Mass Changes of Outlet Glaciers Along the Nordensjköld Coast, Northern

Home , Cape Longing, Edgeworth Glacier, Prince Gustav Channel

PUBLICATIONS

Geophysical Research Letters

RESEARCH LETTER Mass changes of outlet glaciers along the Nordensjköld 10.1002/2014GL061613 Coast, northern Antarctic Peninsula, based Key Points: on TanDEM-X satellite measurements • Volume change 2011–2013 of Antarctic Peninsula glaciers based Helmut Rott1,2, Dana Floricioiu3, Jan Wuite1, Stefan Scheiblauer1, Thomas Nagler1, and Michael Kern4 on new technique • Downwasting of most outlet glaciers 1ENVEO IT GmbH, Innsbruck, Austria, 2Institute for Meteorology and Geophysics, University of Innsbruck, Innsbruck, Austria, ongoing 18 years afterLarsen-Acollapse 3 4 • Trend of decrease in ice mass losses Institute for Remote Sensing Technology, German Aerospace Center, Oberpfaffenhofen, Germany, ESA-ESTEC, due to deceleration of glacier ﬂow Noordwijk, Netherlands.

Supporting Information: Abstract We analyzed volume change and mass balance of outlet glaciers of the northern Antarctic • Readme • Text S1 Peninsula over the period 2011 to 2013, using topographic data of high vertical accuracy and great spatial detail, acquired by bistatic radar interferometry of the TanDEM-X/TerraSAR-X satellite formation. The study Correspondence to: area includes glaciers draining into the Larsen-A, Larsen Inlet, and Prince-Gustav-Channel embayments. After H. Rott, collapse of buttressing ice shelves in 1995 the glaciers became tidewater calving glaciers and accelerated, [email protected] resulting in increased ice export. Downwasting of most glaciers is going on, but at reduced rates compared to À previous years in accordance with deceleration of ice ﬂow. The rate of mass depletion is 4.2 ± 0.4 Gt a 1, with Citation: À1 the largest contribution by Drygalski Glacier amounting to 2.2 ± 0.2 Gt a. On the technological side, the Rott,H.,D.Floricioiu,J.Wuite,S.Scheiblauer, T. Nagler, and M. Kern (2014), Mass investigations demonstrate the capability of satellite-borne single-pass radar interferometry as a new tool for changes of outlet glaciers along the accurate and detailed monitoring of glacier volume change. Nordensjköld Coast, northern Antarctic Peninsula, based on TanDEM-X satellite measurements, Geophys. Res. Lett., 41, 8123–8129, doi:10.1002/2014GL061613. 1. Introduction

Received 20 AUG 2014 Disintegration events of the northern sections of Larsen Ice Shelf on the Antarctic Peninsula (API) in 1995 and Accepted 26 OCT 2014 2002 triggered near-immediate acceleration of the outlet glaciers previously feeding the ice shelves, Accepted article online 29 OCT 2014 resulting in major mass losses due to increased ice discharge [Rott et al., 2002; Scambos et al., 2004, 2011]. Published online 21 NOV 2014 À Sasgen et al. [2013] report a mass balance of À26 ± 3 Gt a 1 for the northern API for January 2003 to September 2012, derived from measurements of the Gravity Recovery and Climate Experiment. Shepherd et al. [2012] present mass balance estimates for the Antarctic Peninsula Ice Sheet based on satellite altimetry, gravimetry, and the input/output method. They report a reconciled mass balance estimate for À API of À36 ± 10 Gt a 1 for the period 2005 to 2010. Precise data on volume changes and their temporal trends are essential for assessing the response of the glaciers to changing boundary conditions, identifying processes controlling ice ﬂux, and estimating their current and future contributions to sea level rise [Barrand et al., 2013]. We are interested in spatially detailed observations of volume change and mass balance of glaciers north of the Seal Nunataks where Larsen-A Ice Shelf disintegrated in January 1995 [Rott et al., 1996] (Figure 1). In contrast to Larsen-B Ice Shelf and its tributary glaciers, little attention has been paid to this area, in spite of the fact that the ice shelf disintegrated already 7 years earlier. A detailed analysis of volume change and mass balance of glaciers on the northern Antarctic Peninsula has been performed by Scambos et al. [2014] using a combination of digital elevation model (DEM) differencing and repeat-track laser altimetry from the Ice, Cloud, and Land Elevation Satellite (ICESat). The DEMs difference pairs, based on stereo images of optical satellite sensors, span 2001–2006, 2003–2008, and 2004–2010 for different sections of the study area, and are integrated with ICESat data spanning September 2003 to March 2008. This is an open access article under the We are focusing at API outlet glaciers along the Nordenskjöld Coast which extends along the east coast of the terms of the Creative Commons Attribution-NonCommercial-NoDerivs API between Cape Longing (Figure 1) and Cape Fairweather, located about 40 km south of the Drygalski License, which permits use and distri- Glacier front. In addition, our analysis includes Sjögren-Boydell glaciers, the main API outlet glaciers to bution in any medium, provided the Prince-Gustav-Channel. Surface elevation change is mapped over a time span of about 2 years, extending original work is properly cited, the use is non-commercial and no modiﬁcations from southern winter 2011 to winter 2013. We use topographic data acquired by a new technique, bistatic or adaptations are made. radar interferometry of the TanDEM-X/TerraSAR-X satellite formation [Krieger et al., 2013].

2. Data Base and Methods We apply DEM differencing for retrieving changes in glacier volume and estimating glacier mass balance [Cogley, 2009], using precise, spatially detailed data of surface topography measured by the TanDEM-X (TerraSAR-X add-on for Digital Elevation Measurements) mission. For converting volume change into mass change we use an average density of À 900 kg m 3 (section 2.2 in Text S1 in the supporting information). TanDEM-X employs a bistatic interferometric configuration of the two satellites TerraSAR-X (TSX) and TanDEM-X (TDX) flying in close formation [Krieger et al., 2013]. The satellites form together a single-pass synthetic aperture radar (SAR) interferometer, enabling the acquisition of highly accurate cross- track interferograms that are not affected by temporal decorrelation and variations in atmospheric delay. The primary objective of the mission is the generation of a global high-resolution DEM [Krieger et al., 2013; Rizzoli et al., 2012]. Our analysis of elevation change is based on DEMs derived from TSX-TDX interferograms acquired in 2011 and 2013, separated by a time span of about 2 years. We generated DEMs with 6 m × 6 m grid size from SAR data with a spatial resolution of about 3 m along track and 1.2 m in radar line of À Figure 1. Rate of glacier surface elevation change dh/dt (m a 1) 2011 to sight [Rossi et al., 2012]. The study area 2013 on outlet glaciers along the Nordenskjöld Coast, Antarctic Peninsula. is covered by four separate DEMs, each 1 to 8: basin number. Landmarks: PGC–Prince-Gustav Channel; CL–Cape extending over an area of about Longing; LI–Larsen Inlet; L-A–Larsen-A embayment; CW–Cape Worsley; 30 km × 50 km (Table S1 and Figures S1 – SN Seal Nunatak ice shelf. Background: Landsat image of 31 December and S2 in the supporting information). 2001. Yellow line: coastline November 1995. Red line: coastline on 12 January 2012. Straight white lines: transects on Drygalski (D) and Edgeworth We produced maps of glacier surface (E) glaciers (shown in Figure 3). elevation change (dh/dt)by differencing the DEMs of 2013 with the corresponding 2011 DEMs (Figure 1). For analysis of elevation change, accurate relative vertical registration is crucial. This has been achieved by fine adjusting each of the four 2013 DEMs vertically to the corresponding 2011 DEM at sea level, compensating for the difference in tide level. The accuracy of the vertical registration has been checked at outcrops (nunataks) within the glacier basins (section 2 in Text S1 in the supporting information). A critical issue for producing DEMs by means of across-track SAR interferometry is the 2π phase ambiguity arising from the periodicity of the phase difference between the two SAR images [Rosen et al., 2000]. A phase

À Figure 2. Rate of glacier surface elevation change dh/dt (in m a 1) 2011 to 2013 versus altitude in 50 m intervals (referring to WGS-84) for basins 2 to 7. Green line: hypsometry of analyzed glacier area in km2.

shift of 2π corresponds to a discrete height difference called the height of ambiguity, Ha. The Ha values of our data set range from 34.2 m to 92.5 m for the different DEMs. 2π phase ambiguities, resulting in elevation shifts

corresponding to Ha, are evident in individual DEMs on some steep slopes, in particular along the escarpment separating the glacier tongues from the plateau of the main API ice divide. Therefore, we constrain the analysis of elevation change to the glacier areas below the escarpment, excluding areas above 1200 m. We checked for phase ambiguities by comparing the TDX DEM repeat pass pairs with each other and with the Advanced Spaceborne Thermal Emission and Reflection Radiometer-based Antarctic Peninsula DEM (API-DEM) of Cook et al. [2012]. Areas affected by phase ambiguities are masked out, resulting for each glacier in a contiguous mask for mapping elevation change. Gaps arising from omission of these areas are filled by means of the API-DEM, using the TDX derived dh/dt values in dependence of altitude (Figure 2). The error analysis, taking into account errors in TDX DEM differencing and extrapolation to missing areas, yields À À uncertainties in elevation change (dh/dt) between ±0.29 m a 1 and ±0.36 m a 1 for the various basins (section 2 in Text S1 in the supporting information). For supporting the discussion on temporal changes of glacier behavior we mapped glacier velocities (Figure S3 in the supporting information). We retrieved flow fields for November 1995 and 1999 from 1 day repeat pass interferometric SAR data of the satellites ERS-1 and ERS-2. For several periods between 2008 and 2013 we derived velocity maps from 11 day repeat pass data of the TSX and TDX missions, applying offset tracking. Specifications for the satellite data, the techniques applied, and the analysis of uncertainty are explained by Rott et al. [2011]. The error analysis yields an uncertainty in velocity of ±5%.

3. Spatially Detailed Pattern of Elevation Change The map of surface elevation change is shown in Figure 1. We separate the study region into eight drainage basins for which volume change (Table 1) and dh/dt in dependence of altitude (Figure 2) are derived. The basin outlines inland of the glacier front are adopted from Cook et al. [2012]. We updated the positions of the coastline based on a Landsat image of 12 January 2012. For basin 1 (Cape Longing Peninsula) the easternmost section is not included, because this area is not covered by the available TDX DEMs. Basin 8 south of Drygalski Glacier includes four small glaciers. Three of these glaciers are now draining directly into the ocean; one glacier drains into the remnant ice shelf around the Seal Nunataks. All basins, except Pyke Glacier (basin 3) are losing mass. The altitude dependence of elevation loss, with the loss rates being highest near the glacier front and diminishing up glacier (Figure 2), reﬂects the pattern of stress perturbation after ice shelf collapse which triggered ﬂow acceleration. The lower terminus of Drygalski Glacier, the largest glacier north of the Seal Nunataks (1008 km2 in area), has the largest rate of elevation À change (À11.51 ± 0.49 m a 1 in the elevation zone 50 m to 150 m) and the greatest loss in total ice mass

Table 1. Name of Glacier Basins, Area Covered by dh/dt Analysis, Mean Rate of Elevation Change dh/dt, Volume Change dV/dt (Above Sea Level), and Mass Change dM/dt Observed Mean dh/dt dV/dt Uncertainty dM/dt À À À À No. Basin Name Area (km2) (m a 1) (km3 a 1) (km3 a 1) (Gt a 1)

1 Cape Longing Peninsula 441.6 À0.379 À0.167 ±0.154 À0.150 2 Sjögren-Boydell (SB) 190.2 À2.131 À0.405 ±0.068 À0.364 3 Pyke Glacier 220.2 +0.282 +0.062 ±0.065 +0.056 4 Dinsmoor-Bombardier 200.2 À2.204 À0.441 ±0.060 À0.523a -Edgeworth (DBE) 5 Sobral Peninsula 186.5 À0.861 À0.161 ±0.056 À0.145 6 Cape Worsley Coast 281.8 À3.155 À0.889 ±0.093 À0.800 7 Drygalski Glacier 604.3 À4.006 À2.421 ±0.194 À2.179 8 Seal Nunatak Coast 153.5 À0.788 À0.121 ±0.032 À0.109 Total 2278.3 À1.994 À4.543 ±0.389 À4.214 aIncludes subaqueous mass loss of the ﬂoating section. The names of basins 1, 6, and 8 are not ofﬁcial place names.

À (Table 1). Average surface lowering on the order of several m a 1 is observed on the lower-elevation zones of Sjögren-Boydell glaciers (SB, basin 2), Dinsmoor-Bombardier-Edgeworth glaciers (DBE, basin 4) and Cape Worsley coast (basin 6) (Figure 2). The volume losses on Cape Longing Peninsula (basin 1) and Sobral Peninsula (basin 5) have been modest (Table 1). These basins are composed of several small glaciers. The mass turnover is smaller than on glaciers originating at the API ice divide because of a strong west-east decrease in net accumulation [Turner et al., 2002; Barrand et al., 2013]. A common feature of the dh/dt altitude curves in the various basins is the convergence of dh/dt toward zero at altitudes of about 1000 m. This agrees with the results of Scambos et al. [2014] who report for the glaciers in this region (basins 21 to 25) a mean volume change close to zero for altitudes >1000 m above sea level, based on ICESat data of the period September 2003 to March 2008. Our analysis of 2011 to 2013 volume change shows that the Pyke glacier basin (basin 3, comprising four glaciers forming a single-calving front) in Larsen Inlet is approximately in balanced state. Whereas the Larsen-A and Prince-Gustav-Channel ice shelves disintegrated in January 100 2013-07-11 1995 [Rott et al., 1996], the ice shelf in 80 2011-07-05 Larsen Inlet disappeared already 60 between 1986 and 1989 [Skvarca, 1993]. 40 Detailed features of surface elevation in Elevation [m] 20 E 2011 and 2013 are highlighted in 0246810Figure 3 for transects across the ice Distance [km] fronts of Drygalski Glacier and 400 Edgeworth Glacier. Both glaciers are out 2013-04-21 350 of balance. The glacier tongues are 2011-06-09 heavily crevassed, but the elevation 300 profiles indicate different stages of 250 recession. On Drygalski Glacier the 200 surface rises 250 m over a distance of 5 km inland of the front. Within this 150 Elevation [m] stretch the surface height fluctuates 100 around the mean elevation profile 50 D at a wavelength of about 150 m perpendicular to transverse crevasses. 0246810Further inland the surface is smoother. Distance [km] The shape of the profiles is similar in 2011 and 2013. These features suggest Figure 3. Profiles of glacier surface elevation (in m above WGS-84 ellip- fi soid) 2011 and 2013 across the glacier front on Edgeworth Glacier (E) that the terminus is still rmly grounded, and Drygalski Glacier (D). The location of the profiles is shown in Figure 1. with exception of a strip along the front.

À Table 2. Velocities in m d 1 at Locations Near the Center of the 2013 Glacier Fronts, Derived From Repeat Pass SAR Data of ERS (1995, 1999) and TerraSAR-X (2008, 2009, 2010, and 2013) Month/Year 11/1995 11/1999 10/2008 11/2009 11/2010 11/2013

Sjögren Glacier 0.44 0.96 - 2.77 2.30 1.81 Pyke Glacier 1.12 1.13 1.10 1.10 1.15 1.21 Edgeworth Glacier 0.63 2.40 2.61 2.38 1.84 1.55 Drygalski Glacier 2.63 7.05 5.74 5.90 5.51 7.40

On Edgeworth Glacier the surface rises only 25 m over the first 5 km. The surface height fluctuations along the transect show two different horizontal scales: (i) approximately 100 m corresponding to the separation of crevasses and (ii) undulations of about 1 km wavelength. The amplitude of these undulations and the fragmentation of the terminus increased from 2011 to 2013, associated with frontal advance. At about 5 km distance from the front the surface slope starts to increase and the surface becomes smoother, indicating À grounded ice. The heavily crevassed area covers about 22 km2 and has a mean dh/dt of À0.9 m a 1. From the fragmentation and the surface height above sea level (only 30 to 60 m), we conclude that this section of the terminus is floating. Assuming freely floating ice, the volume loss of the floating section amounts to 0.160 km3, out of which 0.020 km3 in volume are above the water line.

4. Discussion The analysis of volume change shows that 18 years after ice shelf collapse the glaciers draining into the Prince-Gustav-Channel and Larsen-A embayments are still losing mass due to dynamic thinning. The losses are caused by accelerated ice flow which started soon after ice shelf breakup [Rott et al., 2002, 2008]. The spatial pattern and magnitude of elevation change are governed by glacier geometry, subglacial topography, glacier size, and mass turnover. The surface velocities at the present calving gates of Sjögren Glacier, Edgeworth Glacier, and Drygalski Glacier are still up to 4 times higher than at the same points in November 1995 (Table 2 and Figure S3 in the supporting information), another indication of ongoing mass depletion. On the other hand, the frontal velocity of Pyke Glacier changed little from 1995 to 2013 and the ice front retreated only by 0.4 km. This suggests that the Pyke glacier basin has been close to equilibrium state already since 1995. Apparently, the glaciers of the Pyke basin were not much affected by the disintegration of the ice shelf in front. Main reasons are very likely the bed topography and geometry of the glaciers. The sea floor rises from the deepest point in Larsen Inlet toward the glacier front by several hundred meters [Pudsey et al., 2001], and the glacier surfaces ascend by 500 m over the first 10 km. Unlike for Larsen Inlet, bathymetric data show deep troughs extending into the fjords of SB glaciers and DBE glaciers [Pudsey et al., 2001; Evans et al., 2005]. The glaciers of both drainage basins retreated significantly since 1995 (SB by 12 km, DBE by 7 km), tributary glaciers got separated, and the glaciers are now terminating in narrow sections of the fjords (Figure 1 and Table S3 in the supporting information). On Sjögren Glacier À1 the velocity at the point in the center of the current calving front (Uc) decreased from 2.77 m d in October À1 À1 2009 to 1.81 m d in November 2013. On Edgeworth Glacier Uc decreased from 2.61 m d in October 2008 À to 1.55 m d 1 in November 2013. Deceleration of ice flow and decreasing calving cross sections due to retreat in narrow channels and dynamic thinning are the reasons for gradual decrease of ice export and reduced mass losses in recent years. Among the four basins which are made up by several small glaciers (basins 1, 5, 6, and 8), major mass imbalance is observed for basin 6 (Worsley coast). The glaciers of this basin originate at the main API ice divide which along this stretch is separated only 20 km from the Weddell Sea Coast. The unnamed glacier À northwest of Cape Worsley retreated by 4.8 km since 1995 and shows surface lowering up to 13 m a 1 near À1 the front. In 2009 Uc was 0.95 m d , 4 times higher than the velocity at the same position in 1995. Drygalski Glacier accounts for 50% of the mass loss of the study area, triggered by acceleration of ice flow and frontal retreat. The retreat rate was not uniform across the front. The central and southern sections of the front receded by about 5 km between 1995 and 1999 [Rott et al., 2002] and by another 1.5 km since then. The northern section retreated by about 3 km between 1995 and 1999, remained almost stationary until 2011,

and retreated by another 2 km between November 2011 and January 2013. This indicates complex bedrock topography, with various pinning points, resulting in temporal variations of ice velocity (Table 2). The velocity À at a position near the center of the 2013 front was 1.20 m d 1 in January 1993 (precollapse), increased to À À 2.63 m d 1 in November 1995, 10 months after the Larsen-A collapse, and to 7.05 m d 1 in November 1999. In À À À October 2008, November 2010, and November 2011 the velocities were 5.74 m d 1,5.90md 1 and 5.51 m d 1, respectively. The retreat of the northern section of the front after 2011 and glacier thinning triggered another À phase of flow acceleration, resulting in a velocity of 7.4 m d 1 in November 2013. The difference in the mass À À balance estimates of À2.39 Gt a 1 for 2003–2008 [Scambos et al., 2014] and our estimate of À2.18 ± 0.17 Gt a 1 for 2011–2013 is not significant. In view of the recent increase in calving velocity and the rather large thinning rate on the lower terminus, we conclude that the Drygalski Glacier basin will remain a major source for loss in ice masses of the northern API for many years. À Scambos et al. [2014] report a mass change (dM/dt)ofÀ5.63 Gt a 1 for glacier basins 21 to 25, corresponding À to approximately the same glacier area as our basins 1 to 7 with a dM/dt estimate of À4.10 ± 0.39 Gt a 1 for 2011 to 2013. Their analysis covers the time span September 2001 to January 2006 for SB glaciers and À October 2003 to November 2008 for the other basins. The difference of À1.53 Gt a 1 versus our estimate over the period 2011 to 2013 agrees with the trend of decreasing calving velocities in recent years observed on several glaciers. The main difference relates to the glacier basins north of Worsley coast for which Scambos À À et al. [2014] report a mass change of À2.52 Gt a 1 versus our estimate of À1.13 Gt a 1. The total mass change 2011 to 2013 of the analyzed glacier areas, including the subaqueous volume of À Edgeworth Glacier, is À4.21 ± 0.37 Gt a 1 (Table 1). If the mean condition dh/dt ≈ 0 at altitudes >1000 m observed by ICESat [Scambos et al., 2014] is still valid, the contribution of basins 1 to 8 to sea level rise À amounts to 4.05 ± 0.35 Gt a 1. The upper reaches of the glaciers on the central ice plateau, separated from the terminus sections by steep ice and rock cliffs, have not responded dynamically to the signal of flow acceleration initiated at the glacier fronts. Surface mass balance fields generated by the regional atmospheric climate model RACMO version 2.3 [Lenaerts et al., 2012; van Wessem et al., 2014] show for the eight basins À2 À1 over the period 2004 to 2008 a mean surface net balance bn = 930 kg m a , and for 2011–2012 À2 À1 bn = 1001 kg m a , (based on unpublished RACMO data, provided by van Wessem et al. [2014]). This is not a significant difference, suggesting that the decrease in mass depletion rates between approximately 2003 and 2008 by Scambos et al. [2014] and our values are mainly caused by reduced dynamic mass loss.

5. Conclusions 18 years after disintegration of the buttressing ice shelves the glaciers draining into the Prince-Gustav-Channel and Larsen-A embayments are still clearly out of balance. Our estimate of the 2011 to 2013 mass balance for the À API outlet glaciers between Seal Nunataks and Prince-Gustav-Channel is À4.21 ± 0.37 Gt a 1, based on analysis of volume change. Subtracting the estimated mass deficit of floating ice, the estimated contribution to sea À level rise amounts to 4.05 ± 0.35 Gt a 1, corresponding to 17% of the Antarctic Peninsula mass depletion rate reported by McMillan et al. [2014] for 2010 to 2013. The spatial pattern of volume change with the rates of surface lowering decreasing up-glacier reflects the main driving mechanism for mass depletion: dynamic thinning due to stress imbalance initiated at the glacier front. The specific rates of elevation change (dh/dt) differ significantly between individual glaciers, depending on glacier geometry, subglacial topography, glacier À size, and history of frontal retreat. The largest loss in ice mass (2.18 ± 0.17 Gt a 1), elevation losses up to À 12 m a 1 near the glacier front, and high ice flow velocities are observed for Drygalski Glacier, indications that the glacier is still far from balanced state. On the other hand, in the northern section of the study area the mass losses of glaciers decreased significantly since 2003À2008, which agrees with deceleration of glacier flow and decrease of calving cross sections due to glacier thinning. Our analysis of glacier volume change applies differencing of topographic data generated by a new technique, single-pass synthetic aperture radar interferometry (InSAR) of the TanDEM-X mission. This technique is not impaired by temporal decorrelation of the interferometric phase, variations in atmospheric propagation conditions, cloudiness, and variable illumination. It delivers precise elevation data at high spatial resolution. Our investigations demonstrate the capability of the TSX/TDX satellite formation for accurate and detailed mapping and monitoring changes in glacier volume. The technique has the potential of greatly reducing the uncertainty in mass balance of glaciers worldwide.

Acknowledgments References The authors would like to thank A. Cook (Swansea University, UK) for providing Barrand, N. E., et al. (2013), Computing the volume response of the Antarctic Peninsula ice sheet to warming scenarios to 2200, J. Glaciol., – outlines of glacier basins, and M. van 59(215), 397 409, doi:10.3189/2013JoG12J139. – Wessem and M. van der Broeke (Utrecht Cogley, J. G. (2009), Geodetic and direct mass-balance measurements: Comparison and joint analysis, Ann. Glaciol., 50,96 100. University, NL) for providing surface Cook, A. J., T. Murray, A. Luckman, D. G. Vaughan, and N. E. Barrand (2012), A new 100-m digital elevation model of the Antarctic Peninsula – mass balance data generated by the derived from ASTER Global DEM: Methods and accuracy assessment, Earth Syst. Sci. Data, 4, 129 142, doi:10.5194/essd-4-129-2012. ﬂ RACMO model. The TanDEM-X data were Evans, J., C. J. Pudsey, C. ÓCofaigh, P. Morris, and E. Domack (2005), Late Quaternary glacial history, ow dynamics and sedimentation along – made available by DLR through project the eastern margin of the Antarctic Peninsula Ice Sheet, Quat. Sci. Rev., 24, 741 774. ﬂ – XTI_GLAC0457. This work was supported Krieger, G., et al. (2013), TanDEM-X: A radar interferometer with two formation ying satellites, Acta Astronaut., 89,83 98, doi:10.1016/ by the European Space Agency, ESA, j.actaastro.2013.03.008. contract 4000105776/12/NL/CBi. Lenaerts, J. T. M., M. R. van den Broeke, W. J. van de Berg, E. Van Meijgaard, and P. Kuipers Munneke (2012), A new, high-resolution surface mass balance map of Antarctica (1979–2010) based on regional atmospheric climate modeling, Geophys. Res. Lett., 39, L04501, Julienne Stroeve thanks two anonymous doi:10.1029/2011GL050713. reviewers for their assistance in McMillan, M., A. Shepherd, A. Sundal, K. Briggs, A. Muir, A. Ridout, A. Hogg, and D. Wingham (2014), Increased ice losses from Antarctica – evaluating this paper. detected by CryoSat-2, Geophys. Res. Lett., 41, 3899 3905, doi:10.1002/2014GL060111. Pudsey, C. J., J. Evans, E. W. Domack, P. Morris, and R. A. Del Valle (2001), Bathymetry and acoustic facies beneath the former Larsen-A and Prince Gustav ice shelves, northwest Weddell Sea, Antarct. Sci., 13, 312–322, doi:10.1017/S095410200100044X. Rizzoli, P., B. Bräutigam, T. Kraus, M. Martone, and G. Krieger (2012), Relative height error analysis of TanDEM-X elevation data, ISPRS J. Photogramm. Remote Sens., 73,30–38. Rosen, P. A., S. Hensley, I. R. Joughin, K. L. Fuk, S. N. Madsen, E. Rodriguez, and R. M. Goldstein (2000), Synthetic aperture radar interferometry, Proc. IEEE, 88(3), 333–385. Rossi,C.,F.RodriguezGonzalez,T.Fritz,N.Yague-Martinez,and M. Eineder (2012), TanDEM-X calibrated Raw DEM generation, ISPRS J. Photogramm. Remote Sens., 73,12–20, doi:10.1016/j.isprsjprs.2012.05.014. Rott, H., P. Skvarca, and T. Nagler (1996), Rapid collapse of northern Larsen Ice Shelf, Antarctica, Science, 271(5250), 788–792. Rott, H., W. Rack, P. Skvarca, and H. De Angelis (2002), Northern Larsen Ice Shelf, Antarctica: Further retreat after collapse, Ann. Glaciol., 34,277–282. Rott, H., M. Eineder, T. Nagler, and D. Floricioiu (2008), New results on dynamic instability of Antarctic Peninsula glaciers detected by TerraSAR-X ice motion analysis, in Proceedings of 7th European Conference on Synthetic Aperture Radar (EUSAR),pp.159–162 , VDE Verlag, Berlin, Germany. Rott, H., F. Müller, T. Nagler, and D. Floricioiu (2011), The imbalance of glaciers after disintegration of Larsen-B ice shelf, Antarctic Peninsula, The Cryosphere, 5(1), 125–134, doi:10.5194/tc-5-125-2011. Sasgen, I., H. Konrad, E. R. Ivins, M. R. van den Broeke, J. L. Bamber, Z. Martinec, and V. Klemann (2013), Antarctic ice-mass balance 2003 to 2012: Regional reanalysis of GRACE satellite gravimetry measurements with improved estimate of glacial-isostatic adjustment based on GPS uplift rates, The Cryosphere, 7, 1499–1512, doi:10.5194/tc-7-1499-2013. Scambos, T., J. A. Bohlander, C. A. Shuman, and P. Skvarca (2004), Glacier acceleration and thinning after ice shelf collapse in the Larsen-B embayment, Antarctica, Geophys. Res. Lett., 31, L18402, doi:10.1029/2004GL020670. Scambos, T. A., E. Berthier, and C. A. Shuman (2011), The triggering of sub-glacial lake drainage during the rapid glacier drawdown: Crane Glacier, Antarctic Peninsula, Ann. Glaciol., 52(59), 74–82, doi:10.3189/172756411799096204. Scambos, T. A., E. Berthier, T. Haran, C. A. Shuman, A. J. Cook, S. R. M. Ligtenberg, and J. Bohlander (2014), Detailed ice loss pattern in the northern Antarctic Peninsula: Widespread decline driven by ice front retreats, The Cryosphere, in press. Shepherd, A., et al. (2012), A reconciled estimate of ice-sheet mass balance, Science, 338(6111), 1183–1189, doi:10.1126/science.1228102. Skvarca, P. (1993), Fast recession of the northern Larsen Ice Shelf monitored by space images, Ann. Glaciol., 17, 317–321. Turner, J., T. A. Lachlan-Cope, G. J. Marshall, E. M. Morris, R. Mulvaney, and W. Winter (2002), Spatial variability of Antarctic Peninsula net surface mass balance, J. Geophys. Res., 107(D13), 4173, doi:10.1029/2001JD000755. van Wessem, J. M., C. H. Reijmer, J. T. M. Lenaerts, W. J. van de Berg, M. R. van den Broeke, and E. van Meijgaard (2014), Updated cloud physics in a regional atmospheric climate model improves the modelled surface energy balance of Antarctica, The Cryosphere, 8,125–135, doi:10.5194/tc-8-125-2014.