwater percolation. A limited amount of subsurface runoff cannot be ruled out. (3) At higher elevations (> 2100 m a.s.l.), where there is relatively little melt, meltwater refreezes at shallow depth. Comparison of 2013 and 1998 cores shows that the densification from increased melt in recent warm years is restricted to the near-surface layers. We interpret this as the absence of deep percolation; this elevation band retains virtually all meltwater, even in the most extreme melt years. The dataset provides evidence that climatic change triggers strongly non-linear reactions in the processes of meltwater percolation and retention. Our inference of enhanced runoff due to perched ice layers in the lower elevation firn may partially explain extraordinary river discharge observed in western Greenland during the summer of 2012.


Geometric modelling of DISP images and applications in regional Antarctic ice-sheet change estimation since the 1960s

Gang QIAO, Wenkai YE, Marco SCAIONI, Xiaohua TONG

Corresponding author: Gang Qiao

Corresponding author e-mail: qiaogang@tongji.edu.cn

Global sea-level change is one of the major challenges that all the nations are commonly facing, and changes in the Antarctic ice sheet have been playing a critical role in the global sea-level change research field in recent years. Long time series observations of the ice sheet in Antarctica would contribute to quantitative evaluation and prediction of the effects caused by the ice-sheet change. The declassification of cold-war satellite reconnaissance photographs, known as Declassified Intelligence Satellite Photographs (DISP), has provided a direct view of the Antarctic ice sheet’s configuration in the 1960s, greatly extending the limitation of Antarctic surface observations. This paper presents the results of our efforts in geometric modelling of the DISP images, as well as evaluation of ice-sheet changes in the Amery Ice Shelf and Byrd Glacier, Antarctica, during the past half-century. Part of the camera calibration parameters including focal length and lens distortions were collected from the camera calibration reports. The control points were acquired from the recent optical images and DEMs that covered the same region. The initial estimates of the exterior parameters were derived from the orbital ephemeris of specific missions, and they were then refined in a bundle adjustment process. The accuracy of the bundle adjustment was assessed and discussed. The DISP images were then orthorectified using the bundle adjustment results to generate mosaic, and the DEMs in the 1960s were also generated. These products were compared with the recent correspondences in the Amery Ice Shelf and Byrd Glacier, the changes were thus analysed, and the contributions to sea-level changes were discussed.


The effect of ice dielectric properties on the radar-derived basal conditions of West Antarctic ice streams


Corresponding author: David Ashmore

Corresponding author e-mail: d.ashmore@abdn.ac.uk

Recovering Antarctic basal conditions from ice-stream catchment-scale radio-echo sounding surveys is a major goal in the geophysical exploration of subglacial environments. The central objective of such studies is the retrieval of some measure of ice-bed reflectivity, but critically all such estimates are limited by the need to correct accurately for englacial attenuation, a function of both ice temperature and chemistry, which can vary considerably within an ice-stream catchment. Recent research has strongly advocated an approach in which englacial attenuation is modelled from known or estimated ice temperature and chemistry. These models require knowledge of glaciological parameters (bed topography, accumulation, etc.) but also ice dielectric properties (activation energies and conductivities). Ice dielectric properties used are typically mean values from a synthesis of laboratory-based experiments and are associated with large uncertainties. In this study we consider the effects of ice dielectric properties on the derivation of bed-returned power (BRP) in the Weddell Sea Sector in West Antarctica. Using temperature output from a steady-state finite-difference ice-sheet model and typical ice chemical concentrations we present Monte Carlo simulations of englacial attenuation and BRP, in which ice dielectric properties are allowed to vary within these uncertainties. From these analyses we show that in some scenarios the pattern of BRP is strongly influenced by the adopted ice dielectric properties and that errors associated with derived BRP measurements can be significant. We recommend caution when interpreting radar-derived basal conditions at these spatial scales and highlight the need for improved appraisals of the dielectric properties of glacial ice at a range of temperatures.


Using simulations of the Greenland ice sheet to help constrain its contribution to the Last Interglacial sea-level highstand


Corresponding author: Kerim H. Nisancioglu

Corresponding author e-mail: kerim@uib.no

The Greenland ice sheet is losing mass at an increasing rate, making it the primary contributor to global eustatic sea-level rise. Large melting areas and rapid thinning at its margins have raised concerns about its stability. However, it is conceivable that these observations represent the transient adjustment of the fastest-reacting parts of the ice sheet, masking slower processes that dominate the long-term fate of the ice sheet and its contribution to sea-level rise. Using simulated climate data from a comprehensive coupled climate model IPSL CM4, we simulate the Greenland ice sheet during the Last Interglacial with the three-dimensional ice-sheet model SICOPOLIS. The Last Interglacial is a period approximately 130 000–110 000 years before present with Arctic temperatures comparable to projections for the end of this century. In our simulation, the southwestern and northeastern parts of the ice sheet are unstable and retreat significantly. This result is found to be robust to perturbations within a wide parameter space of key parameters of the ice-sheet model, the choice of initial ice temperature, and has been reproduced with climate forcing from a second coupled climate model, the CCSM3. To investigate the contribution of the Greenland ice sheet to the Last Interglacial sea-level highstand we use the Greenland ice-core records. By combining estimated surface elevation changes of the ice for the Last Interglacial period, where available from ice cores, with a large ensemble of ice model simulations we are able to constrain the likely range of ice melt from Greenland during the Last Interglacial. Our best estimate of sea-level contribution from Greenland is compared with independent estimates based on marine proxy records as well as other model studies.


Mass balance of Antarctica: reassessment of grounding line fluxes and basins areas

Mathieu DEPOORTER, Jonathan BAMBER

Corresponding author: Mathieu Depoorter

Corresponding author e-mail: mathieu.depoorter@gmail.com

The mass balance (MB) of Antarctica is crucial to document its contribution to sea-level rise and to diagnose places of dynamic changes around its margins. Many studies from different research groups have used satellite altimetry and satellite gravimetry in order to estimate the continent’s MB, but only one research group has tackled the problem using the output-input method (IOB). In this study we reassess both the grounding line fluxes (GLF) and the surface mass balance (SMB). For the GLF we use a new (complete and up-to-date) grounding line, and more radar ice thickness data than in previous studies. SMB values are re-evaluated in light of new drainage basins definition.


The role of subaqueous melt in grounded ice. Does it really enhance calving?


Corresponding author: Sue Cook

Corresponding author e-mail: sue.cook@unis.no

Studies have shown that water temperatures in the Arctic, and in particular around Greenland and Alaska, are sufficiently high to cause significant rates of subaqueous melt. It is generally thought that as the calving front is undercut by melting below the waterline, stresses in the remaining ice will increase and drive higher calving rates. This mechanism has been hypothesized as the cause of recent terminus retreat in many Greenland outlet glaciers. Using the finite-element software Elmer/Ice to model a grounded tidewater glacier, we examine the effect of undercutting of the terminus in both static and time-evolving simulations. The results indicate that the common understanding that removal of supporting ice weakens the region around the terminus and causes calving may be oversimplified. When the geometry of the model is allowed to adapt in a time-evolving simulation, the response in stress field to undercutting is far more complex than previously assumed.


Using a control method to evaluate the effect of unpinning on the Roi Baudouin Ice shelf, East Antarctica

Lionel FAVIER, Rémy MERCENIER, Sophie BERGER, Reinhard DREWS, Emmanuel WITRANT, Frank PATTYN

Corresponding author: Lionel Favier

Corresponding author e-mail: lionel.favier@ulb.ac.be

Numerous seamounts emerge from the edge of the Continental Shelf in Dronning Maud Land. Those features translate into pinning points when they attach to the otherwise freely floating ice shelves from beneath, leading to buttressing and ice-shelf stability. The observed thinning of ice shelves, enhanced through Circumpolar Deep Water incursions across the shelf, makes pinning points lose their contact, hence increases ice flow speed. Here, we study the effects of de-pinning on the Roi Baudouin Ice Shelf, which is currently buttressed through a small pinning point that has a large effect on the ice dynamics. Using the shallow-shelf approximation as the background model, a control method is applied to infer the pattern of the rate factor A in Glen’s flow law from a very high resolution (50 m) InSAR flowfield, and using local data for the ice surface elevation. The inferred pattern for the rate factor is subsequently used to investigate the consequences of losing contact between the pinning point and the ice shelf.


A widespread increase in West Antarctic snow accumulation and its impact on ice-sheet surface elevation change


Corresponding author: Brooke Medley

Corresponding author e-mail: brooke.c.medley@nasa.gov

Understanding the temporal variability and trends in accumulation and their impact on firn densification is necessary to properly interpret ice-sheet surface elevation changes. While ice-core accumulation records provide insight into the temporal history of accumulation, their relatively small footprint limits their utility for mass-balance studies. Atmospheric models, which are spatially and temporally complete, provide the most comprehensive product for accounting for accumulation variability in elevation interpretation. But what if the firn column is still responding to accumulation changes initiated before the onset of the modeled product? Here we use ice-core records combined with airborne radar data to measure the change in accumulation between 1944–1984 and 1985–2009 and find a widespread increase in accumulation over much of Thwaites Glacier and part of Pine Island Glacier. Three ice cores collected in 2010/11 indicate a likely shift in accumulation around ~1980 towards higher values, indicating firn column thickening. We use a transient firn densification model to assess the timescale and magnitude of the firn column thickening over Thwaites and whether it should be considered in the interpretation of ice-sheet surface elevation changes from this region.


The temporal and spatial variations in the mass balance of Drangajökull ice cap (Iceland) from 1946 to 2011

Eyjólfur MAGNÚSSON, Joaquin Muñoz-Cobo BELART, Hálfdán ÁGÚSTSSON, Finnur PÁLSSON

Corresponding author: Eyjólfur Magnússon

Corresponding author e-mail: eyjolfm@hi.is

We present a record of the specific mass balance (SMB) of Drangajökull ice cap in NW Iceland covering an area of 143 km2 (in 2011). The record spans the period 1946–2011 and is based on DEMs deduced with photogrammetric processing of aerial photographs (1946, 1975, 1985 and 1994) and from lidar observations (2011). The geodetic mass-balance record shows negative SMB of ~–0.5 m w.e. a–1 in 1946–75, followed by periods of positive SMB: ~0.2 m w.e. a–1 in 1975–85 and ~0.3 m w.e. a–1 in 1985–94. Negative SMB of ~–0.6 m w.e. a–1 is derived for 1994–2011. There is a striking difference in the SMB for the western and eastern halves of Drangajökull. We obtain SMB of ~–0.4 m w.e. a–1 for the western part and ~–0.6 m w.e. a–1 for the eastern part in 1946–75, ~0.4 vs ~–0.1 m w.e. a–1 in 1975–85, ~0.6 ~–0.1 m w.e. a–1 in 1985–94 and ~–0.6 vs ~–0.5 m w.e. a–1 in 1994–2011. Values for the entire study period are ~–0.15 vs ~–0.45 m w.e. a–1 for the western and eastern part respectively. The eastern part decreased in area by 20% during the study period while the western shrank by only 4%. There is no apparent explanation for this difference revealed by comparison of summer temperature records from weather stations east and west of the ice cap. Atmospheric data extending back to 1957, produced by downscaling the ECMWF-analysis with the ARW atmospheric model (9 km grid resolution), indicate weak west-east trend in the winter precipitation but in the opposite direction to the observed trend in SMB. In 1975–94 when the difference in SMB between east and west was highest, the modeled winter precipitation was in general close to equal on both sides. The discrepancy between the trend in model winter precipitation and the observed SMB may be partly due to the coarse resolution of the model, as indicated by atmospheric data at 3 km resolution extending back to 1994. The spatial variability in winter precipitation may also not be representative for the actual trend in winter accumulation. The majority of the winter precipitation falls on Drangajökull in relatively cold wind from the northeast making snow drift from east to west probably more common than in the opposite direction. Our study emphasizes the difficulties of extrapolating the SMB from one glacier or glacier part to another even over short distances. Also the need for high-resolution downscaling of atmospheric models for modelling the mass balance of ice caps with similar dimensions as Drangajökull.


Mass-balance observations in northern Greenland and Jutulstraumen, Antarctica

Daniel STEINHAGE, Veit HELM, Susanne COERS, Angelika HUMBERT

Corresponding author: Daniel Steinhage

Corresponding author e-mail: daniel.steinhage@awi.de

The contribution of glaciers to sea-level change can be estimated by balancing the accumulation over a drainage basin and the flux across the grounding line. The mass flux across the grounding line can be estimated by using observations of the ice thickness at the grounding line and the flow velocities. Beside the quantitative contribution in terms of Gt a–1, the changes in the glaciers upstream are important for understanding of the factors causing the change. We have chosen areas in the northern part of Greenland, where glaciers with different kinds of driving factors in dynamics are located and an area in Antarctica in which satellite altimetry shows thickening in the past few years. We surveyed profiles along two glaciers in North Greenland, Hagen Brae and Academy Glacier, and the Jutulstraumen in Antarctica. Here we present the datasets of laser scanning and ice thickness acquired in 2013 and 2014 and an assessment of the causes of elevation change. At locations where we have already done a repeat survey, we present the elevation changes and compare with those obtained using CryoSat-2 data from 2011–2012. Additionally, we present our plans for mass-balance surveys in the next 5 years.


Using palaeoclimate constraints for simulations of the Greenland ice sheet contribution to future sea-level rise

Reinhard CALOV, Alexander ROBINSON, Andrey GANOPOLSKI

Corresponding author: Reinhard Calov

Corresponding author e-mail: calov@pik-potsdam.de

We present simulations with the polythermal ice sheet model SICOPOLIS, which is bi-directionally coupled with the regional climate model of intermediate complexity REMBO, in order to assess the uncertainties in the future contribution of the Greenland ice sheet to global sea-level rise. As a novel element of our model, we include a subgrid-scale ice discharge parameterization, which mimics the ice loss via small-scale outlet glaciers in a heuristic statistical approach. The range of free model parameters is determined via simulations over two glacial cycles using constraints from ice boreholes, in particular the ice elevation drop between Eemian and present-day at the NEEM location. Further constraints are the present-day mass-balance partition (ratio between precipitation and ice discharge) and the error in ice thickness for every gridpoint. The palaeo simulations show a dependence of the present-day initial state of the ice sheet on its shape for the short-term (some hundred years) future response, while the long-term response is practically not affected. This behaviour is mainly caused by simulated ice in the initial present-day state located at position different from observations. Using the large-ensemble model realizations found with our model constraints, we estimate the range of the future contribution of the Greenland ice sheet to sea-level rise under global warming. We address the short-term as well as the long-term perspective. In particular, our predictions enable a basic assessment of the persistence of an acceleration of the ice discharge via the Greenland outlet glaciers.


Monitoring the drainage system in the Greenland ablation zone by means of satellite-borne SAR


Corresponding author: Jan Wuite

Corresponding author e-mail: jan.wuite@enveo.at

Numerous studies indicate a rapid response of the Greenland ice sheet (GIS) to changing boundary conditions. A substantial part of the current mass imbalance is ascribed to dynamic thinning at the ice-sheet margins where meltwater can penetrate and lubricate the ice–bedrock interface facilitating ice flow. Monitoring the characteristics and evolution of the subglacial drainage system during the melt season is important as it influences the contribution of the GIS to global sea-level rise. Direct measurements however are hampered by the inaccessibility of the subglacial drainage system. One way to investigate the system indirectly is by means of satellite-borne SAR. We use recent high-resolution SAR data acquired by the TerraSAR-X and TanDEM-X satellites complemented with SPOT-5 SPIRIT DEM acquired during IPY to derive time series of ice motion, surface deformation and elevation change along the southwest GIS margin. Combining both ascending and descending repeat-pass SAR image pairs we are able to retrieve detailed 3-D surface displacements. The digital elevation models are generated from bistatic interferometric SAR data of the TSX-1/TDX-1 satellite configuration. The datasets are used to investigate short-term to interannual changes in surface topography and 3D displacement and permit us to study glacier hydraulics. We find that horizontal and vertical displacement rates show uplift and acceleration, which are likely related to summer rainfall events. The maps of surface elevation changes show local surface lowering and uplift due to changes in the water system and reveal a pattern of large intra- or subglacial drainage channels. A comparison of surface elevation changes derived from 3-D displacement with elevation changes from DEM differencing show good agreement. The elevation changes between June 2008 (SPOT) and September 2011 (Tandem-X) show significant mass depletion with surface lowering of up to ~10 m at the terminus. This study highlights the great potential of TanDEM-X and TerraSAR-X data for glacier studies.


The contribution of the ice sheets in Greenland and Antarctica to sea-level change in 2011–2012 estimated from CryoSat-2 data

Veit HELM, Angelika HUMBERT, Heinz MILLER

Corresponding author: Veit Helm

Corresponding author e-mail: veit.helm@awi.de

ESA’s CryoSat-2 satellite has been continuously observing Earth’s polar regions since April 2010 by using a sophisticated radar altimeter. This study focuses on the large ice sheets of Antarctica and Greenland and presents first elevation change maps and mass balance based on two years of CryoSat-2 data acquisition. Additionally, we derived new digital elevation models of Greenland and Antarctica, including error maps. Comparisons with ICESat data show that 95% of the DEMs have an error of less then 5 m ± 30 m. Using 2 years of CryoSat-2 data acquisitions we derived maps of elevation change for 2011 to 2012 in 1 km resolution for both Greenland and Antarctica. Negative elevation changes are concentrated at the west and southeast coast of Greenland and in the Amundsen Sea Embayment in West Antarctica (e.g. Pine Island and Thwaites Glaciers) agreeing well with dynamic mass loss observed between 2004 and 2008 using ICESat. In Dronning Maud Land, Antarctica, an accumulation anomaly led to considerable thickening of the ice sheet, influencing the contribution of the ice sheet to sea-level change.


Evolution of ocean freshwater forcing from Antarctic ice shelves over the past 20 years


Corresponding author: Nacho Merino

Corresponding author e-mail: ignacio.merino@lgge.obs.ujf-grenoble.fr

Freshwater run-off from Antarctica plays a crucial role in setting Southern Ocean properties and circulation. In consequence, observed speed-up of Antarctic outlet glaciers and associated increase of freshwater release may have a big influence on ocean dynamics and sea-ice formation. Increase in freshwater forcing is usually neglected or poorly considered by current ocean models but may contribute to explain the observed trends in the Southern Ocean. Studies of the oceanic impact of increase in freshwater run-off are made using ice-sheet mass-balance budgets. However, most ocean models apply an evenly distributed freshwater flux around the Antarctic coastline and do not consider their calving or melting origin. According to glaciological assessments on ice outflow such assumptions are misleading. Starting from a recent estimation of calving and melting fluxes of each ice shelf around Antarctica in 2010, and taking into account mass balance of both grounded ice sheet and floating ice shelves, we propose some possible scenarios of the spatial and temporal evolution of the freshwater run-off during the last two decades.


High-resolution observations of the topographic and dynamic evolution of Helheim Glacier’s calving front over a 2 year cycle


Corresponding author: Suzanne Bevan

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

Understanding how calving glaciers respond to warming oceans and atmosphere is critical to predicting global sea-level rise. The outlet glaciers of Greenland are of particular interest as their contribution to sea-level rise has doubled over recent years owing to widespread accelerations in flow speed and increased calving rates. The actual calving process is still not well understood, with controlling factors likely to vary seasonally and to be dependent on whether the glacier terminus is grounded or floating. Focussing on an individual glacier, Helheim in southeast Greenland, we will present high spatial resolution surface elevation and velocity measurements over the few kilometres closest to the calving front, to form a picture of changes that occur as the front advances, calves and retreats during a 2 year time period. The observations come from interferometric DEMs and feature-tracking of TanDEM-X SAR data acquired every 11 days from June 2011 to August 2013. During the entire period the front moves through a 2–3 km range, being at its most retreated at the end of the record and with calving occurring all year round. Time-lapse animations of the DEMs reveal the development of troughs in surface elevation close to the front. The troughs propagate down flow and develop into the rifts from which calving takes place. It is not uncommon to see extensive regions of rifts within the front 2 km penetrating to depths level with the melange. It is clear that for much of the time the front of the glacier is floating, therefore it is possible to estimate the total ice thickness and, in conjunction with velocity observations, to estimate total ice discharge to the fjord through the time series. The freeboard of icebergs within the fjord adjacent to the front is consistently 60 m or less, indicating tabular calving followed by rotation and break-up. Downstream profiles show a variety of characteristics depending on the relative advancement of the terminus. The final two image dates in the sequence coincide with a NERC Networks of Sensors project which saw a network of GPS tracked nodes deployed across the front of the glacier, and also the acquisition of an airborne lidar survey. Incorporating the lidar-generated DEM in the interferometric processing chain has allowed a vertical accuracy of around 2 m for the TanDEM-X DEMs. Data from the GPS nodes confirm that the glacier is at, or close to, floatation at this time, as a portion of the front of the glacier is sufficiently buoyant to be vertically displaced by ocean tides.


Greenland firn aquifer extent, volume and stability from in situ measurements, airborne radar and modeling


Corresponding author: Richard Forster

Corresponding author e-mail: rick.forster@geog.utah.edu

The accelerating mass loss from the Greenland ice sheet is a major contribution to current sea-level rise (SLR) with increased meltwater runoff responsible for half of this mass loss increase, yet the mechanisms and timescales involved in allowing surface meltwater to reach the ocean remain poorly understood. The recent discovery of an extensive liquid water reservoir within the Greenland ice sheet firn further complicates the relationship of melt to SLR, since the aquifer system may be either buffering or accelerating SLR. We use a combination of in situ measurements and firn modeling to estimate the Greenland firn aquifer area to be 70 000 ± 10 km2 with an estimated 140 ± 20 Gt water or 0.4 mm sea level equivalent. Repeated airborne radar surveys from NASA’s Operation IceBridge indicate the shape of the undulations in the upper boundary of the aquifer is stable similar to an unconfined terrestrial aquifer but there may be subtle changes in its extent. In April 2013, our team drilled through the aquifer for the first time to gain an understanding of firn structure constraining it, to estimate the water volume within the aquifer, and to measure the temperatures and densities above, within and below the aquifer. The temporal extent of the aquifer is investigated in areas where airborne radar data and/or the ground-based radar data were acquired along the same track and at the same time in both spring 2011 and spring 2013. Despite the intense surface melt of Greenland during the 2012 summer, only a slight increase of the water level in the firn aquifer is observed (+0.7 ± 0.3 m). The fate of this stored meltwater is currently unknown as the aquifer can be isolated from the englacial water network. However drainage into nearby crevasses seems likely based on combining visible imagery with airborne radar profiles, which indicate crevasses coincide with absence of the aquifer, but more work is required to test this hypothesis. Our team will return to the drill site in April 2014 to make additional measurements.


Glaciological inversions and their regularization


Corresponding author: Robert Arthern

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

One of the key advances that has allowed better simulations of the large ice sheets of Greenland and Antarctica has been the use of inverse methods. These have allowed poorly known parameters such as basal slipperiness, bed elevation, or ice viscosity to be constrained using a wide variety of satellite observations. Inverse methods used by glaciologists have broadly followed one of two related approaches. The first is minimization of a cost function that describes the misfit to the observations, often accompanied by some kind of explicit or implicit regularization that promotes smallness or smoothness in the inverted parameters. The second approach is a probabilistic framework that makes use of Bayes’ theorem to update prior assumptions about the probability of parameters, making use of data with known error estimates. Both approaches have much in common and questions of regularization often map onto implicit choices of prior probabilities that are made explicit in the Bayesian framework. In both approaches questions can arise that seem to demand subjective input. What should the functional form of the cost function be if there are alternatives? What kind of regularization should be applied, and how much? How should the prior probability distribution for a parameter such as basal slipperiness be specified when we know so little about the details of the subglacial environment? Here we consider some approaches that have been used to address these questions and discuss ways that probabilistic prior information used for regularizing glaciological inversions might be specified with greater objectivity.


Beyond back-stress: model experiments investigating the role of gradients in longitudinal stress modulating glacier flow


Corresponding author: C.J. van der Veen

Corresponding author e-mail: cjvdv@ku.edu

There is growing consensus in the glaciological community that observed changes in outlet glacier dynamics in Greenland and Antarctica can be attributed to perturbations at the ice–ocean interface. Several studies have suggested increases in longitudinal stress gradients as a trigger for causing subsequent large-scale changes in flow dynamics. Many outlet glaciers underwent changes in the extent of their floating ice tongue coincident with a sudden acceleration and thinning. In this canonical view, a perturbation in the balance of stresses at or near the grounding line gives rise to subsequent acceleration up-glacier; without the support or ‘back-stress’ provided by the ice shelf, the ice sheet rapidly relaxes to a new equilibrium characterized by a lower surface elevation throughout the drainage area. Numerical modeling studies appear to support this theory. Numerical models used to investigate potential physical forcing mechanisms responsible for dramatic glacier changes have become increasingly more sophisticated in recent years, and now include more of the processes believed to be important. In particular, flowband models incorporate explicit inclusion of gradients in longitudinal stress, as well as physically based treatment of iceberg calving at the glacier terminus. Application of these models to specific Greenland outlet glaciers has shown that observed glacier changes can be modeled satisfactorily by selecting the appropriate combination of essentially unconstrained model parameters. However, because of the multitude of non-linearly interacting processes included in these model simulations, it remains somewhat unclear as to what is the cause of glacier change and what may be feedback mechanisms amplifying initially small perturbations. The fundamental question that has been captivating – and arguably dividing – the glacier-dynamics community since publication of the pioneering papers on the stability of marine glaciers by John Mercer and Hans Weertman some 50 years ago, remains unanswered: What role do ‘back-stress’ and longitudinal stress gradients play in regulating glacier flow? To investigate this glaciological stalemate we present results from numerical simulations for idealized geometries, comparing results from flowband models that include longitudinal stress gradients and those based on a balance between driving stress and resistance from basal and lateral drag.


Ungrounding of an ice-stream pinning point due to ocean melting


Corresponding author: Daniel Goldberg

Corresponding author e-mail: dan.goldberg@ed.ac.uk

Pinning points play in important and sometimes overlooked role in the mass balance of an ice stream and the position of its grounding line. Buttressing due to grounding on a topographic rise can stabilize an ice stream which might otherwise be subject to collapse, and loss of contact with the pinning point can lead to a rapid shift in grounded balance. An important question, however, is how contact is lost with the pinning point in the first place. Submarine ice-shelf melting is often cited as an underlying cause; yet the causal link is necessarily indirect, since melting cannot act directly under the ice that is grounded on the pinning point. Rather, ocean heat must act through melting of the shelf, thinning the ice that is grounded on the pinning point by changing the stress pattern within the shelf. Using a coupled model of ice and ocean dynamics, I examine the process by which ocean melting can lead to the ungrounding of a pinning point and the subsequent retreat of the marine ice stream. I show that, in an ice shelf that is otherwise unconfined, there is a limited area over which ocean melting has importance, despite the relatively high melt rates that develop elsewhere. Thus, as the relative importance of the pinning point increases, the ability of ocean melting to impact the grounded ice stream diminishes.


A high-resolution dynamical downscaling of regional climate data for Hansbreen glacier (Svalbard)

Roman FINKELNBURG, Jaime OTERO, Francisco NAVARRO, Marco MÖLLER, Malgorzata BLASZCZYK, Bartlomiej LUKS, Tomasz BUDZIK

Corresponding author: Roman Finkelnburg

Corresponding author e-mail: roman.finkelnburg@upm.es

Spatially distributed modeling is a valuable tool to obtain information on the current state and to better understand the past and future of glacier mass fluctuations. However, much of the uncertainty about the present state of mass balance of ice masses and their local response to climate change is connected to uncertainty in forcing data. Simple statistical methods using for instance elevational gradients combined with digital elevation models are widely used to generate model forcing data based on observations at one or several automatic weather stations (AWS). In contrast, numerical weather prediction (NWP) models could be used to dynamically downscale and thus distribute atmospheric forcing on glaciers physically consistent. We present a dynamical downscaling approach for the Hansbreen glacier using the Polar WRF 3.4.1 model. A dataset from TU-Berlin – the European Arctic Reanalysis – comprising atmospheric fields, sea surface temperature and sea-ice concentration on 2 km resolution is used as input data for the downscaling on Hansbreen glacier. We used a digital elevation model and land cover data of 40 m resolution derived from airborne and satellite observations instead of the standard USGS datasets. The albedo ranges for respective land classes and snow have been determined from MODIS and field observations. The resulting dataset on 350 m horizontal resolution is compared with field observations in and around Hansbreen.


Estimating longwave atmospheric emissivity in the Canadian Rocky Mountains


Corresponding author: Samaneh Ebrahimi

Corresponding author e-mail: sebrahim@ucalgary.ca

Incoming longwave radiation is an important source of energy contributing to snow and glacier melt. However, estimating the incoming longwave radiation from the atmosphere is challenging due to the highly varying conditions of the atmosphere, especially cloudiness. We analyze the performance of some existing models including a physically based clear-sky model by Brutsaert (1987) and two different empirical models for all-sky conditions (Lhomme and others, 2007; Herrero and Polo, 2012) at Haig Glacier in the Canadian Rocky Mountains. Models are based on relations between readily observed near-surface meteorological data, including temperature, vapor pressure, relative humidity, and estimates of shortwave radiation transmissivity (i.e. clear-sky or cloud-cover indices). This class of models generally requires solar radiation data in order to obtain a proxy for cloud conditions. These are not always available for distributed models of glacier melt, and can have high spatial variations in regions of complex topography, which likely do not reflect the more homogeneous atmospheric longwave emissions. We therefore test longwave radiation parameterizations as a function of near-surface humidity and temperature variables, based on automatic weather station data (half-hourly and mean daily values) from 2004 to 2012. Results from comparative analysis of different incoming longwave radiation parameterizations showed that the locally calibrated model based on relative humidity and vapour pressure performs better than other published models. Performance is degraded but still better than standard cloud-index-based models when we transfer the model to another site, roughly 900 km away, Kwadacha Glacier in the northern Canadian Rockies.


Hydrologic budgets for ice streams on the West Antarctic Siple Coast


Corresponding author: Poul Christoffersen

Corresponding author e-mail: pc350@cam.ac.uk

Satellite observations have revealed active hydrologic systems beneath Antarctic ice streams including subglacial lakes, but the sources and sinks of water within these systems are uncertain. This is problematic because flow of water along the bed may be a crucial factor in maintaining the fast flow of the ice streams. Here we use numerical simulations of ice streams to estimate the generation, flux and overall budget of water beneath five ice streams on the Siple Coast. We estimate that 47% of the total hydrologic input (0.98 km3a–1) to Whillans (WIS), Mercer (MIS) and Kamb (KIS) ice streams comes from the ice-sheet interior, and that only 8% forms by local basal melting. The remaining 45% comes from a groundwater reservoir, an overlooked source in which depletion significantly exceeds recharge. Of the total input to Bindschadler (BIS) and MacAyeal (MacIS) ice streams (0.56  km3 a–1), 72% comes from the interior, 19% from groundwater and 9% from local melting. This contrasting hydrologic setting modulates the ice streams flow and thus the Antarctic contribution to sea-level change.


An assessment of the 20th century global mean sea-level rise in historical runs of CMIP5 ocean–atmosphere general circulation models


Corresponding author: Benoit Meyssignac

Corresponding author e-mail: benoit.meyssignac@legos.obs-mip.fr

Global mean sea level is a major indicator of climate change. Its variability integrates over time many changes that affect the different components of the climate system including the ocean, the cryosphere and the land surface. In this study we use it as an indicator to assess the ability of CMIP5 atmosphere–ocean general circulation models (GCMs) to reproduce the 20th century climate variability. We compare estimates of the 20th century sea-level rise (SLR) from GCMs with observations of sea level (from tide gauge records and satellite altimetry) and its contributors. The objective is to select the CMIP5 GCMs which simulate properly the observed 20th century SLR and to propose new projections of the SLR over the 21st century on this basis. In the first part of our work we estimate the different contributors to the 20th century SLR (thermal expansion of the ocean, glacier melt, Greenland SMB and land water storage) from outputs of CMIP5 GCM historical runs (which span the period 1850–2005). We particularly focus on the land ice contributions (glacier melt and Greenland SMB) because unlike other contributors they cannot be directly computed from GCMs outputs and need a further modelling step. Indeed, glaciers are too small to be modelled by the coarse CMIP5 GCMs. To estimate their contribution to sea level we use an offline glacier model developed by B. Marzeion forced by the regional surface air temperature (SAT) and precipitation outputs from CMIP5 GCMs. The Greenland SMB contribution to sea level is not correctly modelled either by CMIP5 GCMs because their coarse horizontal resolution limits their capability of capturing essential SMB changes on the narrow ablation zone. To overcome this problem, we use two different methods: (1) a downscaling method developed by M. Geyer, which gives the Greenland SMB from the SAT, the precipitation and the snowmelt outputs of the CMIP5 GCMs; and (2) an equation developed by X. Fettweis, which gives the Greenland SMB based on GCM global mean SAT output (calibrated against the Greenland regional climate model MAR). In the second part of our work, we compare the GCMs estimates of the different contributors to the 20th century global mean SLR with observations. We also sum up the different contributions and we compute the total SLR from each of the CMIP5 GCMs over 1900–2000. We find that only a few models give estimates of global mean sea level that are in agreement with all the observations.


Response of the Greenland ice sheet to climate changes


Corresponding author: Christine S. Hvidberg

Corresponding author e-mail: ch@gfy.ku.dk

The Greenland ice sheet responds to climate changes by a number of mechanisms. Climate changes affect directly the surface mass balance and ice discharge at the ice–ocean interface, and the ice sheet responds dynamically to retreat at the margin and changes in surface slope and flow in outlet glaciers. The Greenland ice sheet has overall retreated since the Last Glacial Maximum to its present configuration. The retreat has not been steady but occurred on shorter and longer timescales due to different processes. It is not yet known which processes are most important in determining the ice-sheet response to climate change in a warming climate and the contribution to future sea-level changes. In this work, we investigate the response of an ice-sheet model to climate forcing using a range of different model setups. We have used the Parallel Ice Sheet Model (PISM) to construct an ensemble of 1412 simulations of the evolution of the Greenland ice sheet from the Last Glacial Maximum to present day. The model was run with a range of different configurations of internal model parameters controlling ice-sheet flow and basal conditions combined with different climate forcing schemes. We have evaluated the ensemble of simulated present ice-sheet configurations against a set of data constraints. These data constraints include present-day observations from satellites as well as ice-core records and borehole temperatures. The evaluation shows that the model setups that simulate the present-day observations are not the same as the model setups that simulate the ice-sheet evolution constrained by ice-core records. The ensemble was used to investigate the range of responses to past climate forcing since the Last Glacial Maximum. We have used the modeled time evolutions of volume change together with the applied climate forcing from each simulation. We investigate then the relation between the volume response and the climate forcing in the ensemble, as well as the influence of individual model parameters and forcing schemes. Our preliminary results suggest that the response is mainly related to the forcing, and less sensitive to model parameters. Finally, we have investigated the range of responses within the ensemble to future climate scenarios in order to see which processes and key parameters of the ice-sheet model are important for predictions of the future ice-sheet contribution to sea-level rise.


Including the effect of wind in the simulation of snow accumulation patterns over an Alpine glacier: implication for seasonal mass balance


Corresponding author: Ilaria Clemenzi

Corresponding author e-mail: clemenzi@ifu.baug.ethz.ch

Snow accumulation is an essential component of glacier mass balance. In mass-balance models, however, it is often represented in a simplified way, neglecting the effect that gravitational redistribution and wind transport have on the spatial distribution of snow. In this study, we investigate the contribution of snow transport induced by wind to the snow-cover patterns over one glacier at the end of the accumulation season. We account for snow wind interaction processes such as the effect that wind exerts on precipitation, deposition and the removal of already deposited snow. For this, we reconstruct the snow depth patterns at the end of one accumulation season over Haut Glacier d’Arolla in the Swiss Alps using a physically based snow transport model. We force this model with high-resolution wind fields (25 m) obtained from WINDS, a mass-consistent wind model designed for complex terrain. Wind fields accounting for the effect of the local topography on the air flow are generated through a nesting procedure using initial wind fields with lower resolution. Boundary wind conditions are provided by the predictions of the non-hydrostatic limited-area atmospheric model COSMO 2.2. In addition, we simulate the accumulation patterns with the common assumption of spatially constant winds. The snow depth patterns simulated with the two different wind field settings are then compared and validated with distributed snow depth from a lidar digital elevation model available at the test site for the end of the accumulation season 2011. Mass balance and runoff obtained with the two settings are also compared. A significant improvement in snow depth patterns is achieved when topographically modified wind fields are considered. Hence, wind re/distribution contributes in a significant way to the generation of snow depth variability at our study site. Accounting for wind effects also results in different mass-balance and runoff patterns, which result from a combination of multiple effects, as snow is moved from upper areas to lower-lying glacier sections prone to melt but is also accumulated in sheltered areas where shading reduces ablation. These effects seem important at the scale of one alpine glacier and can only be accounted for by a physically based description of processes combined with high-resolution wind fields.


An interactive ice-sheet model for education

Martin O’LEARY

Corresponding author: Martin O’Leary

Corresponding author e-mail: m.e.w.oleary@gmail.com

Much of modern glaciology is focused on the develoment and application of large-scale numerical models. These models incorporate a wide range of processes at a high level of fidelity. Consequently, in order to work with these models, users often require considerable computational resources, as well as a high level of technical skill. These factors reflect the priorities of a research environment. In an educational context, or for simple experiments designed to build one’s intuitions, the priorities are very different. Ease of use is much more important, as is direct access to a range of outputs. Simple representations of individual processes may be more useful than a more complex model of interactions. These priorities lead to a very different kind of model, many of which exist, but few of which are publically available. Here we present a web-based interactive model of the Greenland and East Antarctic ice sheets, based on the popular GRANTISM model by Frank Pattyn. Using Javascript, the model code runs in the user’s web browser, allowing for a great deal of interactivity. Both the model forcings and state can be inspected and modified in ‘real time’, and on even modest modern computers and handheld devices, simulations of tens of thousands of years of evolution run in a matter of seconds. Because of the web-based approach, no technical knowledge is required to operate the model, but the results are easily exported for further analysis using more traditional tools.


Quantification by synchrotron X-ray topography of ice lattice distortion evolution during compression


Corresponding author: Armelle Philip

Corresponding author e-mail: armelle.philip@ujf-grenoble.fr

Deformation of polycrystalline ice leads to strain and orientation heterogeneities in each crystal due to its strong viscoplastic anisotropy. These intra-granular inhomogeneities play an important role in processes like strain-hardening, recrystallization or fracture. To understand the heterogeneity inception, ice crystals were deformed by compression and investigated by X-ray Bragg imaging at the European Synchrotron Radiation Facility (ESRF). First, this study required development of a technique to get very high quality crystals, and a cold cell able to contain a compression device compatible with the ESRF-BM05 beam line. Then, a new method was worked out that combines rocking curve imaging with section and pinhole topography, in order to measure lattice distortions in all three spatial dimensions. The three-dimensional rocking curve imaging (3D-RCI) allows us to quantify the angular misorientations and the mosaicity with a spatial resolution of about 50 × 50 × 50 μm3. The angular misorientation resolution is in the few micro-radians range. In the present paper 3D-RCI is used to follow the inception of the deformation process of one of the grains of a three-grained ice polycrystal: from a quasi-perfect crystal with an initial dislocation density less than 100 cm cm–3 up to a subgrain creation beginning from the triple junction. The knowledge of the whole crystalline orientation in the crystal volume has allowed us to calculate the gradient of lattice rotation and therefore the curvature tensor field in the grain. This latter tensor is associated with the density of geometrically necessary dislocation that increases during the deformation.


Towards an adaptive resolution multiscale model of Pine Island Glacier ice-shelf ocean cavity


Corresponding author: Adam Candy

Corresponding author e-mail: adam.candy@imperial.ac.uk

Recent observational studies have helped to constrain estimates of the melt behaviour underneath Pine Island Glacier (PIG). Generally, however, observations are limited due to the relatively inaccessible and inhospitable environment. A solid ice cover, up to many kilometres thick, bars access to the water column, so that observational data can only be obtained by inference from above, drilling holes through, or launching autonomous vehicles beneath, the ice. This is further exacerbated by the fact that results of these recent studies have implied a significant proportion of the melting (~80%) occurs in networks of sub-kilometre-scale basal channels close to the grounding line, some of the most inaccessible parts of sub-ice-shelf ocean cavities. Accurately representing these small-scale processes in conventional ocean models is a huge challenge even in focused regional studies, and will not be possible in global coupled climate simulations in the near future. We present the development of a new model of PIG that is capable of resolving the range of scales necessary to evaluate the melt distribution and forming processes that dominate. This is built on the fluidity model that simulates non-hydrostatic dynamics on meshes that, like the model of Timmermann and others (2012), can be unstructured. In this case, the grid can be unstructured in all three dimensions and use an anisotropic adaptive-in-time resolution to optimize the mesh and calculation in response to evolving solution dynamics. The parameterization of melting in this model has been validated in idealized cavity domains and a validation is underway for the dynamic treatment of the ice–ocean interface. The model is not limited to a vertical coordinate system, which enables it to accurately represent ice fronts, and small shallow features. We will discuss the development of this model of PIG; including the cavity domain, conforming to appropriately filtered boundaries generated from data collected during the BAS Autosub 2009 expedition, and the simulation of non-hydrostatic dynamics to date. This model has the potential to capture the high spatial variation seen in melt rates in the small-scale channels, and as a result provide valuable insights into the physical processes driving the observed large melting and modulation of ice–ocean interactions at kilometre scales.


The response of the Juneau Icefield, Alaska, to climate change

Regine HOCK, Aurora ROTH, Brian ANDERSON, Florian ZIEMEN

Corresponding author: Regine Hock

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

Glaciers in southeast Alaska are subject to the highest mass loss rates in Alaska. Here we model the response of the Juneau Icefield to climate change. The Juneau Icefield is the fifth largest icefield in the Western Hemisphere covering 3900 km2, including more than 40 large outlet glaciers. Due to its close proximity to the ocean it is located in a maritime climate setting. We model the icefield using two models of different complexity. The first model is an enhanced temperature-index model including potential direct radiation, which applies volume–area scaling to model glacier retreat. The second model applies an energy balance to model the surface mass changes and includes a two-dimensional flow model to update the glacier extent and hypsometry. The models are forced with gridded climate data and calibrated using available point observations of surface mass balance and surface ice velocity fields derived from offset tracking based on Landsat images. Both models produce results in good agreement with the available observations. However, preliminary results indicate that volume projections are sensitive to the choice of the approach taken in updating the glacier geometry in response to the surface mass changes, indicating that proper accounting of the elevation mass-balance feedback is important for glacier projections.


Beyond back-stress: complete surface reconstruction and change detection of Kangerlussuaq and Helheim Glaciers in East Greenland


Corresponding author: Toni Schenk

Corresponding author e-mail: afschenk@buffalo.edu

Over the last two decades many outlet glaciers in Greenland have undergone rapid changes involving glacier thinning and flow acceleration. To elucidate possible physical causes that may have initiated these changes it is important to reconstruct detailed histories of elevation and velocity change. Of particular importance is to establish whether the onset of thinning preceded speed-up or vice versa. While records of surface velocity generally have high temporal resolution, reconstructing similar temporal records of elevation change is more challenging because of the many different sensors used to collect surface elevation data, including satellite laser altimetry observations from ICESat and airborne laser altimetry data from ATM and LVIS. Additionally, stereoscopic DEMs from satellite stereo imaging systems have also become available. Generally, these DEMs are confined to the coastal regions of ice sheets and are one to two orders of magnitude less accurate than laser altimetry points. On the other hand, their spatial coverage is nearly continuous while repeat altimetry is restricted to narrow profiles or swath widths of a few hundred meters. Thus, it makes sense to fuse these disparate datasets to obtain surface elevation and change rate histories over nearly continuous regions. We have recently developed SERAC for calculating surface elevation and elevation change. Originally developed for ICESat laser data, we have extended SERAC to include airborne laser altimetry data from ATM and LVIS. In this study we extend the calculation of elevation change histories by including stereoscopic DEMs in SERAC. Since stereoscopic DEMs are known to have systematic errors it is important to detect and remove those errors. We introduce a novel approach to match the DEMs to the laser altimetry points and to compute correction vectors that will be used in a subsequent height adjustment of the DEMs. The feasibility of this unique fusion approach is demonstrated at Kangerlussuaq and Helheim Glaciers in East Greenland. These glaciers exhibited a short episode of spectacular thinning, retreat and speed-up during the last decade, often used as a basis for developing predictive models of future dynamic mass loss of the Greenland ice sheet. Here we present the first detailed reconstruction of elevation change histories of these glaciers, combined with surface velocity to investigate the processes responsible for initiating and sustaining the rapid changes.


Organized band-like structures in driving and basal stresses of the Antarctic and Greenland ice sheets


Corresponding author: Olga Sergienko

Corresponding author e-mail: osergien@princeton.edu

Analysis of recently released high-resolution ice-sheet-wide datasets reveals the band-like spatial organization of driving stresses of the Greenland and Antarctic ice sheets. Narrow ~5–20 km bands of very high (300–400 kPa) driving stress, aligned approximately transverse to ice flow, are alternate with swaths of very low (~10–50 kPa) driving stresses. Computations of the ice deformational velocities indicate that these driving-stress spatial structures are characteristic of areas with basal sliding. High spatial resolution inversions for basal traction using full-Stokes mechanical models performed at selected locations on the both ice sheets show the presence of spatial organization in the basal shear on similar scales; however, narrow bands of high basal shear are not aligned with those in the driving stress, but rather oriented at an angle which has some consistency for individual outlets but varies between outlets. We argue that the regularity is due to a non-linear pattern-forming process in the basal zone of the ice sheet due to coupling between ice flow, water flow and perhaps till flow as well. It is likely that successful representation of this pattern-forming process in ice-sheet models is a crucial diagnostic of successful representation of basal process modelling.


Ice-shelf melt channels as a pattern-formation process and their sensitivity to oceanic and glaciological forcings


Corresponding author: Olga Sergienko

Corresponding author e-mail: osergien@princeton.edu

The widespread presence of melt channels under a number of Antarctic ice shelves gives rise to the question of their formation processes, evolution and sensitivity to oceanic and glaciological parameters. Numerical simulations with a fully coupled ice-shelf/sub-ice-shelf cavity model are presented, which show that melt channels form in the vicinity of the grounding line; they develop and evolve as a result of advection and deformation of ice-shelf flow. These channels seem to be formed by a pattern-forming instability, as their spacing is not directly controlled by lateral variations prescribed in the model. They form in circumstances where ice-shelf melt rates represent a substantial contribution to the ice-shelf mass balance. Sensitivity studies show that the geometric characteristics of the melt channels are more sensitive to the oceanic forcing, while the location of their formation is controlled by glaciologial parameters.


Response of interior ice to changes at the ice-sheet margin


Corresponding author: Michelle Koutnik

Corresponding author e-mail: mkoutnik@uw.edu

The margins of Greenland and Antarctica are currently exhibiting rapid and significant changes, which have a measurable effect on the flow of ice upstream. However, we do not yet know how much of the interior ice reservoir can be tapped by a dynamic margin, and how fast interior ice could be evacuated: How sustained does the rate of ice discharge and area of margin change need to be over time to significantly change the geometry of an ice-sheet interior? To address this question as a community we rely on whole-catchment or continent-wide (3-D) ice-sheet models. There have been major recent efforts to compare results from continent-wide models of Greenland and Antarctica to address ice-sheet sensitivity to changes in the ocean and atmosphere, especially at the ice–ocean boundary, and to estimate the future ice-sheet contribution to sea level (e.g. SeaRISE project). Comparison of numerous different models was a key part of this effort, which has produced a valuable archive of model runs. We use available archived model runs for West Antarctica, analyzing 2-D transects following flowlines that extend from select margin positions to the interior across the divide. Along these transects we compare the ice response using runs from different models under the same forcing in order to capture the range in modeled ice-sheet behavior, especially due to the range in model treatment of grounding-line evolution. For specific positions on these transects we assess the upstream propagation of velocity and thickness changes that initiate near the margin. This work is motivated by our need to merge understanding from 3-D models that represent large areas at relatively low resolution in our applications of limited-domain 2-D models at relatively high resolution. Ice-sheet models that limit the calculation domain (e.g. in the vicinity of the ice divide) can be run at higher resolution over longer timescales and are used for computationally expensive problems, including inverse problems and in support of ice-core interpretation. However, the interior boundaries of these limited-domain models are poorly constrained, and need to evolve consistent with the larger ice sheet. We show how behavior characteristics from 3-D models help constrain these 2-D problems, and how inferences from 2-D models can inform 3-D ice-sheet constructions.


The global GLAC-1c deglaciation chronology, meltwater pulse 1-a, and a question of missing ice


Corresponding author: Lev Tarasov

Corresponding author e-mail: lev@mun.ca

Unlike other global deglacial chronologies, GLAC-1c includes a posterior distribution of deglacial chronologies (currently for the three largest ice sheets) relative to observational constraints. As such, confidence intervals can be defined. Another distinguishing feature is that Eemian ground surface elevation is determined from that which is needed (to a first-order approximation) to dynamically achieve present-day topography after the last glacial cycle. The Eurasian and North American components are from recently completed Bayesian calibrations of the 3-D MUN glacial systems model (GSM), while the Antarctic component is from an initial scored ensemble of 3344 runs with the MUN/PSU GSM. Both versions of the GSM include thermomechanically coupled glaciological ice-sheet models, visco-elastic bedrock response with the VM5a Earth rheology, various climate representations, and a range of components to enable comparison of model output against observational records. The calibration and scoring is against a diverse and large set of constraint data, including relative sea level (RSL), marine limits, strandline elevations, present-day rates of uplift, and the current configuration of the Antarctic ice sheet. The ice sheet for each continent is independently calibrated against near-field constraints. The interim Greenland chronology is from an earlier glaciological model hand-tuned against RSL data. Meltwater pulse 1-a contributions are predominantly from North America (9–13 m eustatic sea level equivalent (mESL)). Eurasia contributes 2–4.5 mESL, while the Antarctic contribution is less than 2 mESL. An ongoing issue is an apparent shortfall of at least 10 mESL when model results are compared against far-field RSL datasets. Such a shortfall is not new. An examination of the evolution of past geophysically constrained global deglacial ice sheet and ice load reconstructions will reveal a reliance of sticking extra ice where there’s no data in order to fit far-field constraints. As new data has arisen, this extra ice load has been sequentially shifted to new regions. I will describe the key observational and physical constraints that limit continental ice volumes in GLAC-1c, and finish with a few ideas of how this shortfall may be resolved.


Three-dimensional simulations of Glaciar Zongo (Bolivia, 16°S) over the next century, based on different IPCC scenarios


Corresponding author: Marion Reveillet

Corresponding author e-mail: marion.reveillet@lgge.obs.ujf-grenoble.fr

In Bolivia, the Cordillera Oriental gathers more than 1800 glaciers covering an area of about 450 km2. Water resources from these glaciers are of major concern for the Bolivian Altiplano where the capital of Bolivia, La Paz, is located. Glaciar Zongo is located about 30 km from La Paz and flows on the southeast side of the Huayna Potosi which culminates at 6088 m a.s.l. This glacier has been monitored within the context of the GLACIOCLIM Observatory since 1991 and has an extensive network of glaciological, hydrological and meteorological measurements. Studies show a significant retreat since the last maximum of the Little Ice Age, with an acceleration since the late 1970s in relation to climate changes. A field campaign conducted in August 2012 allowed measurement, in different accessible parts of the glacier, of the ice thickness using ice-penetrating radar in order to map the glaciers bedrock. This bedrock mapping allows a detailed study of the glacier’s dynamics in order to project its evolution over the next century. In this study, the three-dimensional thermomechanically coupled full-Stokes model Elmer/Ice, mainly developped by the LGGE (EDGE team), is used and adapted in order to simulate the evolution of Glaciar Zongo. First, simulations have been performed over the current period (1997–2006) with the available field data used as input parameters (surface topography, bedrock, surface mass balance, surface velocities). After validation over this current period, using other field data (mapping of the outline of the ablation area measured by DGPS) and volume changes calculated for the period 1997–2006, simulations of the evolution over the next century were conducted. The atmospheric temperatures for the study area (15–18°S and 68–70°W) from different models included in the CMIP5 project are used considering three IPCC scenarios over the next century (the two extreme scenarios RCP 2.6 and 8.6, and the scenario RCP 6.0 considering as intermediate). Based on these temperature changes the future annual postition of the ELA is calculated, which enables us to know the shape vertical mass-balance profile used as input for the model. According to intermediate scenarios (RCP 6.0), Glaciar Zongo would lose about 60% of its current volume over the next century, with most of this decrease in the first half of 21th century.


Derivation of a multitemporal set of transient snowlines for Alaska’s glaciers

Christian KIENHOLZ, Anthony ARENDT, Regine HOCK

Corresponding author: Christian Kienholz

Corresponding author e-mail: christian.kienholz@gi.alaska.edu

Estimates of snow accumulation are important for assessing seasonal glacier mass balance and calibrating or validating mass-balance models. However, due to the large spatial variability typical of many glaciers, traditional stake networks are often not dense enough to accurately estimate the spatial distribution of snow cover. Previous studies indicate that accumulation distribution can be derived from transient snowlines through backward melt modeling, but the automated derivation of snowlines remains challenging. Here, we seek to (1) develop an algorithm that automatically identifies transient snowlines from remote-sensing imagery, (2) locate transient snowlines for ~90 000 km2 of glaciers in Alaska and neighboring Canada, and (3) investigate spatial and temporal variability, as well as trends, in snowline elevations. We have developed a largely automated workflow to derive transient snowlines through spectral classification of Landsat satellite imagery. Digital elevation model sampling along the snowlines, followed by a statistical analysis of the resulting snowline elevations, allows quantification of variability and trends from individual glaciers to entire regions. Glacier center lines provide the means to examine the variability for individual glacier branches, and a recently completed glacier inventory allows comprehensive interpretation of the results. Preliminary results from southern Alaska indicate large spatial variability of snowline elevation for individual glacier branches of the same glacier, with transient snowline elevations varying > 500 m within distances of less than 1 km. On a regional scale, we find significant trends related to the glaciers’ distance from the coast, with snowlines 100 km inland more than 1000 m higher than snowlines along the coast.


Observational evidence for transient fast flow at the South Pole: basis for modeling complex initiation scenarios in the last glacial cycle

Marie G.P. CAVITTE, Donald D. BLANKENSHIP, Jesse V. JOHNSON, Duncan A. YOUNG, Sasha P. CARTER, Gail R. GUTOWSKI, Martin J. SIEGERT, Charles S. JACKSON

Corresponding author: Marie G.P. Cavitte

Corresponding author e-mail: mariecavitte@gmail.com

Many parts of the East Antarctic ice sheet remain to be explored, in particular the South Pole region where the lack of satellite data has limited our understanding of a key part of the East Antarctic interior. The South Pole has been assumed to be a very stable part of the ice sheet, which drove the siting of the IceCube neutrino detection experiment. However, airborne radar data collected by UTIG between the South Pole and the Transantarctic Mountains has revealed a history of fast flow transients as evidenced from the disturbed radar layer record. We further constrain the spatial and temporal extent of these perturbations in the ice-sheet flow through the tracing and analysis of a set of extensive radar layers. This radio-stratigraphy has been age correlated with the IceCube dust age–depth record and the SPRESSO core timescale; vertical variations in age–depth across the South Pole region indicate enhanced flow at the height of the Last Glacial Maximum, between 50 ka and 10 ka. The South Pole radio-stratigraphy shows distinct ice-stream margins with an asymmetry suggestive of a complex temporal evolution of the margins. Paleo accumulation rates reconstructed from the radar layers via 1-D strain modeling show anomalous linear accumulation features, suggestive of localized enhanced melt rates, consistent with the presence of subglacial lakes in the area. The dynamic relationship between accumulation rates and melt rates due to ice streaming is used to reconstruct potential ice streaming velocities in the past. We numerically investigate a variety of ice-stream geometries and potential source regions as well as the spatial evolution of their margins through time, using a transverse 2-D flow model. We hypothesize the distribution of basal melt corresponding to the thermal evolution scenarios for the South Pole transients’ source regions. Presence of fast flow reaching deep into the interior of the Antarctic ice sheet has crucial ramifications for ice-sheet stability, with implications for rapid sea-level rise. The South Pole’s complex flow history is an important piece of our understanding of the Weddell Sea Embayment evolution in the last glacial cycle.


A current perspective on modelling subglacial hydrology


Corresponding author: Gwenn Flowers

Corresponding author e-mail: gflowers@sfu.ca

Models of glacier hydrology were initially developed to quantify the glacier contribution to the timing and volume of runoff. They have now come into the mainstream with the recognition that the retention, evacuation and distribution of water at the glacier bed is a fundamental control on ice dynamics, and thus on much of what we find interesting and observable at the ice surface. Early attempts to model the glacier drainage system borrowed principles from groundwater hydrology, with calculations of fluid potential and construction of flow nets used to determine hypothetical fluid pathways through the glacier system. Subsequent work focused on the conceptual and mathematical description of individual elements of the basal drainage system: ice-walled conduits, cavities formed in the lee of bedrock obstacles, water sheets, canals incised into subglacial sediment and bedrock channels. In some cases, natural experiments in the form of glacier surges or outburst floods inspired new conceptual models and furnished the data with which to test them. These experiments revealed the dynamic nature of the drainage system in which morphological transitions between fast/channelized and slow/distributed elements play an important role. The next generation of models sought to incorporate key features of the above elements in a continuum framework. This required compromise in the explicit representation of individual drainage elements, or other restrictive assumptions. Such compromise, however, enabled the first hydrologically coupled ice-flow models to be developed. The latest developments have seen many of the earlier obstacles overcome through a two-dimensional numerical framework that allows a slow (continuum) drainage system to interact with a network of discrete channels in a manner that permits realistic and objective transient evolution of drainage system morphology. With growing interest in whole-ice-sheet models of basal hydrology, near-term developments will require parsimonious representation of drainage-system elements while retaining the salient influences of hydrology on basal processes. Future efforts should consider whether a more unified treatment of hard- and soft-bed hydrology is warranted, as well as how best to represent basin-scale drainage networks in continental-scale models.


Oscillatory subglacial drainage dynamics

Christian SCHOOF

Corresponding author: Christian Schoof

Corresponding author e-mail: cschoof@eos.ubc.ca

Changes in subglacial drainage systems are known to alter ice flow speeds, and are therefore relevant to ice discharge and surface lowering in valley glaciers and the marginal areas of the Greenland ice sheet. Research in this area has focused heavily on the effect of drainage channelization, and on the effect of water input variations on basal water pressure through the time lag between instantaneous water supply and the evolving capacity of the drainage system to accommodate that water. In this presentation, I investigate dynamic changes in drainage systems that can occur in the absence of any variations in water input. Motivated by field observations at a valley glacier in the St Elias Mountains, Yukon Territoty, I show how interactions between water storage and the evolution of the drainage system can give rise to self-sustaining oscillatory pressure variations. These are similar to jökulhlaups but do not require a localized water storage body such as a lake, and can instead be driven by water storage that is distributed along the flow path. I show that water pressure oscillations driven by water storage in a lake occur only when the drainage system is channelized, while distributed storage allows oscillatory drainage even when the drainage system consists of linked cavities. The characteristics of the oscillations differ between channelized and linked-cavity system: channelized systems give rise to pressure oscillations that are in phase along the flow path, while oscillations in linked-cavity systems take the form of waves. Key to oscillatory drainage in both cases is that there is water supply above a lower threshold, and that the drainage system is sufficiently long. I also discuss possible implications of these water pressure oscillations for the dynamics of the overlying ice mass.


Investigating controls on submarine ice melt for an idealized Greenlandic outlet glacier fjord system


Corresponding author: Andrew Sole

Corresponding author e-mail: a.sole@sheffield.ac.uk

Recent observations suggest that the delivery of warmer ocean waters to the Greenland ice sheet margin may be the major driver of the observed thinning as a result of enhanced submarine melting and increased rates of iceberg calving. However, while ocean waters off the coast of Greenland have warmed in the last decade, it is unclear how or whether these warm waters actually access the front of glaciers which are often located at the head of long narrow fjords tens of kilometres from the warm ocean shelf waters. Here we use the non-hydrostatic MIT general circulation model to investigate the key controls on water circulation and submarine ice melt within an idealized Greenlandic outlet glacier fjord system. The model domain incorporates a 60 × 6 km by 600 m fjord, bounded by a large glacier at its head and a 30 × 20 km portion of coastal ocean at its mouth. This idealized geometry is scaled on many of the large outlet glacier fjord systems along Greenland’s east coast and enables us to examine the effect of controls on fjord circulation in the absence of confounding fjord-specific factors. We simulate the effects on submarine melting of varying ice-sheet runoff, tides, coastal ocean properties, alongshore winds and fjord geometry (sill depth, fjord sinuosity and fjord entrance shape).


Initialization of an ice-sheet model for present-day Greenland and use in assessing the impact of iceberg calving

Victoria LEE, Antony PAYNE, Stephen CORNFORD, Andrew TAYLOR

Corresponding author: Victoria Lee

Corresponding author e-mail: v.lee@bristol.ac.uk

We generate realistic initial conditions for an ice-sheet model that uses adaptive mesh refinement and a calving law for the Greenland ice sheet (GrIS). These allow the ice-sheet model to assess sea-level rise due to changes in ice dynamics. GrIS has been losing mass at an increasing rate over the last two decades and a significant proportion is due to dynamic thinning of narrow outlet glaciers. It has been suggested that this thinning maybe linked to changes in the stress balance at the glaciers’ front caused by calving. We solve a control problem with a weighted cost function to constrain coefficients of viscosity and basal drag using observed geometry and surface velocities. Using parameter fields generated by the control problem, we allow the ice surface to adjust to anomalies caused by uncertainties in bedrock elevation in a time-dependent run forced by a constant present-day climate with fixed calving fronts. The model uses a base grid with a resolution of 8 km which it refines to 2 km around the margins and is able to further refine to 500 m in areas of fast-flowing ice to capture the behaviour of the main outlet glaciers. The initial state of GrIS is obtained by repeating the time-dependent run with a calving model that determines the position of the calving front as the point where surface and basal crevasses penetrate the full thickness of the ice. We then investigate the effects of a sudden change in the calving rate using 100 year runs with different values of the crevasse water depth parameter. We compare changes in the modelled ice velocity of the main outlet glaciers with observations. We estimate a contribution to sea level from ice dynamics using these perturbation experiments.


Estimating Antarctic ice sheet surface mass-balance contribution to future sea-level rise using the regional atmospheric climate model MAR


Corresponding author: Cécile Agosta

Corresponding author e-mail: cecile.agosta@gmail.com

The Antarctic ice sheet surface mass balance (SMB) is a significant contribution to sea-level changes which may mitigate the rise in sea level in a warmer climate, but this term is still poorly known. The Antarctic SMB cannot be directly deduced from global climate models (GCMs) because of their too low resolution (~100 km) and their unadapted physic over cold and snow-covered areas. That is why the use of a regional climate model (RCM) specifically developed for polar regions is particularly relevant. We present here new estimations of the Antarctic SMB changes for the 20th and 21st centuries at 40 km resolution with the MAR (Modèle Atmosphérique Régional) RCM. Recent studies showed that large-scale forcing from GCMs was the main source of uncertainty for RCM-deduced SMB, thus we first present a careful analysis of the CMIP5 GCMs (used in the AR5 IPCC report) compared with the ERA-Interim reanalysis over the Antarctic region, from which we could select the less biased large-scale forcing for MAR. We thus show the Antarctic SMB evolution as modeled with MAR forced by ACCESS1-3 for RCP 4.5 and 8.5 greenhouse gas scenarios. We evaluate our outputs by comparing MAR forced by ACCESS1-3 and ERA-Interim for the 1980–2000 period to more than 2700 quality-controlled observations and to surface meteorological data from the READER database. We then give SMB change estimations for the 21st century together with an analysis of uncertainties coming from the MAR model, the GCM forcing and the greenhouse gas scenarios.


Sliding velocity fluctuations and subglacial hydrology over the last two decades, Argentiere glacier, Mont Blanc area


Corresponding author: Luc Moreau

Corresponding author e-mail: moreauluc@club-internet.fr

The dynamic behavior of temperate glaciers is largely driven by the sliding velocities. However, the relationships between the sliding velocities and the subglacial processes are poorly understood. It is mainly due to the lack of subglacial observations. Hydroelectric developments beneath the Argentiere glacier have established a subglacial observatory that allows us to directly study the complex interactions between the sliding velocity and the subglacial runoff. The sliding movement has been monitored almost continuously since 1990. The surveys have been carried out thanks to a bicycle wheel in contact with the basal ice and fixed to the bedrock. The annual sliding velocities have decreased over the last two decades concomitantly with the surface mass balance and the thickness decrease of the glacier. In addition, the records reveal strong seasonal variations which have been compared to the subglacial runoff. Short-term acceleration events are associated with periods of rapidly increasing water inputs to the subglacial drainage systems.


A combination of Envisat radar altimetry repeat-track and crossover results of Greenland surface elevation changes


Corresponding author: Rakia Meister

Corresponding author e-mail: rame@space.dtu.dk

Envisat radar altimetry has provided a way of remotely measuring surface height changes of the Greenland ice sheet for the period 2002–2012. In this work, previously derived dH/dt using a crossover method, which provides data in the interior of the ice sheet, is combined with a newly developed repeat-track method. This study is the first of its kind to explore the use of repeat-track radar altimetry data to derive Greenland surface height changes. Because of the large footprint of the Envisat RA-2 altimeter (on the order of 5 km), it has previously been assumed that towards the edges of the ice sheet, where steep slopes exist, radar altimetry data cannot be used to resolve changes with sufficient accuracy. The repeat-track method makes use of all available observations and hence increases the number of measurements compared with the crossover method, thus increasing the spatial extent of measurements close towards the edge of the ice sheet. The high spatial resolution of the repeat-track method can be combined with the comparatively low errors of the crossover method. Thus, a new, reliable radar altimetry estimate of ice-sheet surface height changes is presented. We show that the repeat-track method provides results similar in pattern to the dH/dt rates obtained from ICESat laser altimetry. During the final 2 years of Envisat’s operation, the satellite was in a repeat orbit. However, the attitude was not controlled during that time which causes the ground tracks to drift. For the repeat-track method, this requires a different analysis approach than the one implemented for the first 8 years. We will report on the intricacies of deriving surface height changes from altimetry; on developing different approaches for a repeat-track analysis; and on combining results obtained from different methods.


Seismic observations of Beardmore Glacier, Antarctica


Corresponding author: J. Paul Winberry

Corresponding author e-mail: paul.winberry@gmail.com

We report initial results of our project to determine the dynamic sensitivity of outlet glaciers to forcing at the grounding lines in order to assess the contribution of East Antarctica to sea-level rise when the surrounding ice shelves collapse. In this project, the specific focus is Beardmore Glacier, an outlet glacier that discharges into the Ross Ice Shelf. Goals are to address the following: What is the magnitude and spatial pattern of basal resistance? What are the bed conditions? Does basal motion occur through sliding over a hard bed or deformation of basal sediments? What controls the spatial and temporal patterns of sliding and basal drag? How will collapse of the Ross Ice Shelf affect the flow of outlet glaciers? How fast and how much ice could be drawdown from East Antarctica? In this presentation, we focus on initial active-source seismic observations. First results from seismic measurements revealed that ice in the vicinity of camp is up to 3100 m thick. Surface elevation of camp is 910 m; the bed of the glacier there is more than 2200 m below sea level. Surface velocities in this region are ~300 m a–1. Such a deep trough has also been detected on Byrd Glacier by a recent CReSIS airborne radar survey. These deep troughs exert strong influence on the dynamics of outlet glaciers. In contrast, ice thickness on the slow-moving (~50 m a–1) ‘sticky spot’ is only 900–1000 m. Change in ice thickness is the primary control on the observed pattern of surface velocity.


A simple damage evolution law for ice shelves


Corresponding author: Jeremy Bassis

Corresponding author e-mail: jbassis@umich.edu

Basal melting and iceberg calving are the primary mechanisms responsible for transferring mass from the ice shelves to the ocean. Although the connection between basal melting and ocean forcing is clear, the effect of ocean forcing on iceberg calving remains more controversial with conflicting hypothesis about whether a warming ocean will increase or decrease future iceberg production. Previous theories of iceberg calving have either invoked fracture mechanics to explain the fracture process that precedes calving or sought to use damage mechanics to simulate the bulk behavior of fractures within the ice. Here we use a perturbation analysis to link the creep deformation of brittle crevasses after failure to bulk ‘damage’ within the ice shelf. This allows us to derive a damage evolution law that links brittle failure with creep failure and can be directly applied to simulate the progressive increase or decrease of damage within ice shelves. One of the advantages of this formalism is that it accounts for the effect of surface meltwater related hydrofracturing on damage progression. Moreover, the theory provides an explicit link between damage accumulation and the large-scale melting or refreezing regime of the ice shelf. For example, we find that marine ice accretion within crevasses substantially diminishes damage. In contrast, we find that large basal melt rates can slowly increase damage, but over decadal and longer timescales. This suggests that regions like Pine Island and Thwaites Glaciers that have experienced enhanced basal melt may also experience enhanced calving in the coming decades. Crucially, our results are sensitive to melt/accumulation rates at the scale of an individual crevasses, suggesting that small-scale thermodynamic interaction between crevasses and the ocean may play a pivotal role in the stability of ice shelves.


Beyond back-stress: dynamics of Greenland outlet glaciers from altimetry record


Corresponding author: Beáta Csathó

Corresponding author e-mail: bcsatho@buffalo.edu

Comprehensive satellite monitoring of the Greenland ice sheet (GrIS) has revealed increasing mass loss since the late 1990s. Dynamic processes have been responsible for as much as half of this estimated loss, including ice flow adjustments to past climate variations and contemporary atmospheric and oceanic forcings. Here, we present a reconstruction of GrIS outlet glacier elevation changes, derived from NASA’s 1993–2013 satellite and airborne laser altimetry observations by using the Surface Elevation Reconstruction and Change (SERAC) method. The altimetry record shows that continuing dynamic thinning provides a substantial contribution to Greenland mass loss. For most of the outlet glaciers, rates of elevation change decrease toward the interior, consistent with dynamic thinning initiated at lower elevations. To facilitate interpretation, we divided the dynamic adjustment patterns into a few distinct groups. The results show that members of the different glacier groups are not confined to specific regions and that nearby glaciers can exhibit very different temporal behavior. We detected synchronous elevation changes over large distances confirming the importance of regional climate forcing for controlling recent mass changes. However, outlet glacier dynamics also exhibits large spatial and temporal variability, casting doubt on models that attribute observed flow acceleration and thinning to a single mechanism, such as disintegration of floating ice tongues, or intrusion of warm ocean waters. Rather, these observations suggest that the response of individual glaciers to external forcing is more involved and may depend on local factors such as bed topography, size of the drainage basin or the evolution of the subglacial hydrology. Rather than attempt to analyze the behavior of each outlet glacier, we focus on the limited number of distinct glacier responses derived from the altimetry record and seek to explain these behaviors. To gain a better understanding of how these glaciers respond to external forcings, we examine the evolution of surface elevation, velocity and calving front position, and geometric factors controlling ice dynamics (glacier bed, fjord depth, grounding line locations).


Exploring links between the North Water polynya and the mass balance of Devon Island ice cap


Corresponding author: Laura Edwards

Corresponding author e-mail: loopylol@hotmail.com

Major uncertainties surround future estimates of sea-level rise attributable to mass loss from the polar ice sheets and ice caps. Understanding changes across the Arctic is vital as major potential contributors to sea level, the Greenland ice sheet and the ice caps and glaciers of the Canadian Arctic Archipelago, have experienced dramatic changes in recent times. Most ice mass loss is currently focused around the periphery of these regions where land ice meets the ocean where ice acceleration, thinning and increased calving have been observed. Polynyas are areas of open water within sea ice, which remain unfrozen for much of the year. They vary significantly in size (~3 to > ~85 000 km2 in the Arctic), recurrence rates and duration. Despite their relatively small size, polynyas strongly impact regional oceanography and play a vital role in heat and moisture exchange between the polar oceans and atmosphere. As a result they have the potential to influence air masses reaching nearby glaciers and ice caps creating a maritime climate which could impact their accumulation and surface melt and hence their thickness and mass balance. This work analyses changes in sea-ice extent and concentration in the region of the North Water polynya using passive microwave satellite data and annual mass-balance estimates from the neighbouring Devon Island ice cap, Canadian Arctic (obtained from the University of Alberta), for the period 1980–2006. The results of these analyses will be presented.


Relations between rain falls and sliding velocity of a temperate glacier on its bedrock: Argentiere glacier, Mont Blanc massif, France


Corresponding author: Luc Moreau

Corresponding author e-mail: moreauluc@club-internet.fr

We suggest here highlighting the narrow link between hydrology and sliding velocity of a temperate glacier. According to the intensity of the precipitation measured near Argentiere glacier at 2350 m a.s.l. (Mont Blanc massif, Alps, France), the moment in the melt season, the state of the ice rock interface and suglacial cavities, the mass can be disturbed and the sliding accelerate. It is important to know this phenomenon because it is often the source of the under glacial hydrography change, and this situation already passed in the past. And when the torrent is catch for hydroelectricity, it is important to anticipate on these change of under glacial route when we can. We also bring some figures of the response time between the precipitation and the acceleration!


Surface elevation change by GPS in North Greenland, 2007–2012

Christine S. HVIDBERG, Lars B. LARSEN, Dorthe DAHL-JENSEN, Susanne L. BUCHARDT

Corresponding author: Christine S. Hvidberg

Corresponding author e-mail: ch@gfy.ku.dk

The interior of the Greenland ice sheet have been stable since 2000, but observations by GRACE have shown that the areas losing mass have been spreading north along the west coast from the south, and since 2007 affected the interior of North Greenland. It is not known how these changes affect the interior flow pattern at the divide. We present results from a surface GPS survey near the NEEM drill site (77.45°N, 51.06°W) in North Greenland, covering 6 years of observations from 2007 to 2012. The NEEM drill site is located at the main ice divide in North Greenland, between Petermann Glacier in the north and Baffin Bay Glaciers in the northwest. A strain net was established around the NEEM site in 2007 and resurveyed with GPS each year until 2012. The strain net at NEEM consists of a reference pole and 12 poles placed in three diamonds at distances of 2.5, 7.5 and 25 km, respectively, from the reference pole. Additional poles are located at the ridge, ~50 km upstream from NEEM. Processing and analysis of the GPS data show that the ice flow along the ice divide is W-NW with an average horizontal surface velocity at NEEM of 5.83 ± 0.3 m a–1 along the ridge. Preliminary analysis shows surface strain rates at NEEM to be (-0.2 ± 0.6)·10–5 a–1 (longitudinal) and (11.5 ± 0.3)·10–5 a–1 (transverse), i.e. flow is divergent and approximately constant along the ridge. Measurements of surface height at all the poles provide consistent observations of the mean rate of surface elevation changes over the strain net. The mean rate of surface elevation change has varied from year to year, with a lowering of the surface between 2007 and 2009, followed by a slightly rising surface from 2009 to 2012, resulting in an overall slightly lowering surface from 2007 to 2012 with a rate of –0.03 ± 0.07 m a–1. These data provide validation data for satellite observations of elevation change and ice velocity. The spatial pattern of the surface elevation change was also investigated and compared with the surface mass balance to discuss trends across the divide and a possible migration of the divide position. We compare our results with satellite observations of surface elevation change. We also set up a 2-D ice-sheet flow model to assess how sensitive the ice divide position is to changes at the margin, and discuss the mass balance and ice dynamics of the ice sheet in this region.


Present-day and future projections of the surface mass balance of the Greenland ice sheet using the EC-Earth and HIRHAM5 global and regional climate models

Ruth MOTTRAM, Guðfinna AÐALGEIRSDÓTTIR, Fredrik BOBERG, Jens Hesselberg CHRISTENSEN, Ole Bøssing CHRISTENSEN, Peter LANGEN, Marianne Sloth MADSEN, Christian RODEHACKE, Synne HØYER SVENDSEN, Martin STENDEL, Shuting YANG

Corresponding author: Ruth Mottram

Corresponding author e-mail: rum@dmi.dk

We present new results from the EC-Earth and HIRHAM5 climate models which we use to calculate the present-day surface mass balance (SMB) of the Greenland ice sheet and to make projections of SMB over the coming century. The HIRHAM5 regional climate model (RCM) has a new sophisticated snow scheme, including important processes such as refreezing in the snowpack and an enhanced albedo parameterization, in order to calculate SMB directly within the model. The simulations were run at very high horizontal resolution (0.05° or ~5 km), using the ERA-Interim reanalysis as forcing for the present day (from 1979 to 2013) and the EC-Earth global climate model (GCM) running the RCP4.5 and 8.5 scenarios for forcings at the boundaries for three time slices over the 20th and 21st centuries (1990–2010, 2040–2060, 2080–2100). Validation of the present-day model output indicates that HIRHAM5 represents the present-day climate well, particularly the precipitation distribution and melt area extent when compared with satellite observations and ice-core records. This is also confirmed by comparison with observations of air temperature, snowpack temperature and longwave and shortwave radiative fluxes from automatic weather stations (AWS) in the PROMICE and GC-Net networks on the ice sheet, as well as the coastal DMI weather stations around Greenland. The output from runs driven by the EC-Earth climate model on the other hand shows a cold bias in Greenland at the present day, particularly in the RCP4.5 scenario simulation, which has a corresponding effect on SMB over the course of the 21st century.


Impact of observationally constrained Earth rheological uncertainty on Eurasian ice sheet evolution


Corresponding author: Lev Tarasov

Corresponding author e-mail: lev@mun.ca

We examine the sensitivity of the evolution of the Eurasian ice sheet complex over the last glacial cycle to uncertainties in the viscous structure of the Earth. The analysis is with a sample of 3-D MUN glacial systems model (GSM) runs from the calibrated GLAC-1c distribution. The observational constraints for the Bayesian calibration of the GSM included relative sea-level data, present-day rates of uplift, and fits to a geologically inferred deglacial margin chronology. The GSM incoporates a fully coupled visco-elastic Earth response module (onion type model with spherically symmetric Earth rheology and density structure). The distribution of Earth viscosity models derives from analysis of GSM decay-time fits to the central Fennoscandia Angermanalven relative sea- level record. Given the over-riding uncertainty in glacial cycle climate forcing, we consider the question: do uncertainties in Earth rheology need consideration for glacial cycle modelling? Across our 95% confidence interval for Earth rheology, the ice volume response ranges up to 20% during glacial stadials.


Seismicity from hydraulic fractures in glaciers and ice sheets


Corresponding author: Brad Lipovsky

Corresponding author e-mail: lipovsky@stanford.edu

Improving estimates of future sea-level rise will require an improved understanding of the basal ice dynamics of the ice sheets. Because the presence of liquid water greatly affects the dynamics of these bodies of ice, a passive seismic method for determining the spatial and temporal distribution of liquid water within ice bodies is highly desirable. With these goals in mind, we investigate the role of fluid-filled fractures in creating seismicity observed in glaciers and ice sheets. We model the resonant modes of a hydraulic fracture, and demonstrate how this model can be used to estimate the length and aperture of hydraulic fractures from passive seismic data. We imagine a thin millimeter-wide water-filled fracture in ice. The system, initially at rest, is perturbed and responds through excitation of normal modes of the hydraulic fracture. Perturbations to the conduit could result from the unsteady advance of a crevasse tip or fracture events in nearby ice. We present a linearized analysis that accounts for quasi-static elasticity of the fracture wall, as well as laminar viscous fluid drag, inertia, and compressibility. We limit attention to slow waves (phase velocity less than the sound wave speed), symmetric fracture perturbations, and thin fractures (fracture width much smaller than wavelengths of interest). Two frequencies characterize the coupled fluid-elastic system. Low-frequency oscillations below a characteristic viscous frequency are highly attenuated. High-frequency oscillations above a characteristic elastic frequency experience small motions of the fracture wall and therefore little seismic energy. For a 1 mm water-filled fracture in ice, the viscous and elastic frequencies are ~1 Hz and ~100 kHz. This wide range encompasses many passive seismic observations. In this frequency range there exists a dispersive guided wave known as a crack wave that exists due to restoring forces from elastic wall deformation. We estimate hydraulic fracture length and width for harmonic seismic events reported in the literature, specifically, Bakaninbreen, Bering Glacier, Ice Stream C and Ice Stream E. Fractures have widths on the order of several millimeters. Fracture lengths vary from several meters in Bakaninbreen to several tens of meters in Ice Stream E. These estimates of conduit aperture require fracture water pressure on the order of several kPa. Because we do not account for all sources of attenuation, these estimates are taken to be lower bounds.


Branch, shift or step? Transitions in planform development of distributed sheet-like water systems

Timothy CREYTS

Corresponding author: Timothy Creyts

Corresponding author e-mail: tcreyts@ldeo.columbia.edu

Subglacial drainage is critical to basal lubrication and slip of glaciers and ice sheets over their beds. Whether basal water is distributed over broad areas of the bed or coalesces into a few main channels determines the degree and extent of sliding. Subglacial water that is deep relative to bed roughness with large distributed extents can increase sliding. Shallow, coalescent water systems can aid coupling of ice and bed and lower sliding rates. Switches between any two states in space or time can change sliding rates dramatically. These can correspond to drainage that changes in stepwise fashion with increasing water depth. As a result, the spatial and temporal adjustment of subglacial hydraulic systems can be an important control on ice discharge. Water evacuation is governed by gradients driving water flow and hydrologic transmissivity. Transmissivity is governed by effective pressure and basal melt rates. Effective pressure controls closure rates of the overlying ice into the water system. Melt rates, however, act in the opposite sense and open the subglacial water system. These functions are not necessarily monotonic nor is it necessary that they are smooth. Such multivalued transmissivity functions result from the rates of water system closure and melt. We include a broad and an along-path water balance that shows how these multivalued functions can lead to variations in melt rate. Distributed and channelized subglacial networks with broad patterns in the basal drainage network are shown to have time-varying evolution of subglacial effective pressure. Switches between different branches of the water depth relationship correspond to either the establishment or shut-down of a ‘connected’ drainage system and yield either more or less hydraulically conductive areas of the subglacial system. We show how along-path water discharge affects water depth and results in spatial switches from one drainage state to another. We conclude by discussing state behavior relative to subglacial conditions beneath present ice sheets and ice streams and relate this to ice discharge from sliding.


The fast full-Stokes method (FFS) applied to the Greenland ice sheet


Corresponding author: Josefin Ahlkrona

Corresponding author e-mail: josefin.ahlkrona@it.uu.se

In order to efficiently do numerical modeling of ice-sheet/ice-stream/ice-shelf systems, we develop a method to automatically and dynamically combine different approximation levels of the equations governing ice dynamics, within the same ice sheet. The method is called FFS (fast full Stokes) and is implemented in the widely used code Elmer/Ice (http://elmerice.elmerfem.org). Solving the exact equations describing ice flow (the full-Stokes equations) is computationally demanding and for many applications, such as paleoglacial simulations and uncertainty quantification, it is not possible today. Therefore, a very common approach is to apply approximations such as the shallow-ice approximation (SIA). However, while the SIA is fairly accurate in large parts of the interior of an ice sheet it fails to properly model marginal ice-sheet dynamics, ice streams, and coupling to ice shelves. We use the SIA in areas where it is sufficiently accurate, and solve the exact equations elsewhere. Determining in which areas the SIA is appropriate beforehand is hard, and we therefore implement an error estimation to automatically assign the areas in which it can be used. By this method, the areas where the SIA is applied are allowed to change with time. We apply the FFS to the Greenland ice sheet and compare with solving the full-Stokes equations over the whole ice sheet, with regard to accuracy and simulation time.


Iceberg capsize hydrodynamics and the source of glacial earthquakes


Corresponding author: Mac Cathles

Corresponding author e-mail: mcathles@umich.edu

Warming in the recent past has led to an increase in large ice mass loss events from the Greenland and Antarctic ice sheets such as the catastrophic collapse of ice shelves on the Antarctic Peninsula, and the calving of cubic-kilometer-scale icebergs in Greenland’s outlet glaciers. Both calving styles involve the fracture and then subsequent capsize of unstable icebergs. The latter has been identified as the source of long-period seismic events classified as glacial earthquakes, while a seismic signal emanating from the collapse of an ice shelf has not yet been observed. The ability to partially monitor polar mass loss through the Global Seismographic Network is quite attractive, yet this goal necessitates an accurate model of a source mechanism for glacial earthquakes. A simple relationship between iceberg mass and the measured seismic signal has proved difficult to develop from in situ observations or numerical models and has not yet been found. To address this, we use a laboratory-scale model to measure aspects of the post-fracture calving process not observable in nature. Our results show that a combination of mechanical contact forces (pushing on the terminus) and hydrodynamic pressure forces (pulling on the terminus) are generated by the capsize of an iceberg adjacent to a glacier’s terminus. These forces produce a dipolar strain which is reminiscent of a single couple seismic source. The mass of the iceberg in our physical model affects both the total force on the terminus, the shape of the force history, and the magnitude of the torque on the terminus. This relationship is complicated by a dependence also on the geometry of the calving front – the torque is significantly less for experiments where the depth of iceberg capsizing is less than the water depth. We show the measured force-time histories for experiments with a range of iceberg aspect ratios (equivalent to changing the mass of the capsizing iceberg) and water depths in our laboratory-scale model, and offer an interpretation of how these signals would look when scaled and observed using a seismometer at various distances from the source.


Analysis of different snow dunes on the slopes of Dome A, East Antarctica

Marco SCAIONI, Yixiang TIAN, Xiaohua TONG, Rongxing LI, Bo SUN

Corresponding author: Yixiang Tian

Corresponding author e-mail: tianyixiang@tongji.edu.cn

The analysis of optical (Landsat and MODIS) and radar satellite images (RADARSAT) allowed us to recognize different kinds of snow dunes on the slopes of Dome A in East Antarctica. Investigation of such morphological structures is quite relevant because of its influence on surface mass balance and snow deposition and layering. In addition, very huge dunes like those found in this region may also contribute to ice instability. ‘Megadunes’ have been detected in this region. Here such dunes have similar geometric characters (amplitude and wavelength) with respect to large megadune fields in inner East Antarctica, but their spatial extent is quite limited. In addition, every small groups of dunes do not feature homgeneous amplitude, as can be measured from ICESat laser altimetry data. Usually, the first dune in the uphill direction has a larger amplitude than others, a fact that is not common in other large megadune fields and required in-depth investigations. Another kind of dune can be found in the area roughly delimited by 77°–80°S latitude. The size of this kind of formation is much bigger than ‘regular’ megadunes, because their wavelength extends up to a few tens of kilometres and amplitude spans over about 18–20 m. While megadunes are mainly a surface process resulting in the stratification of ice layers over years, other larger dunes (we termed them ‘gigadunes’) should be deeper phenomena involving the internal structure of the ice sheet as well. The IPR data gathered during the Chinese expeditions on the PANDA traverse from Zhongshan Station to Dome A are exploited for the analysis of this kind of formation, whose existence and properties are not documented in the literature and do not seem to occur in other areas of the polar continent.


A simple mass-balance model to calculate ice volume changes of Icelandic glaciers for glacial isostatic adjustment modelling


Corresponding author: Tómas Jóhannesson

Corresponding author e-mail: tj@vedur.is

Modelling of glacial isostatic adjustment requires time series of ice volume changes and an estimate of the spatial distribution of the associated ice thickness changes to calculate crustal load variations. We describe a mass-balance model based on daily distributed precipitation and temperature time series for Iceland on 1 × 1 km grids. Temperature fields are derived from measurements at meteorological stations using a seasonal variation of the vertical lapse rate to extrapolate the temperature measurements to the elevation ranges of the glaciers. Precipitation is calculated based on a linear theory of orographic precipitation and snow accumulation is calculated from a snow-rain threshold. Ablation of snow and ice is calculated from degree days with separate coefficients for snow and ice. Mass-balance elevation feedback is taken into account based on the ‘reference-surface’ mass balance introduced by Elsberg and others (2001), using the glacier surface geometry from around 2010, thereby avoiding the need to explicitly calculate glacier geometry changes with a dynamic model. The model will be used to provide crustal load changes for GIA modelling of Iceland since 1890 with the aim of deriving a mutually consistent ice volume history and uplift rate field.


Projecting sea level: challenges faced by ice-sheet models


Corresponding author: Sophie Nowicki

Corresponding author e-mail: sophie.nowicki@nasa.gov

Projecting the future evolution of the Greenland and Antarctic ice sheets is a problem of enormous societal importance as ice sheets influence our future sea levels. This crucial issue is however a non-trivial task, as demonstrated by the Sea level Response to Ice Sheet Evolution (SeaRISE) effort: prescribing simple external forcings to a group of ice-sheet models results in a spread in responses. Understanding the source of the diversity in the model results is therefore crucial in order to reduce the uncertainty in the projection. Just as in any future climate simulation, the analysis presented here demonstrates that the model spread in the SeaRISE effort is due to a number of factors. First is the problem of obtaining an initial configuration for the projection. The two commonly used methods, interglacial spin-up or data assimilation, have both advantages and drawbacks, and will affect the determination of fields that cannot be measured (such as basal slipperiness). Second is the uncertainty in actual observations, which includes but is not limited to surface mass balance, basal topography, ice thickness and surface velocities. An additional issue with these observations is that they can be transient quantities which are not measured at the same time, but ice-sheet models require them to be simultaneous. Third is the uncertainty in the models’ physics and discretization, which is limited by our understanding (or lack of understanding) of crucial processes that often occur at subgrid scale relative to the resolution used by continental ice-sheet models, and thus require parameterization. Grounding line migration and sliding laws are such an example. Fourth is the determination of the future forcing scenarios and their implementation as the external forcing. Unfortunately, as demonstrated in this analysis, all ice-sheet models face these limitations to some degree, so that it is extremely difficult to identify a set of models and projections that should be trusted in preference to others. One model might be more suitable for assessing the impact of a warmer atmosphere because of its initialization procedure, but its deficiencies in capturing grounding line migration, for example, might make its projections for oceanic forcing unreliable. More work is thus required to evaluate individual ice-sheet models’ skills in projection, and to reduce the uncertainties in ice-sheet projections.


Big data for advanced ice-sheet modeling


Corresponding author: Eric Rignot

Corresponding author e-mail: erignot@uci.edu

Current uncertainties in projections of sea-level rise from ice sheets and glaciers are embarrassingly large because we are lacking large-scale, high-resolution models coupled with the ocean, sea ice and the atmosphere, and constrained by massive data assimilation to minimize the impact of unresolved physics. Progress is being made in those directions, however, under the impetus of new remote-sensing observations that reveal rapid, significant changes affecting the outlet glaciers and that help constrain ice-sheet models in a new, effective way; furthermore, high-resolution, higher-order physics ice-sheet models are being developed and applied to ice-sheet-wide problems with encouraging results; ice–ocean–sea-ice–atmosphere coupled models are also starting to emerge; and efforts are made on the remote-sensing side to alleviate our most significant knowledge gaps such as bed topography and ice thickness, and sea-floor bathymetry beneath ice shelves, along glacial fjords and on continental shelves, to name a few. While one might easily argue that ice sheets and glaciers will remain fundamentally unpredictable, I will discuss current progress and frameworks that are likely to provide significant advances in projecting ice-sheet and glacier evolution. A most fundamental aspect of this new framework, however, is that the problem is of a multidisciplinary nature, beyond glaciology.


What can we learn about glacier flow by observing and modelling tidally induced motion?

G. Hilmar GUDMUNDSSON, Sebastian H.R. ROSIER, Keith MAKINSON, Matt A. KING

Corresponding author: G. Hilmar Gudmundsson

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

One of the arguably more surprising recent discoveries in glaciology is the observation that ocean tides can affect flow of ice streams tens of kilometres upstream from the grounding line. The impact can be substantial. On Rutford Ice Stream, tides cause about 20% peak-to-peak change in flow velocities. A further intriguing aspect of these observations is the fact that the tidal response is sometimes strongest at tidal frequencies where tidal forcing is negligible. We will give an overview of observations of tidally induced motion on ice streams surrounding the Ronne–Filchner ice shelf, and from the ice shelf itself. A consistent feature seen at all locations is a strong tidal response at the Msf (14.7 days) tidal frequency, even though the Msf tide is absent in the ocean tides in the area. Recent 3-D full-Stokes modelling supports the idea that this kind of tidal response can be generated through a non-linear interaction between tidal stresses and basal motion. Including migrating grounding line in a visco-elastic model increases the size of the signal.


Introduction and analysis of a benchmark dataset for tidewater glacier retreat


Corresponding author: Ethan Z. Welty

Corresponding author e-mail: ethan.welty@colorado.edu

What triggers and regulates the dynamic response of marine-terminating glaciers to climate forcing is poorly understood, representing one of the largest sources of uncertainty in predicting future sea-level rise. Alaska’s Columbia Glacier is a dramatic example of such dynamic instability; since entering into rapid retreat in the early 1980s, calving losses at the 1000 km2 glacier have accounted for upwards of 1–2% of new water globally. Uniquely long and detailed time series of ice surface elevations, velocities and terminus position – assembled from aerial, satellite and time-lapse photography – reach back to steady-state conditions (1976), through the onset of retreat, to the present day. Hoping to further the development of realistic calving laws in ice flow models, we discuss our progress to construct a standardized benchmark dataset of glacier calving from Columbia Glacier’s rich but unwieldy data history, and present a statistical analysis of the relationship between ice flux and terminus geometry over the varied 35 year retreat. Finally, we discuss the frequency distributions of calved iceberg area, calving event volume, and time between consecutive calving events as additional benchmarks of modeled calving statistics.


Monitoring ocean heat variability around the Greenland ice sheet: insights from numerical models


Corresponding author: Ian Fenty

Corresponding author e-mail: Ian.Fenty@jpl.nasa.gov

Ocean warming has been hypothesized as a cause for some of the recent observed Greenland ice sheet mass loss and glacier destabilization, both of which have clear implications for the rates of global sea-level rise. Warming subsurface waters circulating in front of marine-terminating glaciers are expected to increase submarine melting at the ocean–ice interface which may then, in turn, trigger an acceleration, thinning and retreat of the ice stream. While in situ hydrographic observations in the seas to the southwest and southeast of the ice sheet clearly indicate warming over the past two decades, the paucity of ocean measurements near glacier termini, in the narrow fjords, and above the continental shelves over the same time period has limited our ability to effectively test the ocean warming hypothesis. Progress towards relating ocean and Greenland ice sheet variability is challenging because the subsurface ocean temperature field on the continental shelves and within many fjords is likely largely dominated by high-frequency (hourly-daily) and short spatial scale (1–10 km) variations arising from ocean dynamical processes (e.g. mesoscale and sub-mesoscale eddies, tides and convection). Drawing robust inferences about lower-frequency (e.g. seasonal, annual and interannual) subsurface ocean temperature trends from sparse measurements of such a field requires well-constrained prior estimates of the field’s underlying variance. In this talk, we present recent work towards quantifying the variability of the subsurface temperature field above Greenland’s continental shelves near several important glacial fjords using a set of five numerical ocean models of increasing horizontal resolution (30 to ~1 km). Analysis of the simulated ocean variability field reveals the existence of a set of special regions on the shelves that exhibit relatively little high-frequency variability and capture lower-frequency warming signals. We argue that in situ sampling in these regions may prove useful for the long-term monitoring of the waters in contact with the Greenland ice sheet.


The curious distribution of a pervasive depositional marker at ~17.5 ka in West Antarctica


Corresponding author: Robert Jacobel

Corresponding author e-mail: jacobel@stolaf.edu

We report results from radar surveys acquired in the trunk of Kamb Ice Stream and on Siple Dome that depict an as yet unexplained spatial distribution of reflections from an internal radar horizon seen widely across the West Antarctica ice sheet (WAIS). Previous radar studies have detected this layer over an area in excess of 250 000 km2 throughout the central WAIS. Results from the Byrd ice core link the radar reflection to excess acid deposited from a volcanic eruption approximately 17.5 ka ago, and the same chemical signature has recently been observed at nearly the same time in a tephra layer in the WAIS Divide core. Observations reported here extend evidence for this horizon into the lower trunk of Kamb Ice Stream and suggest that it is likely present in areas downstream into the Ross Ice Shelf. However, the layer is not detected in ice flowing from Siple Dome into Kamb Ice Stream, nor does it appear to be present in the Siple Dome ice core. This suggests a peculiar pattern in its spatial distribution and places constraints on the source of the eruption. Here we discuss hypotheses about its source location, the mechanisms of its distribution, and implications for ice-sheet mass balance.


Modelling glacier retreat after ice-shelf collapse


Corresponding author: Jan De Rydt

Corresponding author e-mail: janryd69@bas.ac.uk

Satellite measurements have shown the consistent and ongoing speed-up and retreat of glaciers that were once buttressed by the collapsed Larsen B ice shelf. Understanding the response of grounded ice to ice-shelf collapse is a prerequisite to future predictions of sea-level rise, as other ice shelves such as Scar Inlet or the Larsen Ice Shelf are weakening due to changing atmospheric and ocean conditions. We present preliminary model results for a number of sensitivity experiments that aim to simulate the response of glaciers to the collapse of Larsen B. For this purpose we use Ua, a state-of-the-art shallow shelf model with grounding line resolving capabilities. The model is initialized to observed pre-2002 conditions with the ice shelf in place, and transient runs are done that study the response to a weakening and removal of the ice shelf. Model results are compared with a novel dataset of ice velocities, which provides the most comprehensive overview of dynamical changes after the collapse. In addition, we investigate the glacier response to the collapse of Scar Inlet, a remnant of the Larsen B ice shelf, which has been suggested to have weakened in recent years.


Subglacial water flow and sediment interaction beneath ice streams


Corresponding author: Teresa Kyrke-Smith

Corresponding author e-mail: teresa.kyrke-smith@earth.ox.ac.uk

The nature of the subglacial water flow beneath glaciers and ice sheets is a long-standing problem. Various types of flow have been proposed, ranging from film flow to Röthlisberger channels, linked cavity flow, canals, and patchy films. We consider the coupled behaviour of subglacial water flow over a saturated, deformable till, which moves by till shear, till squeeze and bedload transport. The water layer is fed by subglacial melting and is initially presumed to take the form of a Weertman–Creyts–Schoof rough-bedded film, in which the ice is supported by larger clasts, but there is a millimetric water film which submerges the smaller particles. We present a model of the coupled flows of ice, water and sediment, and we show that in certain cases the model predicts the formation of shallow streams or canals, with a typical depth of the order of centimetres. These canals are stable features once formed.


The first remotely sensed full-depth measurements of ice-sheet flow


Corresponding author: Jonathan KINGSLAKE

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

We describe the first geophysical technique capable of remotely measuring the vertical velocity of ice through to the beds of ice sheets. We validate the radar technique at several ice divides where theory predicts strong spatial gradients in ice velocity, showing that the observed velocity patterns agree with the predictions of current ice flow models. The data are also consistent with the geometry of internal isochronous layers, which depends on the ice flow history and is mapped using a different radar technique. Vertical velocities measured using our new technique can be used to improve ice rheology parameterizations and to calculate the mean horizontal velocities required for mass flux calculations. This work demonstrates the feasibility of large-scale mapping of three-dimensional velocity fields which could substantially improve ice-sheet mass-balance calculations.


The POLENET-ANET integrated GPS and seismology approach to understanding glacial isostatic adjustment and ice mass change in Antarctica

Terry WILSON, Michael BEVIS, Stephanie KONFAL, Richard ASTER, Julien CHAPUT, David HEESZEL, Douglas WIENS, Sridhar ANANDAKRISHNAN, Ian DALZIEL, Audrey HUERTA, Eric KENDRICK

Corresponding author: Terry Wilson

Corresponding author e-mail: wilson.43@osu.edu

The POLENET-ANET project is simultaneously resolving crustal motions, measured by GPS, and Earth structure and rheological properties, mapped by seismology. Measured vertical and horizontal crustal motion patterns are not explained by extant glacial isostatic adjustment (GIA) models. These models have ice histories dominated by ice loss following the Last Glacial Maximum (LGM) and rely on 1-D Earth models, with rheological properties varying only radially. Seismological results from POLENET-ANET are revealing significant complexity in lateral variation in Earth properties. For example, crustal thickness variations occur not only across the East-West Antarctic boundary, but also between crustal blocks within West Antarctica. Modeling of mantle viscosity based on shear wave velocities shows a sharp lateral gradient from high to low viscosity in the Ross Embayment, a much more gradual gradient in the Weddell Embayment, and very low viscosities below Marie Byrd Land and the Amundsen Sea Embayment (ASE). Remarkable vertical and horizontal bedrock crustal motion velocity magnitudes, directions and patterns correlate spatially, in many aspects, with Earth property variations mapped by seismology. Within the ASE, extremely high upward velocities are flanked by subsiding regions – neither predicted by GIA models. Given the thin crust and low mantle viscosity, it is likely that this is not an LGM signal, which would have already relaxed, and uplift due to the elastic response to modern ice mass change clearly is important. As in other regions where rapid GIA-induced uplift has been measured, the crustal velocities in the Amundsen Embayment may also record a viscoelastic response to ice loss on decadal–centennial timescales. Along the East-West Antarctic boundary in the Ross Embayment, GIA-induced horizontal crustal motions are toward rather than away from the principal ice load center, correlating spatially with the strong lateral gradient in mantle viscosity. In the Weddell Embayment region, where crustal thickness is intermediate between East and West Antarctica and mantle viscosity values are moderate, crustal motions show the best match with predictions of GIA models. It is clear that lateral variations in Earth properties fundamentally control the isostatic response to ice mass changes in Antarctica. Ongoing integrated seismic-GPS studies are critical to developing the next generation of GIA models.