Ground penetrating radar system for measuring deep ice in Antarctica using software-defined radio approach

José Uribe, Rodrigo Zamora, Sebastián Pulgar, Jonathan Oberreuter, Andrés Rivera

Corresponding author: José Uribe

Corresponding author e-mail: juribeparada@cecs.cl

This work shows the design of a radar instrument specifically developed for cold ice measurements in polar regions. The instrument is based on a high-performance software-defined radio (SDR) platform, employing sampling frequencies of 800 MHz for digitizing and 1600 MHz for arbitrary signal generation. A flexible multifrequency radar instrument is built on this SDR system, that comprises: 1) a deep ice chirp-pulsed radar working at central frequency of 155 MHz, with 20 MHz of bandwidth, 1 kW of maximum transmitting power and including synthetic aperture radar in post-processing; and 2) a high-resolution shallow-ice frequency-modulated continuous wave radar, which operates from 200–700 MHz and with 100 mW of output power. We used this instrument in December 2017 during a ~460 km over-snow campaign conducted to Subglacial Lake CECs (79°15′ S 87°34′ W), West Antarctica. We measured a maximum ice thickness of ~2700 m using the deep ice radar and ~150 m of snow/firn thickness obtained in high detail by the shallow ice radar. A detailed description of the radar instrument will be presented, together with some results and their comparisons with data previously obtained in the same area with different radar systems. The instrument presented here shows an improvement in signal-to-noise ratio and along-track resolution of bedrock and englacial structures over previous radar measurements, and also the system provides better adaptability for future parameter changes or upgrades.


Review of 55 years of Russian radio-echo sounding investigations in Antarctica

Sergey Popov

Corresponding author: Sergey Popov

Corresponding author e-mail: spopov67@yandex.ru

Russian (former Soviet) investigations in Antarctica began on 13 February 1956 when Mirny Station was founded. The first radio-echo sounding (RES) tests in Antarctica were carried out in January 1964. The first airborne RES survey on the regular net was carried out in February 1968 in Enderby Land. After that positive experience, the same work was repeated on the wide coastal area between 50° E and 100° E in the framework of Operation Amery (1971–74). This was the first large Russian multidisciplinary long-term scientific project, which included many different geophysical methods. After an interruption, regular airborne surveys including RES were recommenced in this area in 1985 and continue until the present day. The same investigations were also carried out in the Pensacola Mountains area (West Antarctica), on the Filchner–Ronne Ice Shelf in 1980–82, and in Coats Land, Dronning Maud Land and Enderby Land in 1985–91. In the same period (1987–90) a wide regional survey of inland East Antarctica was also carried out. Ground-based RES investigations were started on the band of Mirny–Vostok scientific and logistic traverses, mostly in the 1970s and 1980s. They were continued after the discovery of Lake Vostok in 1993. RES investigations were carried out in 1998–2008 to study this unique feature. They resulted in the discovery of a number of lakes. Since 2008, ground-based RES investigations have been carried out in the track of the new logistic traverse route Progress–Vostok. These were completed in 2013. All these ground-based studies were followed by glaciological measurements (AARI) and geodetic observations (TUD, Dresden, Germany). Special Russian GPR investigations were begun in 2012 in collaboration with INGV (Rome, Italy). They were carried out in conjunction with ground-based RES along the logistic traverse Progress–Vostok to study the internal structure of the snow–firn layer. These have been continued until the present time to acquire scientific results and solve the applied tasks. Mostly they are related to investigations of crevasses and sea ice. One of the main results has been the resumption of the snow-runway at Mirny Station. The other is the study of the vast depression in Dålk Glacier (Progress Station area, East Antarctica) and finding a new path to connect the station and airfield broken by the depression. This study was funded by RFBR according to the research project No. 17-55-12003 NNIO_a.


Overview of the low-frequency ice penetrating radar system survey conducted to Subglacial Lake CECs, West Antarctica

Rodrigo Zamora, José Uribe, Sebastián Pulgar, Jonathan Oberreuter, Andrés Rivera

Corresponding author: Rodrigo Zamora

Corresponding author e-mail: rzamora@cecs.cl

In December 2017, Centro de Estudios Científicos (CECs), conducted a survey at the Subglacial Lake CECs area, West Antarctica (79°15′S 87°34′W). The over snow survey used a deep-penetrating low-frequency radar system developed by CECs that was mounted on sledges pulled by snowmobiles allowing measuring near 260 km of the central West Antarctic plateau. The system is an impulse radar working at a central transmission frequency of 2 MHz having three main components: a 4 kV power transmitter that works at 1 kHz pulse repetition frequency (PRF); a digital acquisition system ADQ214, with an analog to digital converter working at 400 MHz of sample rate, with a resolution of 14 bits and 256 traces on average; and a dual-frequency GPS receiver used for georeferencing the whole system. Transmitter and receiver antennae were resistively loaded wire dipoles. The components were installed inside peli-cases, each one having its own power and solar-panel charging system, including batteries allowing 8 hours of a continuous survey. Each component was installed on sledges tied together forming a 140 m long convoy. The convoy was pulled by snowmobiles moving at 8–10 km h–1 surface velocity. Ice thickness up to ~2700 m was reached. The upper hundreds of meters of the internal structure of the ice were mapped in detail, permitting the identification of isochronous layers with a vertical resolution of ~40 m. A comparison with previous surveys (2010 and 2016) offers major improvements in performance and reliability.


The origin of ice-shelf channels at Institute and Möller ice stream grounding zones, West Antarctica

Martin Siegert, Hafeez Jeofry, Jilu Li, Prasad Gogineni, Neil Ross

Corresponding author: Martin Siegert

Corresponding author e-mail: m.siegert@imperial.ac.uk

Channels carved upwards into the underside of ice shelves have been shown to emanate from grounding lines, and originate from a well organized subglacial hydrological system allowing water to exit the ice sheet at a point source. Such a system has been shown to be associated with both soft- and hard-bedded landforms, which act to route basal water. Both the Möller and Institute ice streams in the Weddell Sea sector of West Antarctica are associated with ice-shelf channels. Here, using radio-echo sounding data, we investigate how these channels are formed by upstream subglacial conditions. For the Institute ice stream, we find that the hard-bedded landform explanation holds, where basal water is channelled beside the landform and, when mixed with cavity water, flows upwards into the corrugation developed by upstream ice flow around it. For the Möller ice stream, however, we have discovered a new mechanism for ice-shelf channelling. Subglacial water flows in a well organized manner along the base of a deep but otherwise smooth basal trough. This trough also moulds the ice-sheet base, such that at the point of flotation it is characterized by a notable downward-facing mound. When the water exits the ice sheet, after mixing with ice-shelf water, it flows upwards beside the mound and etches an ice-shelf channel offset laterally from the axis of the upstream basal tough. These results demonstrate that, while ice-shelf channels are associated with rough beds at the grounding line, there are at least three ways in which they are generated.


Comparing numerical ice-sheet model results with geophysical measurements in the Weddell Sea sector of West Antarctica

Martin Siegert, Hafeez Jeofry, Neil Ross, Jilu Li, Prasad Gogineni, Antony Payne, Steph Cornford

Corresponding author: Martin Siegert

Corresponding author e-mail: m.siegert@imperial.ac.uk

Numerical ice-sheet models commonly match surface ice velocities with InSAR measurements by reducing basal drag. While the resulting ice-sheet solution mimics the flow and form of the actual ice sheet, few attempts have been made to examine whether there is geophysical evidence in support of reduced basal drag where the model requires it. As such, geophysical validation of ice-sheet models has yet to be undertaken. Here, we examine radio-echo sounding data from the Weddell Sea Sector of West Antarctica to investigate how a results from a well established ice-sheet model compare with measurements of the basal environment. The Weddell Sector contains the Institute, Möller and Foundation ice streams, each having distinct characteristics: the trunk of Institute Ice Stream lies over a wide region of wet unconsolidated sediments, where basal drag is likely very low; the Möller Ice Stream lies over a relatively rough bed, where basal drag is likely larger; and the flow Foundation Ice Stream is heavily controlled by deeply incised channels. We find the ice-sheet model deals with each ice stream system well, matching its known geophysical characteristics with reasonable estimates of basal drag. However, we also find that ice velocities do not match perfectly in some locations, and that adjustment of the boundaries of low basal drag, in line with geophysical evidence, should lead to improved model performance.


AWI’s airborne ultra-wideband radar for sounding and imaging of ice sheets and shelves

Olaf Eisen, Tobias Binder, Nils Dörr, Steven Franke, Veit Helm, Angelika Humbert, Daniela Jansen, John Paden, Daniel Steinhage, Stephen Yan

Corresponding author: Olaf Eisen

Corresponding author e-mail: oeisen@awi.de

Over the last years AWI implemented a new ultra-wideband radar, developed by the Center for Remote Sensing of Ice Sheets (CReSIS), on AWI’s polar airplane Polar 6, a Basler BT-67. After terminating the overall system tests and calibration/validation survey in 2015, the system has been in operational use in Greenland and Antarctica for several field seasons. We will present an overview of the results and experiences obtained over the last 2 years to illustrate the system performance in terms of achievable specifications for imaging and sounding ice sheets, and will discuss the requirements and opportunities for logistic deployment in Greenland as well as Antarctica.


Basal roughness of the East Antarctic Ice Sheet and indications for the basal thermal state

Anna Winter, Daniel Steinhage, Franz-Fabian Bellot, Eythor Gudlaugsson, Thomas Kleiner, Angelika Humbert, Olaf Eisen

Corresponding author: Olaf Eisen

Corresponding author e-mail: oeisen@awi.de

Basal motion of ice sheets represents a large uncertainty in ice-flow models, as it dependends on the roughness and material of the subglacial bed and the occurrence of water. It is the part of the total flow speed that can change most rapidly and can therefore facilitate rapid variations in the dynamic behaviour of ice sheets. In this study we investigate the subglacial properties of the East Antarctic Ice Sheet by statistically analyzing the roughness of the bed topography, which is inferred from radio-echo sounding measurements. The roughness analysis with two roughness parameters enables a classification of the subglacial landscapes below the ice sheet. The roughness parameters are correlated with the flow speed of the ice and modeled basal temperatures. The observed relationships lead to the conclusion that one of the roughness parameters might indicate the thermal condition at the base of the ice sheet. If confirmed by further studies, this could be used as an additional method for predicting the basal thermal regime.


Seeing through the ice: perspectives on a half-century of achievement and progress with ice-penetrating radar

Robert Jacobel

Corresponding author: Robert Jacobel

Corresponding author e-mail: jacobel@stolaf.edu

While it was recognized during World War II that radar waves could be used to penetrate ice, the first scientific airborne deployments were made in the 1970s by the TUD–SPRI–NSF collaboration. Following these continental–scale surveys in Antarctica and Greenland that sought primarily ice thickness information, the next decades saw more dedicated surveys at sites chosen for their glaciological significance, with radar often acquired together with other airborne geophysical data. More recently synthetic aperture recording and processing, together with more sophisticated pulse-compression techniques, have led to radars with greater resolution and dynamic range, allowing not only ice thickness but basal properties to be explored. Also improved depiction of internal reflection horizons with greater fidelity and at greater depth has enabled these to be used in siting ice cores and to study ice dynamics and ice-sheet evolution based on their deformation. Recently the attenuation of radar waves with depth has been used to infer ice-sheet internal temperature. Parallel efforts with impulse radars designed to operate at lower frequencies to circumvent the problem of scattering losses from water inclusions were first made in surface-based surveys in the 1980s, allowing penetration of temperate ice in mountain-glacier environments. Later improvements with folded dipole antenna designs have resulted in highly portable and commercially available impulse radars at higher frequencies that have been useful in firn and shallow-ice studies. Both high- and low-frequency surface-based impulse radars have been used effectively for tracing internal layers with high precision to study ice dynamics and ice-sheet evolution on smaller spatial scales. A more recent development for surface-based studies has been phase-sensitive radar (ApRES) to make very precise range measurements and monitor their changes over time. These have been used, for example, to measure ice-shelf basal melt rates and vertical strain rates in ice sheets, but their utility is just becoming recognized. Beyond the ice on earth, surface-penetrating radars have been deployed to sound the subsurface of the moon and more recently to explore the polar caps of Mars. Missions to the Jovian planets and their icy satellites are in the planning stages. Clearly, the last 50 years have brought remarkable progress and results, and it is exciting to contemplate the science questions that will be addressed in the future.


Ice thickness and bedrock topography Mac. Robertson, Princess Elizabeth and Wilhelm II Lands (East Antarctica) according to the Russian data collected from 1985 to 2018

Sergey Popov, Olga Soboleva, Alexander Kiselev, Valery Masolov

Corresponding author: Sergey Popov

Corresponding author e-mail: spopov67@yandex.ru

The first Russian multidisciplinary investigations in the area of Mac. Robertson and Princess Elizabeth lands (East Antarctica) were conducted during 1971–74 (Operation ‘Aimery’). They included airborne radio-echo sounding (RES) and seismic reflection on the Amery Ice Shelf. The first theories about the glacier structure and subglacial morphology were formed. After a short interruption airborne RES investigations were resumed in 1985 and since 1986 this type of study has been carried out with a 5 km distance between profiles. In the period 2004–14 ground-based RES was carried out in the region of the Mirny–Vostok and Progress–Vostok logistic traverse routes. Data were provided by geodetic and glaciological observations. Ground-based RES data, as the most reliable, complemented the aerogeophysical dataset. The collected data helped to reveal the main features of the ice-sheet structure and the bedrock topography of the extensive coastal area of East Antarctica, including Mac. Robertson, Princess Elizabeth and Wilhelm II lands, and inland about 500 km. In addition, the research revealed a system of subglacial reservoirs in the area of the abandoned Russian station Pionerskaya. Ice-thickness and bedrock-topography maps, which were created during the research, are demonstrated in the presentation.


Noise removal and interpretation of Antarctic ice-penetrating radar data using variational mode decomposition

Xueyuan Tang, Siyuan Cheng, Jingxue Guo

Corresponding author: Xueyuan Tang

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

Ice-penetrating radar is a geophysical method for characterizing the subglacial processes conditions and bed properties of the polar ice sheets. However, noise (strip and ring) from internal instrument and environmental interference is often involved in the radar data, which limits the interpretation of subglacial conditions. Here, using aerogeophysical radar data from the Antarctic ice sheet, we introduce a variational mode decomposition (VMD) that enables a better suppressing of noise, based on the electromagnetic theory of radar scattering. VMD is a decomposition method of adaptive and quasi-orthogonal signals, which decomposes airborne radar data into multiple frequency-limited quasi-orthogonal eigenmode functions (IMFs). The IMFs containing the noise from multi-scale components of the IMFs are removed and the remaining IMFs are then reconstructed. Model tests of forward simulation and ice-penetrating radar data processing show that the VMD method effectively removes interference noise in the data and improves signal-to-noise ratio. Radar scattering reflections are associated with internal layers can be traced well from the noise reduction data. Verification tests of VMD radar data demonstrate that VMD provides an effective model for noise removal and can improve our understanding of the relationship between englacial properties and ice-sheet dynamics.


Passive radio sounding for glaciological investigations of subsurface processes

Sean Peters, Dustin Schroeder, Winnie Chu, Davide Castelletti, Mark Haynes, Andrew Romero-Wolf

Corresponding author: Sean Peters

Corresponding author e-mail: stpeters@stanford.edu

For the last five decades, active radar sounders have been the principal tool to characterize subsurface conditions of ice sheets, but they have been limited by the need to transmit their own powerful electromagnetic signal. If one could eliminate the need to actively transmit a signal for echo detection, this could significantly reduce the overall design complexity, power consumption and cost of glaciological measurements. Here, we demonstrate this as a reality by presenting for the first time ‘passive’ radio sounding to measure the thickness of an ice sheet without transmitting a signal. In this work, we present the results from our field testing using the Sun as a signal of opportunity at Store Glacier in Greenland, where a passive receiver sitting on the surface of the ice sheet records the Sun’s direct path and reflected path off the ice–bed interface. With the received data, we then use our developed autocorrelation-based technique to obtain an estimate of ice thickness. In addition to discussing our passive radar system and its performance, we evaluate its potential to provide critical time-series of glaciological measurements, such as basal melt monitoring, basal reflectivity, englacial water storage and vertical velocities. We conclude by discussing the geographic regions where these observations can best be performed. Similar to the impact that ambient noise seismology had on seismic monitoring, a receive-only passive radar sounder promises to serve as a low-resource method for continuous multi-year observations of glaciers and ice sheets at a continent-size scale.


Radar-detected englacial sediments

Kate Winter, John Woodward, Neil Ross, Stuart Dunning, Andrew Hein, Matthew Westoby, Shasta Marrero, David Sugden, Martin Siegert

Corresponding author: Kate Winter

Corresponding author e-mail: k.winter@northumbria.ac.uk

Glaciers can entrain and transport sediments rich in essential nutrients, like silica and iron, from continental sources to the ocean, where deposition could enhance marine primary productivity. A lack of data from geophysical instruments capable of detecting englacial sediment has led to limited knowledge on the acquisition, transfer and distribution of debris-rich ice in Antarctica. Here, we use ground penetrating radar and airborne radio-echo sounding to detect, and assess the controls on, sediment entrainment and transfer in Antarctic ice flows. We identify sediment reflectors near the glacial surface in katabatic-wind-scoured blue ice areas, and further down the ice column, along partially buried mountain ranges and subglacial bedrock bumps. These sediments are entrained at the thermal boundary between cold- and warm-based ice and transported through the ice by compression towards the mountain front, and regional ice flow towards the coast. Our radargrams highlight the sensitivity of debris entrainment processes and transportation mechanisms to internal and external forcings – related to sediment availability, ice flow and ice temperature. As these controls vary spatially and temporally, changes in the climate and/or internal instabilities in the glacial system could modify sediment sources, alter debris entrainment mechanisms and revise englacial transportation routes – with resultant implications for patterns of continental-sediment-derived nutrient deposition in the Southern Ocean.


A polarimetric coherence method to determine ice-crystal orientation fabric from radar sounding: application to the NEEM ice-core region

Thomas Jordan, Dustin Schroeder, Davide Castelletti, Jilu Li, Jørgen Dall

Corresponding author: Thomas Jordan

Corresponding author e-mail: tom.jordan@bris.ac.uk

The orientation distribution of ice crystals in the polar ice sheets – the ‘ice fabric’ – is an important component of ice rheology. The anisotropy of present-day ice fabric provides a record of past ice deformation, and, in turn, influences the viscosity of ice during future deformation. Our understanding of ice fabric is primarily informed by measurements from ice cores, which tend to be located at ice divides. Remote sensing is, therefore, necessary to place observational constraints upon the spatial development of ice fabric and to assess its role in different flow regimes. Polarimetric radar sounding provides a means to measure ice-fabric anisotropy due to the associated dielectric anisotropy of polar ice. Past radar-sounding analysis methods have focused upon polarimetric power to infer fabric properties. However, this approach can suffer from ambiguity and it is complicated by the combined effect of birefringent propagation and anisotropic scattering upon power. Here we describe a phase-based ‘polarimetric coherence’ method to determine ice fabric from a typical ground-based radar sounding survey (single-polarized data as a function of azimuth). The coherence method provides a direct way to measure the polarimetric phase shift associated with birefringent propagation and from this we can determine horizontal fabric properties: the prevailing crystallographic axis and the asymmetry/strength of the fabric. The new method is demonstrated using MCRDs (Multi-channel Coherent Radar Depth sounder data) from the NEEM ice-core region in Greenland. It is cross-validated using a combination of polarimetric matrix backscatter simulations, ice-fabric data from the NEEM ice core, and comparison with other radar studies. The analysis is consistent with a conventional model of ice deformation at an ice divide, with minor horizontal asymmetry present and the greatest horizontal concentration of crystallographic axes orientated near-parallel to the ice divide. Importantly, rather than being reliant upon quad-polarized data, the coherence method exploits track orientation to generate orthogonal polarization pairs. It can, therefore, be up-scaled to orthogonal cross-over points in non-polarimetric radar-sounding surveys and place constraints upon ice fabric over extended regions of the polar ice sheets.


Estimation of ice fabric within the Whillans Ice Stream using polarimetric phase-sensitive radar sounding

Thomas Jordan, Dustin Schroeder, Cooper Elsworth, Jørgen Dall, Matthew Siegfried

Corresponding author: Thomas Jordan

Corresponding author e-mail: tom.jordan@bris.ac.uk

The net alignment of ice crystals in the polar ice sheets is referred to as the ice fabric. Ice fabric anisotropy provides a record of past ice deformation (strain history) and can also have a pronounced effect upon present-day ice flow. Ground-truth knowledge of ice fabric exists only at a limited number of ice-core sites, which tend to be located at ice divides. Geophysical measurements – both radar and seismic – are used to infer fabric properties in other regions of the ice sheets, and are essential to constrain spatial variation in ice fabric and test ice-dynamic hypotheses. Seismic measurements have recently been applied to ice-stream environments, revealing the presence of developed ice fabrics. To date, radar-sounding surveys of ice fabric have focused upon ice divides, domes and rises, and have yet to be applied to ice-stream environments. We used ground-based polarimetric measurements from ApRES (Autonomous phase-sensitive Radio-Echo Sounder) to investigate ice fabric near to the shear margin of the Whillans Ice Stream, West Antarctica. Our data analysis employs a new phase-based ‘polarimetric coherence’ method to estimate horizontal ice-fabric anisotropy (orientation and asymmetry/fabric strength), and this is compared with an analysis of polarimetric power. At mid-ice depths (z ~ 150–350 m) we infer that the prevailing crystallographic axis is orientated near-perpendicular to the flow direction, whereas in the near-surface (z ~ 10–50 m) it is orientated near-parallel to flow. Furthermore, in the near-surface, we infer that the ice fabric asymmetry increases toward the center of the ice stream, which we interpret as a compression signal that is likely related to the recent deceleration of the Whillans Ice Stream. This study serves as a practical guide for using ApRES in future polarimetric ground surveys of ice fabric. In particular; we show that the polarimetric correlation strength should be used as a guide to select which sections of the ice column can be interpreted and compared spatially.


Measuring and modelling the effects of ice-fabric anisotropy on oblique radio-wave propagation at the South Pole Ice Core Experiment (SPICE) for neutrino detection

Thomas Jordan, Dave Besson, Andrew Romero-Wolf, Dustin Schroeder

Corresponding author: Thomas Jordan

Corresponding author e-mail: tom.jordan@bris.ac.uk

Polar ice exhibits dielectric anisotropy due to the presence of anisotropic crystalline fabric. Polar ice therefore behaves as a birefringent material where double refraction and propagation time delays occur for different radio-wave polarizations. Understanding these effects is essential to estimating the sensitivity of neutrino detectors that are embedded in the Antarctic ice column. These neutrino detectors have a relatively similar frequency sensitivity (~100-1000 MHz) to radar sounding systems, and therefore insight from existing radar-sounding models of ice fabric anisotropy and wave propagation can be used to inform neutrino detection. Here we consider a model–data comparison for oblique radio-wave propagation relevant to the Askaryan Radio Array (ARA) as part of the South Pole Ice Core Experiment (SPICE) for neutrino detection. In December 2018, a custom high-amplitude radio-frequency transmitter lowered into the 1700 m SPICE ice core provided approximately 150K triggers for the five-station, in-ice ARA experiment radio receiver stations, ~2–5 km away. The receiver stations exhibit a time delay for different wave polarizations, indicating the presence of a birefringent material. To better quantify, and predict, the observed behavior we have used SPICE ice-core fabric data to forward model the dielectric tensor of the ice fabric and construct an oblique propagation model. In turn, this model enables us to predict polarization time delays for oblique radio-wave propagation scenarios that are relevant to neutrino detection. The data–model comparison reveals three-dimensional ice-fabric information that is not obtainable from monostatic (nadir) radar sounding or from the SPICE ice core (which does not reference azimuthal fabric orientation). We discuss the consequences of the data–model comparison for three-dimensional fabric inversion from bistatic (off-nadir) radar sounding and constraining past ice deformation patterns in the South Pole region. Additionally, implications for the design of a next-generation radio neutrino experiment, and techniques for utilizing these results to maximize sensitivity, will be discussed.


Subglacial hydrology and newly identified subglacial lakes in the Ellsworth–Whitmore Mountains, West Antarctica

Felipe Napoleoni, Stewart S. R. Jamieson, Michael J. Bentley, Neil Ross, Andrés Rivera, Guisella Gacitúa, José A. Uribe, Andrew M. Smith, Alex M. Brisbourne

Corresponding author: Felipe Napoleoni

Corresponding author e-mail: felipe.a.napoleoni@durham.ac.uk

Recent studies have demonstrated the existence of nearly 400 subglacial lakes across Antarctica, 14% of which are located beneath the West Antarctica Ice Sheet (WAIS). Other studies have shown that some of these subglacial lakes are well connected and are highly dynamic. This has significant implications, because subglacial water can play a fundamental role in ice-sheet dynamics, with implications for ice-sheet mass balance and sea-level rise. Despite the importance of improving knowledge of subglacial hydrology in West Antarctica, particularly in the onset areas of Pine Island and Thwaites glaciers, very little is known about the subglacial hydrology of the Ellsworth–Whitmore Mountains (EWM), and even less is known about how it may have evolved in response to past ice-sheet configurations (e.g. over glacial–interglacial cycles). Previous work in the EWM has been dominated by the characterization of subglacial lakes Ellsworth and CECS. We characterize the subglacial hydraulic conditions of the EWM block to determine the presence of subglacial water and to assess the potential influence of this water on the glaciology of several major WAIS catchments (i.e. Pine Island Glacier, Rutford Ice Stream and Institute Ice Stream). We reanalyse British Antarctic Survey PASIN ice-penetrating radar data to map saturated sediments and subglacial lakes. Incorporating uncertainties related to bedrock temperature and geothermal heat flux, we calculate bed reflection power (BRP) to identify potential subglacial aquatic environments, leading to the identification of >20 new subglacial lakes. In locations of high BRP interpreted as indicative of the presence of subglacial water we classify features using the methodology of Carter and others (2007). Using GIS-based analysis of hydropotential based on new ice thickness measurements, Cryosat-2 ice surface DEM data and ice-sheet model output for snapshots during the Pleistocene we assess the connectivity of these potential lakes during both present and past ice-sheet configurations. This analysis reveals present-day hydrological flow paths through the EWM, and demonstrates that routing of subglacial water through parts of the mountain range can reverse over glacial–interglacial cycles (e.g. from the Weddell Sea towards the Amundsen Sea sector). This work will improve our understanding of the subglacial hydrology of the EWM, and can be used to inform future subglacial-lake access experiments in the EWM (e.g. of Subglacial Lake CECS).


Wire-mesh dipole antenna design for the ApRES radio-echo sounder at HF band (20–40 MHz)

Jonathan D. Hawkins, Lai Bun Lok, Paul V. Brennan, Keith W. Nicholls

Corresponding author: Jonathan D. Hawkins

Corresponding author e-mail: uceeawk@ucl.ac.uk

A prototype ApRES sounder has been developed at HF band (20–40 MHz) to extend the capabilities of the ApRES to sound through deeper or wetter ice and to be practical for imaging areas of ice sheets using ground-based synthetic aperture radar techniques. However, the reduction in system operating frequency introduces challenges in the antenna design, while maintaining the relatively low-cost and high-portability characteristics of the ApRES system. A rollable wire-mesh dipole design is investigated as a solution to achieve an octave bandwidth, electrically small and lightweight antenna suitable for deployment in remote regions of glaciological interest. We first establish the suitability of the theoretical limits on maximum dimensions for antennas operating across this frequency range, before using these limits as a benchmark for the design process. A systematic design approach is presented, undertaken and validated using simulations and laboratory measurements of a one-tenth scaled model of the antenna at VHF. Our initial results validate the use of a wire mesh as an analogue for a solid conductor and show good agreement between simulation and laboratory measurements; however the realized antenna gain is lower than expected with a modest value of –1.2 dBi. Based on these preliminary results we are now continuing to verify the radiation pattern of the antenna, implement an appropriate and tuneable matching network before applying this design method to realize a full-scale implementation of the antenna for the HF band.


Status and prospects of passive sounding with radio-astronomical sources

Andrew Romero-Wolf, Dustin Schroeder, Sean Peters, Bruce Bills, Donald Blankenship, Lorenzo Bruzzone, Bruce Campbell, Leonardo Carrer, Cyril Grima, Essam Heggy

Corresponding author: Andrew Romero-Wolf

Corresponding author e-mail: Andrew.Romero-Wolf@jpl.nasa.gov

The ability to passively sound extraterrestrial ices with radio-astronomical sources has the potential to enhance active radar instruments as well as produce low-cost sounding instruments for Solar System exploration. A concept for passively sounding Jupiter’s icy moons was developed at the Jet Propulsion Laboratory in 2014 and subsequently developed in collaboration with Stanford University. Since then, the technique has evolved to a number of proposed applications ranging from sounding Earth’s cryosphere to the lunar regolith and Solar-System icy bodies. In this presentation I will review the passive sounding concept, its development over the past 5 years and future prospects.


Automated detection and categorization of Antarctic basal units using radar sounding data: demonstration in Institute Ice Stream, West Antarctica

Madison Goldberg, Dustin Schroeder, Davide Castelletti, Neil Ross, Martin Siegert

Corresponding author: Madison Goldberg

Corresponding author e-mail: madisongoldberg@college.harvard.edu

Over the past several years, ice-penetrating radar surveys have repeatedly found evidence of visibly distinct structures near the ice–bedrock interface, a region that has conventionally been considered echo-free. These basal units warrant investigation for a number of reasons: many are of unknown composition and origin, characteristics that could lend substantial insight into near-bed processes. They have also been shown to affect the stratigraphy of the surrounding ice column, indicating their significance in ice-sheet dynamics beyond the bedrock. In order to enable improved characterization of these features, we developed and applied an algorithm that allows for automatic detection. We used a tunable, layer-optimized SAR processor to distinguish the basal units from the bedrock, englacial layers and ice sheet surface. We theorize about the relation between the basal units’ echo characters – specifically, their response to our processor – and their composition. We investigate one basal unit in particular, in West Antarctica’s Institute Ice Stream, and assess the probability of three hypothesized modes of formation. Through this case study, therefore, we introduce a technique that can simultaneously detect these structures and impose constraints on their provenance.


Passive radio sounding with Jupiter's radio emissions to correct for Europa's ionospheric distortion

Sean Peters, Dustin Schroeder, Mark Haynes, Andrew Romero-Wolf

Corresponding author: Sean Peters

Corresponding author e-mail: stpeters@stanford.edu

Jupiter’s icy moon Europa is an exciting candidate for planetary radar sounders, as it potentially hosts a global subsurface ocean. In order to unambiguously confirm the subsurface-ocean hypothesis and constrain the thickness of ice on Europa, previous work has highlighted the potential to use a radar sounder as an altimeter to measure Europa’s love number, in addition to its primary objective of attempting to measure the subsurface. However, Europa’s ionosphere presents a significant challenge as it distorts the radar signal by creating pulse broadening and signal delay. This effect severely impacts the precision of the altimetric measurements necessary to constrain the amplitude and phase of gravitational tides on Europa and accurately measure its love number. While using the relative delay between two active radar channels of different center frequencies is a common reference-based method to correct for ionospheric distortion, the high-frequency (HF) active band for a dual-frequency planetary sounder such as the Radar for Europa Assessment and Sounding: Ocean to Near-surface (REASON) is susceptible to domination by the bursts of Jupiter’s radio emissions when the radar is on the sub-Jovian side. At these times, the HF radar will be severely degraded alongside the slightly dispersed very-high-frequency (VHF) channel for performing radar sounding of Jupiter’s icy moon as well as reference-based ionospheric distortion corrections of the VHF altimetric measurement. To address this, we assess the potential to use Jupiter’s radio emissions to correct for ionospheric distortion with a passive radio-sounding approach that has been developed on Earth to estimate the thickness of ice sheets using the Sun. Here, we discuss the abilities and requirements for a dual-frequency radar system using a passive HF mode alongside a VHF radar to correct for ionospheric distortion, and we compare its performance to that of an active HF radar system. The results of our simulations with a receive-only HF mode show that, even in the presence of significant total electron content (TEC) in Europa’s ionosphere, the passive radar approach not only corrects for distortion but also performs at a level similar to that of an active HF system in terms of TEC estimation and range-error reduction of VHF radar measurements.


More than 40 years of ice radar soundings in Germany

Heinrich Miller, Daniel Steinhage

Corresponding author: Heinrich Miller

Corresponding author e-mail: heinrich.miller@awi.de

For the past 40 years radar surveys were carried out at different institutions in Germany. In this review we also aim to show how systems were developed and adapted to the requirements posed by the glaciological questions. We will show examples from temperate glaciers as well as the polar ice sheets and how this leads up to the most recent radar surveys. Comparisons between data from earlier systems and the most recent ones will allow to demonstrate the advances in resolving internal structures as well as properties of ice/basement interface.


What radar reveals about crystal orientation: a study from the Greenland Ice Sheet

Ole Zeising, Olaf Eisen, Ilka Weikusat, Nicolas Stoll, Angelika Humbert

Corresponding author: Ole Zeising

Corresponding author e-mail: ole.zeising@awi.de

Deformation processes dominated by dislocation activity within ice sheets take place on small scale: ice crystals are effectively re-oriented to minimize resistance when deformation takes place. The analysis of crystal orientation fabric (COF) of c-axes in ice cores is a well-established technique to investigate these processes within ice sheets. Due to the extensive infrastructure required for drilling and protracted analysis of ice cores, the amount of information of COF of the Antarctic and Greenland ice sheets is limited, both in aerial coverage and in depth resolution. Indirect measurements, such as geophysical techniques, provide complementary information. Depending on whether the COF is isotropic or anisotropic, a radar signal is propagating differently in terms of angle of incidence and polarization, and partially reflected when COF properties change along the direction of travel. We study the ability of phase-sensitive radar measurements to infer an overall pattern of COF by comparing our results to COF derived from the EastGRIP ice core, drilled into the North East Greenland Ice Stream. If radar measurements allow the revelation of information about the COF as analyzed ice cores do, this would provide important additional information on the (an)isotropy at locations where no ice core is available. Furthermore, it has the ability to offer a quasi-continuous spatial coverage and to greatly improve our understanding of the evolution of anisotropy along ice-flow trajectories, from ice divides to the calving front of outlet glaciers.


Glacier thickness estimations using glaciological constraints and ground-penetrating-radar data

Lisbeth Langhammer, Andreas Bauder, Melchior Grab, Hansruedi Maurer

Corresponding author: Hansruedi Maurer

Corresponding author e-mail: hansruedi.maurer@erdw.ethz.ch

Advanced knowledge of ice thickness distributions within glaciers is of fundamental importance for several purposes, such as water-resource management and studying impacts of climate change. Ice thicknesses can be modeled using ice-surface features and mass-balance considerations, but the resulting models can be prone to considerable uncertainties. Alternatively, it is possible to measure ice thicknesses, for example, with ground-penetrating radar (GPR). Such measurements are typically restricted to a few profiles, with which it is not possible to obtain spatially unaliased subsurface images. We developed the Glacier Thickness Estimation algorithm (GlaTE), which optimally combines modeling results and measured ice thicknesses in an inversion procedure to obtain overall thickness distributions. Properties and benefits of GlaTE are demonstrated with three case studies performed on different types of alpine glacier. In all three cases, subsurface models could be found that are consistent with glaciological modeling and GPR data constraints. Since acquiring GPR data on glaciers can be an expensive endeavor, we additionally employed elements of sequential optimized experimental design (SOED) for determining cost-optimized GPR survey layouts. Our results indicate that a relatively large amount of data can be acquired before the benefit–cost curves enter into the area of diminishing returns, where it becomes increasingly expensive to obtain further information. Only at one out of the three test sites was this level reached.


Producing multi-decadal observations of grounding line change in East Antarctica with archival radar data

Emma MacKie, Michaela Murray, Andrew Pollack, Dustin Schroeder

Corresponding author: Emma MacKie

Corresponding author e-mail: mackie3@stanford.edu

Change in grounding-line position is one of the most tangible ways to diagnose glacier retreat but, until recently, multi-decadal observations of grounding-line positions have not been available for Antarctica. Between 1967 and 1979, airborne RES surveys were taken of Antarctica in a collaboration between the Scott Polar Research Institute (SPRI), the National Science Foundation (NSF) and the Technical University of Denmark (TUD), known as the SPRI–NSF–TUD surveys. These surveys are the oldest observations of internal and subglacial features in Antarctica and provide the opportunity for comparison with modern ice-penetrating radar campaigns in order to study temporal changes in the ice sheet over the last 50 years. Data was collected in the form of ‘Z-scope’ (radargram) and ‘A-scope’ (power return over time) format. These data were recorded on 35 mm film, making it difficult for such comparisons to be made. But the recent digitization of this archive enables the observation of temporal changes in subglacial and englacial conditions. We discuss the methods used to position and analyze these data and present preliminary results for half a century of grounding-line observations in East Antarctica. We also present a preliminary investigation of the A-scope data and how it can be used to study basal properties.


Geostatistical simulations of subglacial topography used to study paleo and modern bed condtions in the Amundsen Sea sector

Emma MacKie, Dustin Schroeder

Corresponding author: Emma MacKie

Corresponding author e-mail: mackie3@stanford.edu

Comparing swath bathymetry data with ice-penetrating radar data is one of the only direct ways to place modern subglacial conditions in a historical context and make inferences about the evolution of ice sheets and morphology. However, radar and bathymetry have different ranges of coverage so such comparisons are not rigorous. Gaps in radar data are typically filled by kriging, creating unrealistically smooth bed topography and limiting morphological interpretations. We have developed a technique for making these comparisons and demonstrate its capabilities in the Amundsen Sea. We do this through simulating radar surveys over bathymetry data in the Amundsen Sea sector so that we have an equivalent dataset to compare with radar data. Geostatistical simulations are used to generate synthetic topographic information between the flight lines. HiCARS, PASIN and MCoRDS radar bed picks over the Thwaites and Pine Island glaciers are used to generate geostatistical simulations of modern subglacial topography. We compare the seafloor and subglacial topographic simulations and demonstrate the morphological differences between the two. We simulate subglacial basins, which play an important role in hydrological processes. We also simulate topography beneath the ice shelf and grounding line to improve grounding-zone interpretation and observe changes in geomorphological patterns. The use of geostatistics to improve radar interpretation helps characterize bed regions for models and provides insights into both the contemporary and paleo bed conditions during retreat.


Surface-based multi-channel radar systems for ice-sheet measurements

Jie-Bang Yan, Joshua Nunn, Siva Gogineni, Charles O’Neill, Christopher Simpson, Ryan Taylor, Linfeng Li, Shashank Wattal, Daniel Steinhage

Corresponding author: Jie-Bang Yan

Corresponding author e-mail: janunn@crimson.ua.edu

The feasibility of sounding of ~1 km-thick ice at UHF frequencies has recently been demonstrated using a 1 W airborne radar that was originally designed for mapping near-surface accumulation layers. As compared to conventional ice sounders operating at VHF, it is much easier to integrate a wideband and electrically large aperture on airborne and spaceborne platforms to obtain significantly improved vertical resolution and spatial selectivity to suppress surface clutters that could mask ice-bottom return. In order to fully validate the concept of ice sounding at UHF frequencies as well as to image internal layers within the bottom 10% of ice, a surface-based eight-channel radar system operating from 600 to 900 MHz with a very high sensitivity has been designed to image and map ice sheets of more than 3 km thick. The high sensitivity is achieved through the use of 100 W power amplifiers together with an electrically large ultra-wideband UHF monopole antenna array arranged in a Mill’s Cross configuration that extends over 16 m × 17 m. The UHF radar system, together with a back-up VHF radar system operating from 170 to 230 MHz with a peak transmit power of 600 W, was deployed to the East Greenland Ice-coring Project (EGRIP) site in Summer 2018 to demonstrate the concept and collect experimental ice-sheet sounding and imaging data. Both radar systems, sharing the same digital waveform generator and digitizer, were installed and operated inside a tracked vehicle. For the first time, we have successfully demonstrated UHF radar sounding of ice of ~2.6 km thick. We will present the design of the two radar systems and the radar data collected from EGRIP. We will also describe the plan to deploy the radar system to Antarctica to assist the search of the ‘oldest ice’.


Improving constraints on englacial temperature and water distribution using an autonomous phase-sensitive radio-echo sounder (ApRES) and a bistatic software defined receiver

Nicole Bienert, Dustin Schroeder, Sean Peters, Matthew Siegfried

Corresponding author: Dustin Schroeder

Corresponding author e-mail: bienert@stanford.edu

The behavior, evolution and sea-level contributions of outlet glaciers and ice streams are highly dependent on the location and stability of shear margins. However, the physical processes governing their spacing, formation and migration are not well understood. Recent modeling studies have indicated that shear-margin behavior may be controlled by elevated temperatures and water content in the bottom half of the shear margin, while other studies predict that ice shear margins exist over concentrated subglacial water channels. To test these hypotheses and improve our understanding and modeling of processes governing shear-margin evolution, we are developing a direct measurement technique to map the temperature distribution and water content in the cross-section of a shear margin. The system also has the potential to identify ice-fabric orientation in and around the shear margin to discern ice viscosity and how the ice moved in the past. The bi-static measurements are conducted with an ApRES acting as a transmitter and a low-cost software defined radio (SDR) as a receiver. A method was developed to attain high SNR gain through coherent summation of pulse compressed signals without radio synchronization. We are moving towards a multistatic, fully SDR design because it enables 3-D map generation and makes the system useful for a wider variety of applications. Here, we present simulation studies and preliminary test results from Store Glacier, Greenland, and the Siple Coast, Antarctica. In the fall of 2019 this approach will be applied to the eastern shear margin of Thwaites Glacier, which separates it from the Pine Island catchment and is the only topographically unconfined shear margin in the rapidly changing Amundsen Sea Embayment of West Antarctica.


Radar scattering in firn and the implications for orbital ice sounding

Riley Culberg, Dustin M. Schroeder

Corresponding author: Riley Culberg

Corresponding author e-mail: culberg@stanford.edu

Radar sounding from orbit has the potential to provide the spatial and temporal data coverage required to address many open questions in subglacial processes and ice dynamics. While the success of the Martian sounding missions demonstrates that many of the associated engineering challenges are surmountable, the results of high frequency and high-altitude airborne experiments suggest that particular care is required to optimize the system design for the terrestrial environment. One significant concern is that bright radar scatterers in polar firn may contribute off-nadir returns that obscure much weaker echoes from the bed or englacial layers. We have developed two glaciologically constrained models of electromagnetic scattering in dry firn and show that scattering behavior is dominated by the thin film interference of reflections at interfaces of varying density, whereas volume scattering from air inclusions plays a negligible role. Additionally, we model the angular scattering behavior with altitude for the ice surface and these near-surface layers. We apply these models together to constrain the choice of system center frequency and bandwidth for orbital ice sounding.


How to hack Your ApRES

Nicole Bienert, Dustin Schroeder, Ha Tran, Michaela Murray

Corresponding author: Dustin Schroeder

Corresponding author e-mail: Dustin.M.Schroeder@stanford.edu

Several methods are evaluated for improving the performance of the autonomous phase-sensitive FMCW radar (ApRES). First, the artificial curvature of the bed in 2-D images is addressed by decreasing the sidelobes through improvement of the antenna array factor. The numbers of antenna elements, antenna spacing, transmit power distribution and applied phase shifts are optimized to decrease the sidelobes while keeping the beam-width narrow. To validate simulated results, experiments are conducted to validate simulated results by altering the digital beam-forming code and increasing the number of antenna elements. The number of antennas is increased by externally adding power splitters, switches and power amplifiers to the Tx/Rx ports of the ApRES. The effects of increasing transmitted power are evaluated separately because normal ApRES measurements could benefit from increased transmit power if the system is not being used for a full year deployment. Currently, A low-cost software defined radio (SDR) radar system is being investigated as an alternative to the ApRES because it allows for greater frequency and waveform diversity, making the system useful for a wider variety of applications.


Constraining the englacial and basal thermal state in Dome Fuji, East Antarctica, with radar attenuation models

Alex Miltenberger, Eliza Dawson, Dustin Schroeder

Corresponding author: Alex Miltenberger

Corresponding author e-mail: ammilten@stanford.edu

Estimating the thermal state within a glacier and at the bed is important for many glaciological applications. Examples include oldest-ice reconnaissance, ice-sheet dynamics and more. However, widespread measurements of thermal profiles in glaciers are not available simply because currently temperatures can only be measured in drill holes. We propose using radar data, particularly attenuation, to overcome this limitation and constrain the thermal state away from drill holes. First, the problem will be formulated in a Bayesian framework, outlining each of the key steps. After describing the methodology, we will show the results of a case study in Dome Fuji, East Antarctica. Finally, we discuss how this approach can be applied to other parts of Antarctica.


DeepBedMap: a super-resolution deep neural network for resolving the bed topography of Antarctica

Wei Ji Leong, Huw Horgan

Corresponding author: Wei Ji Leong

Corresponding author e-mail: weiji.leong@vuw.ac.nz

To better resolve the bed elevation of Antarctica, we present a novel deep convolutional neural network that produces realistic terrain given multiple remote-sensing data inputs. Our super-resolution DeepBedMap neural network model is trained on scattered regions in Antarctica where high-resolution groundtruth bed elevation grids are available, and later used to generate high-resolution bed topography in areas less well surveyed. DeepBedMap improves on previous interpolation methods by not restricting itself to a low spatial resolution (1000 m) BEDMAP2 raster image as its prior. It takes in additional high-spatial-resolution datasets, such as Antarctic ice-surface velocity and surface elevation maps, which can be used to better inform the subglacial bed’s topography even in the absence of direct ice-penetrating radar survey data. Our DeepBedMap model is based on an adapted Enhanced Super Resolution Generative Adversarial Network architecture, chosen to minimize the per-pixel elevation error while producing realistic topography. The final product is a four times upsampled (250 m) bed-elevation model of Antarctica that can be used by glaciologists interested in the subglacial terrain and also by ice-sheet modelers wanting to run catchment or continent-scale ice-sheet model simulations. We show that DeepBedMap produces a more accurate digital elevation model than a baseline bicubic interpolation product, and also compare it with other synthetic bed elevation models on reference groundtruth survey tracks.


Investigation of subsurface structure at the edge of the ice sheet based on GPR sounding in East Antartica

Jingxue Guo, Xueyuan Tang

Corresponding author: Jingxue Guo

Corresponding author e-mail: guojingxue@pric.org.cn

Ground penetrating radar (GPR) is one of the most effective approaches to detect the shallow structure of an ice sheet and to find the crevasses and cracks. During the 2017/18 season, we used a SIR 20 ground-based pulse radar system to measure with a 100 MHz antenna and a 400 MHz antenna simultaneously at the edge of the ice sheet near Zhongshan Station in east Antarctica, in order to obtain information about the shallow structure of the ice sheet. To improve the signal-to-noise of the data, DC removal, background removal, Butterworth filtering and moving average were applied to the data processing. After conversion of two-way travel to depth, we present the thickness of the ice sheet and the undulation of the topography on the survey profile. Based on the radar data analysis, the shape of crevasses appear on the radar profile and the distribution of crevasses is shown. This study also infers the formation and relationship of the crevasses to bedrock,with the aim of contributing to research and exploration on the ice sheet.


Doppler-based discrimination of radar sounder target scattering properties: a case study of subsurface water geometry in Europa's ice shell

Roger Michaelides, Dustin Schroeder

Corresponding author: Roger Michaelides

Corresponding author e-mail: rmich@stanford.edu

Radar-sounding reflectometry is a powerful method of inferring subsurface geophysical properties of planetary bodies. The upcoming Europa Clipper Mission will employ a dual-frequency radar sounder to image and characterize Europa’s subglacial environment. In terrestrial subglacial radar sounding, characteristic Doppler signatures of radar returns have been used to discriminate between different subglacial features. We propose that analogous techniques can be used in conjunction with traditional reflectometry methods to discriminate between different proposed models of subglacial and englacial water distribution within Europa’s ice shell. We introduce the synthetic aperture radar (SAR) processing considerations that are relevant for such a technique, propose a simple hypothesis test for Doppler-based subglacial target discrimination, and briefly discuss the implications and applicability of such a technique for planetary radar sounding.


Buried ice and sand caps at the north pole of Mars: revealing a record of climate change in the cavi unit with SHARAD

Stefano Nerozzi, John Holt

Corresponding author: Stefano Nerozzi

Corresponding author e-mail: stefano.nerozzi@utexas.edu

The cavi unit is an aeolian deposit of sand and water ice making up a large fraction of the polar deposits in the northern hemisphere of Mars. Previous studies determined an age for this unit of 10–100 Ma. Its stratigraphic position just underneath the north polar cap of Mars implies that its strata record climate conditions and processes during a global climate transition. Prior studies involving imagery, spectrometry and radar sounding determined that the cavi unit is composed of a mixture of water ice and lithic materials. However, there is still no consensus on its precise composition, and it remains unclear which fraction, water ice or sand, dominates. Precise constrains on composition are needed to determine the importance of the cavi unit as a water and sediment reservoir, reconstruct its accumulation history and understand the climatic and other processes that resulted in its observable morphology and stratigraphy. Hundreds of profiles acquired by the shallow radar (SHARAD) reveal a deep underneath the cavi unit, which we interpret as its basal surface. This detection allows us to use a simple inversion technique to determine the dielectric constant (real part of dielectric permittivity) of this unit and, in turn, its bulk composition. Our exercise reveals a low bulk dielectric constant of the cavi unit, compatible with volume ice fractions of 62–88%. The dielectric constant also has substantial spatial variability, with a trend decreasing towards the north pole. Several reflectors appear in the cavi unit northern reaches, indicating the presence of numerous layers with varying compositions. In the northern reaches of the unit, water ice appears to be the dominant fraction at 80–90%. Based on the observation of several internal reflectors, we hypothesize that alternating sandy and pure water-ice layers make up the cavi unit. Several water-ice accumulation models predict substantial water-ice accumulation during periods of low spin axis obliquity before the onset of the north polar cap. We hypothesize that some of this ice was buried and preserved by aeolian sand sheets. We observe similar sand mantles overlying thick ice deposits in visible outcrops. Therefore, we argue that previously accumulated ice caps are not necessarily lost during high obliquity periods, but preserved within sand sheets. Moreover, the high water-ice fraction potentially makes cavi the third largest ice reservoir on Mars.


Observational constraints from englacial layers on fast flow initiation of a West Antarctic ice stream

Elisa Mantelli, Dustin Schroeder, Helene Seroussi, Marnie Bryant, Davide Castelletti, Martin Siegert, Jenny Suckale

Corresponding author: Elisa Mantelli

Corresponding author e-mail: emantell@stanford.edu

The physical processes governing the organization of ice flow into ice streams are of major interest with respect to predictions of future ice-sheet mass loss. A variety of elements have been proposed that might explain ice-flow acceleration. On the one hand observations of contemporary ice streams suggest that fast flow is favored by overdeepenings in bed topography and sediment beds. On the other hand, dynamic processes involving either thermo-mechanical feedbacks within the ice and at the ice–bed contact, or subglacial hydrology have been proposed, which would carry implications with respect to the spatio-temporal dynamics of ice streams. While basal topography and lithology are reasonably well known, dynamic feedbacks remain challenging to constrain with direct observations, and therefore theoretical predictions on the key physical processes controlling fast flow initiation are, to date, largely untested in the real world. In this study we seek to fill this gap in the context of thermo-mechanical feedbacks. These feedbacks come in two basic forms: either ice becomes less viscous at higher temperatures, or sliding between ice and bed is temperature-dependent. In this study our focus is on the latter. First, we briefly review recent theoretical work by one of the authors that describes an instability inherent in ice flow across cold–temperate basal transitions commonly encountered in the onset region of ice streams. This instability relies on the physics of subtemperate sliding, which the theory predicts to occur in extended portions of the ice sheet. Informed by this finding, we turn the attention to a major Antarctic ice stream – Institute Ice Stream – and attempt to characterize the transition from basal no slip to basal sliding observationally. To do so, we follow two lines of evidence: (i) we perform a catchment-scale inversion for bed slipperiness with the Ice Sheet System Model, as constrained by the current geometry and velocity field; (ii) we apply a novel Layer Optimized Synthetic Aperture Radar processing to selected flight lines of the BAS-PASIN 2010/2011 radar survey, and compare the architecture of englacial layers in this region to model predictions. Our results so far are consistent with the hypothesis that the onset region of Institute Ice Stream is characterized by a spatially extended transition from basal no slip to basal sliding, and the observed layer architecture is compatible with an extended region of subtemperate sliding.


Chinese radioglaciological survey and studies in Antarctica in recent decades

Xiangbin Cui, Bo Sun, Lin Li

Corresponding author: Xiangbin Cui

Corresponding author e-mail: cuixiangbin@pric.org.cn

Field ice-penetrating radar measurements in Antarctica and radioglaciological studies were initiated relatively late in China. During CHINARE 21 (Chinese National Antarctic Research Expedition, 2004/05), a Chinese inland traverse team arrived at Dome A in the deep interior of East Antarctica from the coastal Zhongshan Station for the first time, and radar measurements were carried out along the traverse and in Dome A with a detailed grid. Since then, different kinds of radar system have been deployed for both fine-scale surveying networks in the Dome A and along the traverse for multiple scientific purposes, such as mapping subglacial conditions, distinguishing crystal orientation fabric types, measuring internal layers and freeze-on ice, searching for the oldest ice on Earth and mapping the unknown bedrock topography of the Gamburtsev Subglacial Mountains (GSM) and inferring the early origin and evolution of the Antarctic Ice Sheet and the GSM. In 2015, China deployed its first fixed-wing airplane, ‘Snow Eagle 601’, for both scientific and logistical activities in Antarctica. The airborne scientific system was configured with an integrated airborne deep ice penetrating system (HiCARS), GT-2A gravimeter, CS-3 magnetometer, laser altimeter, camera and GPS. Meanwhile, an international campaign of International Collaborative Exploration of the Cryosphere through Airborne Profiling in Princess Elizabeth Land (ICECAP/PEL) was initiated. So far, we have finished four austral seasons of airborne survey in East Antarctica and the coverage includes Princess Elizabeth Land, the biggest data gap in Antarctica, the Amery Ice Shelf, West Ice Shelf, Ridge B, etc. These measurements will help us to study subglacial water and lakes, subglacial canyons and channels and the subglacial hydrological system, basal processes across grounding zones andbathymetry under ice shelves in the region and will have great implications for investigation of ice sheet expansion, stability and the subglacial geology of East Antarctica. In future, we will engage in and make Chinese contributions to more science issues.


A new view of a 1970s radar dataset from Greenland

Nanna B. Karlsson, Louise S. Sørensen, Sebastian B. Simonsen, Jørgen Dall

Corresponding author: Nanna B. Karlsson

Corresponding author e-mail: nbk@geus.dk

The uncertainties related to predicting future mass loss from the ice sheets are partly due to the short observational record. The quantity and quality of observations of the Polar regions are ever-increasing but prior to the 1990s observations are scarce. Here we present a new view of a unique ice-penetrating radar dataset that was acquired over the Greenland Ice Sheet from 1969 to the early 1980s. The dataset was acquired as a part of a large-scale project that ran over multiple years and covered more than 170 000 km of radar flight lines. While the ice-thickness information from the data has subsequently been digitized, the data itself is presently only available as 35 mm films, microfiche copies of the films and enlarged positives. These formats do not render themselves easily to modern digital analysis tools, and thus the dataset has hitherto been underutilized and overlooked. Here we present the first results from digitization of several flightlines acquired in the 1970s. While large navigational errors introduce uncertainties in the exact geolocation of the flightlines, we are able to retrieve valuable information on the state of the Greenland Ice Sheet in the 1970s. This includes information on layer stratigraphy, shear-margin migration and firn aquifers.


Deconvolution of surface-based FMCW ice-sounding radar for mapping internal reflecting horizons and snow/firn accumulation in Antarctica

Shinan Lang

Corresponding author: Shinan Lang

Corresponding author e-mail: langshinan@bjut.edu.cn

We present a novel system response deconvolution algorithm to achieve range-sidelobe reduction and range-resolution improvement of surface-based frequency-modulated continuous wave (FMCW) ice-sounding radar. The proposed algorithm consists of two subparts: the deconvolution strategy for nonideal radar system effect, and the deconvolution process for antenna distortion. The algorithm introduces the homomorphic deconvolution strategy containing the unwrapping method integrated with RVP removal technique to realize the nonideal radar system component estimation to adapt to the range-dependent overlapped imperfection in the beat frequency signal, which considerably reduces operation complexity. In addition, the acceleration R-L deconvolution algorithm is imported to promote noise robustness. The theory analysis and implementation steps of the proposed algorithm are demonstrated, and full-scale simulations with different kinds of radar system imperfection verify the effectiveness and robustness of the proposed algorithm. Data collected by our designed shallow-layers-detection ice-sounding radar (SLDISR) during the 31st Chinese Antarctic Research Expedition (CHINARE 31) and CHINARE 33 show that our algorithm can illustrate clarified IRHs of ice sheets. Compared with the echograms processed by the traditional system response deconvolution algorithm, the proposed algorithm is better able to improve range resolution improvement.


Detailed bedrock topography in the vicinity of the EGRIP ice-core drill site

Steven Franke, Daniela Jansen, Tobias Binder, Veit Helm, Daniel Steinhage, Nils Dörr, John Paden, Ilka Weikusat, Paul Bons

Corresponding author: Steven Franke

Corresponding author e-mail: steven.franke@awi.de

Fast-flowing ice streams are known to contribute a large proportion of continental ice-sheet discharge and, therefore, to sea-level variability. In this study we investigate the internal layering, stratigraphy and bedrock topography of the North East Greenland Ice Stream (NEGIS). In conjunction with physical property data from the EastGRIP ice core, these data enable us to expand insights from a geographically limited dataset to a broader area to gain insights into the complex mechanisms governing ice streams. We acquired airborne radar data at the NEGIS in the vicinity of the EGRIP drill site. The data were collected in May 2018 in different acquisition modes with AWI’s ultra-wideband multichannel radar installed on the AWI Polar 6 Basler BT-67 aircraft. A total area of 16 000 km2 has been mapped with profiles along and perpendicular to ice flow. The area in the immediate vicinity of the drill site has been covered with a profile spacing of 5 km further downstream and upstream the profile spacing is 10 km. The survey area reaches from 150 km upstream to 150 km downstream of the drilling sites, and also includes both shear margins and parts of the slow flowing areas adjacent to the ice stream. Our fine-resolution radargrams provide detailed insights into the bedrock topography, including the ~500 m step just upstream of the drill site. Furthermore, we observe an increase in bedrock roughness perpendicular to ice flow in the downstream direction. We calculate ice thickness and bedrock depth and compare these findings with the BedMachine data set. In the southeast of the NEGIS our radar data show large-scale folded structures that have been detected in previously collected radar data. Our dense grid of radar profiles allows for a detailed 3-D representation of these features, which are orientated parallel to the ice stream.


Common offset velocity analysis of GPR: implications for glacier bed topography retrieval

Richard Delf, Robert G Bingham, Andrew Curtis, Satyan Singh

Corresponding author: Richard Delf

Corresponding author e-mail: r.delf@ed.ac.uk

Ground-penetrating radar (GPR) data are widely used on polythermal and temperate glaciers to sound bed topography and investigate the hydrothermal conditions through detection of englacial radar scattering. Water held within micro- and macro-scale pores and ice grain boundaries in ice at the pressure melting point influences the velocity of radar propagation and results both in incoherent and coherent scatter, the latter often in the form of diffraction hyperbolas. Methods to investigate the water content distribution quantitatively within temperate ice often require the use of multi-offset common midpoint or common source-point survey techniques, which are logistically challenging and expensive. As a result, bed topography estimation is often undertaken using a constant velocity, and, because lateral variations in the velocity field are unaccounted for, errors in topography are likely. Here, we present a semi-automated workflow to estimate an englacial radar velocity field from common offset data and apply the algorithm to GPR data collected on Von Postbreen, a polythermal glacier in Svalbard, using a 25 MHz zero-offset GPR system. We first extract the diffracted wavefield using local coherent stacking to remove scatter and enhance diffractions. We then use the focusing metric of negative entropy to deduce a local migration velocity field from constant-velocity migration panels and produce a glacier-wide model of local (interval) radar velocity. We show that this velocity field is successful in differentiating between areas of cold and temperate ice and can detect lateral variations in radar velocity close to the glacier bed. The effects of this velocity field in both migration and depth-conversion of the bed reflection are shown to result in consistently lower ice depths across the glacier, indicating that diffraction focusing and velocity estimation are crucial in retrieving correct bed topography in the presence of temperate ice.


Distribution and dynamics of Greenland subglacial lakes

Jade Bowling, Stephen Livingstone, Andrew Sole, Winnie Chu

Corresponding author: Jade Bowling

Corresponding author e-mail: j.bowling@lancaster.ac.uk

Few subglacial lakes have been identified beneath the Greenland Ice Sheet (GrIS) despite extensive documentation in Antarctica, where periodic release of water can impact ice flow and direct sampling reveals a unique ecosystem of complex microorganisms. We present an ice-sheet-wide survey of subglacial lakes beneath the GrIS using airborne radio-echo sounding (RES), acquired from Operation IceBridge (1993–2016). We identify 54 candidates from RES, and two hydrologically active subglacial lakes from ice-surface elevation changes. These range from 0.2–5.9 km in length, and are mostly distributed away from ice divides, beneath relatively slow-moving ice. We also observe subglacial lakes concentrated in ‘uncertain’ regions of predicted basal thermal state where ice-sheet models and ice-penetrating radar cannot agree on whether the bed is frozen or thawed. Based on our results and previous observations we suggest three subglacial lake formation zones: (i) stable lakes in northwestern and eastern regions above the equilibrium line altitude (ELA) but away from the largely frozen interior; (ii) hydrologically active lakes near the ELA recharged by surface meltwater; and (iii) small, seasonally active lakes below the ELA, which form over winter and drain during the melt season. These observations provide important constraints on the GrIS basal thermal regime and help refine our understanding of the subglacial hydrological system.


Distribution and dynamics of Greenland subglacial lakes

Richard Delf, Robert G. Bingham, Dustin M. Schroeder

Corresponding author: Richard Delf

Corresponding author e-mail: r.delf@ed.ac.uk

Radar surveys across ice sheets and ice caps typically sound numerous englacial layers which can be treated as isochrones. In principle, isochrone geometry has many applications for both palaeo and contemporary glaciology and climatology, for example extrapolating age-depth information beyond ice-coring sites, reconstructing palaeoaccumulation variability and investigating past ice-sheet dynamics. However, the use of radar-sounded englacial layers for such purposes in Antarctica has been hampered by underdeveloped techniques for characterizing layer continuity and geometry over large distances. Motivated by a new international effort to develop a continent-wide model of Antarctica’s internal architecture, we present an assessment of potential ways forward for characterizing englacial-layer geometry using automated methods of processing. Here we present a range of test-case scenarios, selected from CReSIS MCoRDS datasets over Antarctica, to test the application of automated approaches to englacial-layer model building. These test cases are selected from a range of challenging environments, including broken and disturbed layering with rapid variations in dip, and ice-stream-induced deformation. Manual interpretation has been undertaken to pick internal layers, and metrics of vertical picking error statistics, lateral continuity and dip error are presented to assess the effectiveness of a range of demonstrative, simplified algorithms. These include correlation and maximum-peak tracking, image enhancement and thresholding, and local dip estimations. These interpretations will be made available for the benchmarking of future approaches to automated mapping of internal layers, to assess objectively the uncertainties and potential failure points for each approach.


Swath topography, and the future of polar bed mapping

Nicholas Holschuh, Knut Christianson, John Paden

Corresponding author: Nicholas Holschuh

Corresponding author e-mail: nick.holschuh@gmail.com

Advancement in radar electronics and processing methods, largely developed by faculty and students at the University of Kansas, allow for both along-track and across-track focusing of ice-penetrating radar data. The densely spaced ice-thickness measurements these systems produce can resolve the subsurface with unprecedented detail, allowing us to reevaluate the scope of questions addressed by radioglaciology and forcing us to develop new strategies for system selection, survey design and interpretation. While individual flights produce ~O(25 m) spacing bed topographies across a 2 km swath width, this technology does not eliminate all of the challenges we face for science applications of nadir-focused radar data today:
1) Large discrepancies in observation density within a given survey region are still likely (in the case of swath mapping, when airborne data is collected with line spacing greater than the 2 km necessary for swath overlap). To supply useful products for modelers, this typically results in either (a) degraded observation quality by down-sampling to the lowest possible resolution (the flight line spacing) or (b) information lost as part of the interpolation and smoothing required to make a continuous grid.
2) With increasingly dense coverage in Antarctica and Greenland, improvements to our understanding of bed topography seem incremental, and collecting new topographic information from radar surveys is often less motivated by fundamental glaciological questions. Scientific applications can and should extend beyond improving gridded products for ice-sheet modelers. As community driven aerogeophysical campaigns move forward, it is important that the radioglaciology community identify new priority targets and ice-dynamic questions that can be observationally addressed with swath data.
Here, we present a single application using swath topography at Thwaites glacier to evaluate the diversity of models for glacial bedform development, highlight possible applications of existing swath-capable data and motivate discussion of future swath radar campaigns.


The limits on ice-fabric analysis using radar anisotropy at Hercules Dome, Antarctica

Nicholas Holschuh, Knut Christianson, Benjamin Hills, Eric Steig

Corresponding author: Nicholas Holschuh

Corresponding author e-mail: nick.holschuh@gmail.com

The temporal evolution of ice divides in Antarctica and Greenland is a topic of increasing interest – interpreting existing climate records and targeting future coring efforts requires constraints on the stability of the local flow regime. One way of inferring the subsurface flow history is using ice fabric, the preferential alignment of ice crystal C-axes in response to strain and recrystallization in the ice column. Ice divides have a characteristic fabric profile with depth, with C-axes randomly oriented in the near surface down to a single-maximum vertical fabric for basal ice undergoing pure shear (observed at both GRIP and Dome Fuji). At flank sites, strong girdle fabrics are expected, indicative of uniaxial extension at intermediate depths. Because fabric development is thought to occur at total strains greater than 0.3, divides are slow to overprint fabrics imposed by previous flow regimes. Weak birefringence in ice, integrated over typical ice thicknesses in Antarctica and Greenland, can produce measurable changes to the amplitude and phase of radar reflections as a function of ice fabric, providing a potential to discriminate between long-lived and recently formed divides. The way anisotropy manifests in radar data depends on the system configuration; monopulse, chirped and frequency-modulated continuous-wave (FMCW) systems have different sensitivity to ice fabric as a function of their frequency, chirp duration and bandwidth. In this study, we present the theoretical effects and expected signature of different ice fabrics on radar returns for commonly used radar systems. This is done with the goal of interpreting polarimetry data collected with an autonomous phase-sensitive radio-echo sounder (ApRES) at Hercules Dome during the 2018/19 austral summer. We show that measurements at Hercules Dome exhibit more variability as a function of polarization angle than expected from wave splitting alone, highlighting the challenge of separating the effects of ice fabric from anisotropic scattering off englacial reflectors.


Changes in water chambers beneath ice cauldrons on Mýrdalsjökull ice cap, southern Iceland, detected with repeated 3-D-migrated RES surveys

Eyjólfur Magnússon, Hrafnhildur Hannesdóttir, Finnur Pálsson, Joaquín M.C. Belart, Etienne Berthier

Corresponding author: Eyjólfur Magnússon

Corresponding author e-mail: eyjolfm@raunvis.hi.is

Jökulhlaups originating from beneath ice cauldrons in glaciers are a relevant natural hazard in Iceland. They have threatened people and livestock and caused damage to roads, bridges and power-supply systems as well as ruining meadows and vegetated areas. More than 20 cauldrons have been identified in Mýrdalsjökull ice cap, which covers the infamous Katla volcano. The cauldrons are formed and maintained by geothermal activity at the glacier bed although some may also be subject to volcanic activity. The ice thickness beneath them is typically 200–400 m. Water accumulates beneath the majority of the cauldrons and is typically released annually in small jökulhlaups during midsummer with peak flow usually ranging from a few cubic metres per second to a few hundred cubic metres per second, with the total volume of released water in the range 0.1–10 × 106 m3. In the past 70 years, three jökulhlaups have exceeded these limits. The last one, in 2011, destroyed the bridge over the river Múlakvísl, cutting the road connection along the south coast of Iceland for more than a week. This event provoked the idea of using low-frequency (~2 MHz) radio-echo sounding (RES) to monitor water accumulation beneath the cauldrons. Until then monitoring efforts had focused on the cauldron’s surface elevation and depth. Since 2012, between one and three RES survey profiles have been accurately repeated across most of the cauldrons in Mýrdalsjökull once or twice a year. The repeated RES-profiles (2-D-migrated) reveal clear changes beneath the cauldrons, not observed elsewhere, related to growing or diminishing water chambers. It is, however, hard to obtain a quantitative measure of water volume beneath the cauldrons based on so few survey profiles. In the attempt to quantitatively measure water volume beneath the cauldrons, dense sets of parallel RES profiles (20 m apart) have been repeatedly measured across selected cauldrons. The dense profiling allows full 3-D-migration of the data. The surveys are typically carried out in the spring and again in autumn, following depletion of the water chambers during the annual summer jökulhlaups. Digital elevation models (DEMs) of the glacier bed (20 m × 20 m cell size) are extracted from the survey data and distribution and volume of water beneath surveyed cauldrons is estimated. Results for water distribution and volume in spring are compared with the subsidence of the cauldrons during summer jökulhlaups, based on Pléiades surface DEMs obtained before and after each jökulhlaup.


An ice-penetrating radar system for widespread and rapid thickness measurements of high-mountain glaciers

Hamish Pritchard, Edward King, David Goodger

Corresponding author: Hamish Pritchard

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

Measurements of glacier thickness are very rare in most mountain ranges because radar and seismic field surveys are laborious, slow and expensive, particularly in high mountains and on debris-covered glaciers. Indirect modelling approaches to estimating glacier volumes from the local to the global scale have been developed but their dependence on other poorly known parameters such as surface mass balance, plus the lack of observations for constraint, mean that the uncertainties are large. Estimates of the ice volume in High Mountain Asia, for example, vary by about a factor of 2. Low-frequency radar systems towed over snow are routinely used to survey the thickest areas of the ice sheets but the rough surface of many high-mountain glaciers makes over-snow towing impossible. We have adapted a low-frequency ice-sheet system for airborne survey by helicopter over high, debris-covered mountain glaciers, and we report here the results of field and airborne testing. This new system is capable of transforming our knowledge of glacier thickness on the mountain-range scale, and can therefore greatly reduce uncertainty in predictions of the future summer water supply to glacierized river catchments.


Alternative strategies for synthetic aperture radar focusing of orbital radar sounding measurements

Kirk M. Scanlan, Gregor Steinbruegge, Scott Kempf, Cyril Grima, Duncan Young, Donald Blankenship

Corresponding author: Kirk M. Scanlan

Corresponding author e-mail: kirk.scanlan@utexas.edu

Coherent radar sounders in orbit around Mars (the Mars Advanced Radar for Subsurface and Ionospheric Sounding (MARSIS; 1.8, 3, 4, and 5 MHz center frequency) and SHAllow RADar (SHARAD; 20 MHz center frequency) instruments) have shown themselves to be invaluable tools in the investigation of that planet’s surface and subsurface. Furthermore, two additional sounding experiments are to be deployed on upcoming missions to the Jovian system; the 9 MHz center frequency Radar for Icy Moons Exploration (RIME) instrument on board the Jupiter Icy Moons Explorer (JUICE) and the 9 and 60 MHz center frequencyRradar for Europa Assessment and Sounding: Ocean to Near-surface (REASON) instrument on board Europa Clipper. Fundamental to the interpretation of all these data is proper signal focusing, which leverages the coherent nature of the measurements to improve the along-track resolution of the resulting radargrams. Current strategies in the focusing of MARSIS and SHARAD datasets include the separation of the data into individual Doppler bins (MARSIS) and a delay Doppler technique based on relative distances to the mid-aperture nadir surface within each aperture window (US SHARAD data analysis). While successful for MARSIS and SHARAD, these focusing strategies may not be applicable for REASON because ofthe variable pulse repetition frequency and the dramatic change in the range-to-surface during a single Europa fly-by. To this end, we propose and investigate two alternative orbital sounding SAR focusing strategies using Martian data. The first is based on the matched-filter focusing used in the analysis of HiCARS airborne radar sounding measurements. The second is a modification of the existing SHARAD delay-Doppler focusing algorithm. The matched-filter focuser allows the user to introduce a dielectric material below surface, while the modified delay-Doppler approach interpolates orbital measurements to a constant along-track spacing and makes use of spherical coordinates to accurately account for changes in platform-surface distances within the aperture. While varying the relative dielectric permittivity of the subsurface exhibits little change in the final matched filter-focused product, the modified delay-Doppler approach does appear to produce a sharper final image compared to existing data products. The modified delay-Doppler focuser may also be more suitable for the analysis of REASON data where the range to the surface may vary significantly across an aperture.


New radar-sounding investigations over the hypersaline subglacial lakes beneath Devon Ice Cap, Canadian Arctic

Anja Rutishauser, Donald D. Blankenship, Lucas Beem, Jamin Greenbaum, Duncan A. Young, Cyril Grima, Lauren N. Schwartz, Joel A. Foran, Alison Criscitiello

Corresponding author: Anja Rutishauser

Corresponding author e-mail: rutishau@ualberta.ca

From radio-echo sounding (RES) measurements and the surrounding geological context, a recent study presented the first evidence for the existence of two subglacial lakes beneath Devon Ice Cap (DIC), Canadian Arctic. The lakes are located in a cold-based region of the ice cap and are hypothesized to be hypersaline with their salinity being derived from a salt-bearing evaporite unit that likely outcrops beneath the ice. Subglacial water systems have long been of interest in the search for life beyond Earth as they provide terrestrial analogs for potentially similar ice-covered habitats on other planetary bodies across the solar system. The hypersaline nature and cold conditions of the Devon subglacial lakes makes them particularly tantalizing analogs for brines inferred to exist on Europa and Mars, and are thus compelling targets for in-situ investigations of their biogeochemical conditions. However, due to limited spatial coverage of the geophysical data used to identify the lakes, critical boundary conditions of these lakes and their surrounding subglacial environment remain unknown. Here, we present results from a targeted airborne geophysical survey (RES and laser altimetry) conducted over DIC in spring 2018. We use the RES-derived basal reflectivity and specularity content in combination with the hydraulic potential gradients to re-evaluate the evidence for the existence of the subglacial lakes, derive their full extents and investigate their surrounding subglacial hydrological conditions. Our results support the evidence for one of the subglacial lakes, for which we delineate new shorelines that reveal that the lake is more extensive than the previous dataset showed. Furthermore, the radar-derived basal properties suggest that extensive areas beneath DIC consist of wet beds, which we hypothesize are part of a brine network where water is concentrated in small ponds, thin films or saturated sediments. The second subglacial lake previously hypothesized likely forms part of this brine network rather than being a deep water body. We speculate that the characteristics of the brine network are related to, and partially controlled by the bedrock lithology and topography.


REASON for Europa: data products and algorithms

Duncan Young, Cyril Grima, Gregor Steinbrügge, Kirk Scanlan, Scott Kempf, Donald Blankenship

Corresponding author: Duncan Young

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

The Radar for Europa Assessment and Sounding: Ocean to Near-surface (REASON) is the radar instrument on NASA’s Europa Clipper mission, designed to sound the icy crust of Europa, measure the roughness and compositional properties of the surface, detect the tidal deformation of the shell and search for anomalies while investigating Europa’s ionosphere. REASON is being jointly developed by the University of Texas Institute for Geophysics (UTIG), NASA’s Jet Propulsion Laboratory (JPL) and Johns Hopkins University’s Applied Physics Laboratory (APL). REASON’s diverse data products will be generated by the REASON Science Data System, located at the PI institution at UTIG. Here we describe the science products and high-level algorithms REASON will use to satisfy its science requirements. These products include shallow VHF and full depth HF and VHF sounding radargrams supported by interferometry and clutter simulation, profiles of altimetry for isostasy, backscattered power for inferring roughness and density, the total electron content for identifying plasma anomalies, and ranging to the surface at crossovers to infer tidal deformation. We discuss our approach to using SHARAD and terrestrial data to validate our algorithms.


Geospatial simulations of airborne ice-penetrating radar surveying reveals elevation under-measurement bias for ice-sheet bed topography

Oliver Bartlett, Steven Palmer, Dustin Schroeder, Emma MacKie, Timothy Barrows, Alastair Graham

Corresponding author: Oliver Bartlett

Corresponding author e-mail: ob285@exeter.ac.uk

Airborne radio-echo sounding (RES) surveys comprise the principal method of measuring ice-sheet bed topography, bed conditions and ice-sheet thickness. Bed topography and bed conditions influence overlying ice dynamics, which in turn affects the quantity and rate at which ice is moved to lower elevations where it ablates. Numerical models that aim to predict how ice sheets will respond to climate change require bed topography as an input. Hence, accurate and widespread quantification of ice-sheet bed topography is essential for robust predictions of future ice-sheet response to climate change. Various errors and uncertainties arise from RES surveying where different facets of the ice-sheet environment scatter or attenuate the radar signal, impacting bed-measurement accuracy. In this study, we simulated RES surveys over currently exposed formerly glaciated terrain in the Arctic, in order to quantify the sensitivity of the derived bed elevation to topographic properties such as roughness, relief and land-feature orientation. Our simulations are based on the CReSIS MCoRDS instrument, as the majority of existing RES measurements in the Arctic have been acquired with this system. We present a novel approach for identifying the most probable location of a received radar return at each point along a survey flight line with respect to the geometry of the underlying topography. We then extract the elevation from this location and compare it with the nadir elevation to quantify deviations in measurements. From the findings presented here, we establish (i) the relationship between subglacial roughness, survey geometry and bed measurement deviation, (ii) the probability of the magnitude and sign of deviations for varying terrains and survey geometries and (iii) the influence of deviations on derived geometry of subglacial valleys. The last is to investigate uncertainty on flight-lines used as flux-gates for numerical modelling and mass conservation. Following this, because bed topography datasets are largely interpolated between flight lines, we aimed to quantify how uncertainty from RES surveys carries through into interpolations. We interpolated our simulated surveys to reflect previous mapping methods (kriging) and alternatives (conditional simulation) and compared them with the original topography to quantify uncertainty. Finally, we discuss the wider implications of our findings for future RES survey design and their analytical outputs.


Applying pRES-derived firn densification rates: model tuning and exploration

Elizabeth Case, Jonathan Kingslake

Corresponding author: Elizabeth Case

Corresponding author e-mail: ehc2150@columbia.edu

Firn densification models are used to constrain the age of air bubbles in ice cores and the densification contribution to altimetry measurements of ice-surface height change. As climate change leads to warmer, wetter, higher-precipitation glacial environments, we need to measure and model firn densification’s transient response. This work builds on previously presented research that showed how phase-sensitive radars (pRES) can be used to measure firn densification. To date, most densification models are tuned to static density profiles. The low logistical cost of pRES densification measurements makes fitting to the process (densification) rather than the product (density) a possibility. We investigate the impact of tuning model parameters to a densification profile compared to a density profile at the Fletcher Promontory. The best-fit activation energies change, along with the goodness of fit to density and densification, as measured in an ice core and by the radar. We also use the radar and core data to tease apart model dependence on density, stress and grain size, identifying regimes that might be important to explore in laboratory experiments. Finally, we show that if the shallow-ice approximation holds at the measurement location, we can likely estimate average accumulation from the radar measurements.


What topography will come to light as Swiss glaciers melt away? Surveying the glacier bed with dual-polarization helicopter-borne GPR

Melchior Grab, Lisbeth Langhammer, Andreas Bauder, Lasse Rabenstein, Lino Schmid, Hansruedi Maurer

Corresponding author: Melchior Grab

Corresponding author e-mail: melchior.grab@erdw.ethz.ch

Alpine glaciers have lost substantial parts of their ice volumes during recent decades, and all signs indicate that this process will continue unabated in the future. In Switzerland, and likewise in other mountainous or subpolar regions, glacier melting will have a direct ecologic and economic impact. Consequences are for example expected for the supply of Swiss electricity, which is in major part produced by hydropower, but also for tourism, which is an important economic factor in Alpine regions, or with regard to natural hazards, which will change their pattern in the course of deglaciation. A good knowledge of the bedrock topography is key for developing strategies to deal with risks and new opportunities arising from the ongoing glacier melt. During the past years, we have developed AIR-ETH (Airborne Ice Radar ETH Zurich), a helicopter-borne GPR system from commercially available components, and the data-processing software package GPRglaz for surveying the ice thickness and bedrock topography of Alpine glaciers. During earlier investigations we found that the quality of radar-wave reflections from the bed of valley glaciers strongly depends on the antenna orientation. In order to measure permanently with the ideal antenna orientation during helicopter-based surveying, our GPR platform is designed as a dual-polarization system. It consists of two orthogonal receiver–transmitter antenna pairs which measure in alternating fashion. The data is processed with our in-house Matlab®-based software package. It enables us to effectively filter out the strong artefacts appearing due to the interference of GPR waves with the helicopter. Directivity effects of the antennas can be suppressed by simply summing the signals recorded with the two orthogonal antenna pairs. For migrating the data, either a standard Kirchhoff time migration algorithm can be used or a computationally expensive but more effective reverse time migration algorithm. Up to today we have built a database of around 2500 km of GPR profiles, providing bedrock information of most ice-covered areas in the Swiss Alps. We will present examples to demonstrate the performance of our GPR system. Our goal is to estimate the total ice volume of today’s Swiss glaciers. Therefore, we combine the bedrock information given on the sparse grid of GPR profiles with glaciological modeling techniques, in order to provide continuous ice-thickness maps and estimate the ice volume.


Processing of ground-based ice-penetrating radar data: hints and tricks for difficult data

Edward King

Corresponding author: Edward King

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

In many situations in glaciology useful results can be obtained from ice-penetrating radar systems with straightforward acquisition parameters and minimal processing of the raw data. However there are some circumstances where the desired interpretive objective cannot be reached without more complex processing. Many glaciologists have received little or no training in geophysical data processing and may be unaware of some of the tools available to them that may rescue apparently unusable data. In this paper I will describe four surveys that required extra processing effort before interpretable profiles could be obtained. The first comes from a survey of a deep, narrow valley glacier where the raw profiles are dominated by high-amplitude linear reflections that are radio-wave echoes from the steep sidewalls of the valley. These echoes obscure the reflection from the bed of the glacier. The second shows that data apparently ruined by an error in acquisition setup can still be used. The third uses a data editing trick to make weak internal reflections near the bed of an ice sheet interpretable. The fourth demonstrates the elimination of cultural noise and distortion by topography to show the internal structure of a drumlin.


Is your oversnow radar tough enough? Practical design considerations for system survivability

Edward King

Corresponding author: Edward King

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

Ground-based impulse radar systems remain an important tool in the radioglaciologist’s toolbox where the survey targets are highly detailed bed topography or the tracking of isochronal reflectors through complex internal structure. It is technically relatively easy to assemble such a radar system from mainly commercial off-the-shelf components. However towing a system for thousands of kilometres over terrains such as sastrugi or debris-laden bare ice places extreme demands on the physical robustness of the equipment. The low-frequency impulse radar deployed by the British Antarctic Survey, known as DELORES (DEep LOok Radio Echo Sounder) was first put into the field in Greenland and Antarctica in 2005 and has evolved through several iterations during a total of 25 Antarctic field campaigns and 4 Arctic test periods. During this period many of the central components of the system have remained the same but the way those components are packaged and protected has evolved in order to secure reliable operation while being towed on small sledges behind snowmobiles. In this paper I will illustrate the progress of the design, show what has worked to increase reliability and what apparently good ideas did not survive contact with Antarctic fieldwork.


Development of a subglacial lake monitored with radio-echo sounding: case study from the Eastern Skaftá Cauldron in the Vatnajökull ice cap, Iceland

Eyjólfur Magnússon, Finnur Pálsson, Magnús T. Gudmundsson, Þórdís Högnadóttir, Cristian Rossi, Thorsteinn Thorsteinsson, Erik Sturkell

Corresponding author: Eyjólfur Magnússon

Corresponding author e-mail: eyjolfm@raunvis.hi.is

We present a 5-year record of repeated radio-echo sounding (RES) on a profile grid (200–400 m between profiles) surveyed over the eastern Skaftá cauldron (ESC). ESC is an ice cauldron produced and maintained by powerful geothermal activity (~1 GW) at the glacier bed. Beneath the cauldron and 200–400 m of ice, water accumulates in a lake and is regularly released in jökulhlaups into the river Skaftá, the latest occurring in 2015 and 2018 (maximum discharge ~3000 m3 s–1 and ~2000 m3 s–1, respectively). The record is from 2014–18 and consists of annual measurements obtained in June each year. Comparison of the repeated RES profiles (2-D-migrated) reveals the margin of the lake at different times and enables classification of traced reflections into lake and bedrock measurements. The bedrock measurements include data obtained with the lake close to its minimum size in 2016 and 2017 (<~1 km2 compared to 4.0 km2 in 2015), which allows creation of a fairly accurate digital elevation model (DEM) of the glacier/lake bed, further constrained by two borehole measurements of the lake bed elevation at the cauldron’s centre. The traced lake reflections and comparison with the bedrock DEM enables creation of a lake-thickness map and an estimate of lake volume for each survey. The lake-thickness maps and volumes in June 2015 and 2018 are compared with the surface-lowering pattern and water volumes drained in the jökulhlaups in October 2015 and August 2018. The drained water volume was derived by integrating the surface lowering during the jökulhlaups and adding the estimated volume of crevasses formed during the events. The lowering in 2015 was obtained from TanDEM-X DEMs of 23 September and 10 October, shortly before and after the jökulhlaup. The lowering in 2018 was derived from dense set of airborne altimetry profiles acquired on 9 August, a few days after the jökulhlaup, compared with a DEM in June 2018 (ArcticDEM in July 2017 corrected with dense GNSS profiles in June 2018). The lake volume estimate from the RES data is 240 × 106 m3 in June 2015 but 320 ± 20 × 106 m3 drained from the cauldron in October. In June 2018 a relatively dense RES profile grid (~200 m between profiles) reveals a lake volume of 180 × 106 m3 while 210 ± 30 × 06 m3 drained from the cauldron in August. This comparison demonstrates the applicability of our survey approach to monitor the water accumulation in the lake and thus better constrain potential hazards in jökulhlaups.


Recovery of englacial stratigraphy across Pine Island Glacier: proof of concept for analysing the internal architecture of West Antarctica

Julien A.B. Bodart, Robert G. Bingham, David W. Ashmore, Nanna B. Karlsson, David G. Vaughan, Hugh F.J. Corr

Corresponding author: Julien A.B. Bodart

Corresponding author e-mail: julien.bodart@ed.ac.uk

Concerns over the potential collapse and future contribution of the West Antarctic Ice Sheet (WAIS) to sea-level rise have resulted in significant scientific interest over the last three decades. Central to this issue are dynamic changes to large glacier catchments draining the WAIS, such as the thinning and mass loss observed over Pine Island Glacier (PIG) in recent years. Whilst much work has focused on assessing current changes in the mass balance of PIG, few studies have utilized the geometry of radar-sounded internal layers to reconstruct past changes to the catchment. A key challenge when analysing radio-echo sounding data is the identification and interpretation of the internal layers themselves, which are influenced by choices made in the processing flows or artefacts produced during the data-acquisition phase. Here, we use a novel approach to radar processing that utilizes two data-acquisition modes from the 2004/05 PASIN airborne survey to assess the internal stratigraphy of the PIG catchment. From this, we construct a tentative age–depth relationship based on the tracing of three consistent horizons at varying depths within the ice column. We tie in these horizons with further internal layers traced across PIG’s main trunk and tributaries from GPR data acquired as part of the 2013/14 iSTAR science programme. Lastly, we produce three-dimensional elevation maps of layer stratigraphy to assess the englacial conditions of the glacier catchment. Our results act as a proof of concept that wider layer tracing is possible across a significant proportion of the WAIS.


Signature of NEGIS shear margins in radar stratigraphy

Daniela Jansen, Steven Franke, Niklas Neckel, Tobias Binder, Paul Bons, Olaf Eisen, Veit Helm, Heinrich Miller, John Paden

Corresponding author: Daniela Jansen

Corresponding author e-mail: daniela.jansen@awi.de

The North East Greenland Ice Stream (NEGIS) is an essential part of the Greenland mass balance, reaching up to the central divide of the ice sheet. Its shape is dominated by very distinct shear margins that appear not to be linked to bedrock topography in the upper part of the ice stream. To understand the effective mass transport within an ice stream it is necessary to investigate the nature of the shear margins, in which very localized deformation decouples the inner ice stream from the slower-flowing surrounding ice sheet. We present results from an airborne radar survey with the Alfred Wegener Institute Ultra Wide Band Radar system in the vicinity of the EGRIP ice core location, focusing on the signatures of the shear margins in the radar data. In regions of localized shear, the internal reflections show disturbances in the form of steep undulations, which are intensified with ongoing shear. We show the evolution of these characteristic signatures over the survey area. Our data covers the main trunk and the shear margins over 200 km along the ice stream, in which the flow velocity in the center increases from 12 m a–1 to 75 m a–1. We analyze the change in the shape of the internal reflections in the shear zones in combination with a strain rate field calculated from high-resolution flow velocities derived from TerraSAR-X SAR interferometry. In the vicinity of the EGRIP drilling site, 240 km downstream of the divide, the NEGIS widens and the shear margin in the northwest steps 6 km outward, which is visible in optical satellite imagery as well as in the satellite-derived velocity field. A special focus of our analysis lies on the step in the shear margin whose clearly visible radar-backscatter signatures indicate a newly forming vertical disturbance caused by a draw-down of the layers.


Radar architecture of the Weddell Sea sector of West Antarctica

David Ashmore, Robert Bingham, Neil Ross, Martin Siegert, Tom Jordan, Hugh Corr, Douglas Mair

Corresponding author: David Ashmore

Corresponding author e-mail: ashmore@liverpool.ac.uk

The structural architecture of the Antarctic Ice Sheet, as derived from the geometry of isochronal radar reflectors, records the cumulative effects of ice-sheet accumulation, basal melting and ice-flow dynamics. Measuring this structure informs us on key parameters of the ice-sheet system that are often ignored or heavily parameterized in ice-sheet models. Hence, although internal layers are widely recorded, a lack of reliable classification and consistent measurement between surveys has meant their use in understanding glaciological processes and change has been restricted to regional investigations. In this presentation we investigate the feasibility of tracking a distinctive package of three englacial reflectors from slow-flowing regions across ice streams. Through 3-D visualization of the UK IMAFI PASIN (2010/11) survey of the Institute/Möller ice streams system, we show the extent of these reflectors throughout a large ice-flow catchment. In some areas, it is possible to track the layers continuously from the Weddell–Ross divide over 450 km to within 50 km of the grounding line, and across the ice-stream trunk. We discuss and interpret the resultant layer geometry in terms of their normalized depth, and modern-day catchment-scale patterns of accumulation and ice velocity. We then combine these data with published work regarding englacial layering elsewhere in West Antarctica, and describe links to PASIN surveys of Pine Island Glacier (acquired in 2004/05) and Rutford Ice Stream, Carlson Inlet and Evans Ice Stream (in 2006/07). We show that these highly traceable layers may be widespread throughout central West Antarctica and act as useful anchors for future studies of ice-sheet stratigraphy elsewhere in West and, potentially, East Antarctica.


Surface and basal melting of the Totten Glacier Ice Shelf, East Antarctica

Jamin Greenbaum, Dustin Schroeder, Cyril Grima, Noel Gourmelen, Christine Dow, Feras Habbal, Jason Roberts, Roland Warner, David Gwyther

Corresponding author: Jamin Greenbaum

Corresponding author e-mail: jamin@utexas.edu

Totten Glacier is the largest outlet of ice to the ocean from the Aurora Subglacial Basin (ASB) and is the most rapidly thinning glacier in East Antarctica. Ice-shelf basal melting is driven by warm modified circumpolar deep water (mCDW) that enters the ice-shelf cavity through a system of seafloor troughs connecting to a reservoir of mCDW present on the nearby continental shelf. Totten Glacier is also susceptible to seasonal surface melting owing to its relatively northern latitude. Using airborne ice-penetrating radar (IPR) data we show that the ocean actively melts large channels into the ice-shelf base that grow to over 2 km wide and 350 m deep with steep walls and flat terraces characteristic of rapid melt. We estimate basal melt rate along each flight line over the ice shelf from IPR basal-reflectivity data using a novel approach that corrects for scattering losses due to ice-shelf surface and basal roughness; the new estimates compare favorably to independent satellite-derived estimates but offer higher along-track resolution. We also use the IPR data to show that the near surface of the ice shelf undergoes widespread melting in warm years and that the resulting melt is concentrated in the surface expression of basal melt channels. The natural vulnerability of the Totten Glacier Ice Shelf to surface and basal melting implied here is concerning given recent numerical modeling results indicating that ocean melting and surface melt-induced hydrofracture and ice-cliff failure could cause substantial retreat into the ASB and corresponding sea-level rise.


ImpDAR: an open-source impulse radar processor in Python

Joshua Driscol, David A. Lilien, Benjamin H. Hills, Knut Christianson, Robert W. Jacobel

Corresponding author: David A. Lilien

Corresponding author e-mail: joshuadr@uw.edu

Despite the widespread use of radio-echo sounding data in glaciology, there is no community standard for processing software. This complicates intercomparison of radiometric information between radar systems or even between different datasets collected using the same radar system processed by different operators. An enormous amount of new radar data is being collected, so now more than ever it is important to have dependable, fast, and compatible processing flows for these data. Here, we present ImpDAR, a cross-platform, open-source impulse radar processor and interpreter written in Python. ImpDAR is designed to be a universal processor that is capable of reading data from the common commercial ground-penetrating radars (i.e. GSSI, PulseEkko, and MALÅ) as well as from some of the custom-built high-frequency systems. The core processing algorithms are based on the St Olaf radar processor, which required a Matlab license and contained deprecated, legacy code. Other previously available tools also have significant drawbacks: commercial options are expensive, brand-specific, platform-specific, have limited capabilities, and some lack batch-processing utilities. The processing flow included in ImpDAR is designed to be versatile, permitting the user to choose between multiple filters, geometric corrections, migration routines and display options. It performs several critical, customizable processing steps, such as bandpass and horizontal filtering, time correction for antenna spacing, trace geolocation, migration and elevation correction, among others. After processing data, ImpDAR’s interpreter makes it easy to plot the data and to digitize internal reflecting horizons. There are several additional plotting functionalities to visualize geometry, frequency content and power of reflected waveforms. As an initial test of ImpDAR, we present interpreted profiles from 500 MHz snow radar used at South Cascade Glacier, Washington, USA, and 3 MHz radar deployed at the South Pole subglacial lake. We hope that ImpDAR can become a common resource for impulse radar processing and interpretation in the glaciological community, and therefore request community input to improve this new tool.


Glacier Thickness Database (GlaThiDa) version 3.0

Ethan Welty, Francisco Navarro, Johannes Fürst, Isabelle Gärtner-Roer, Kathrin Naegeli, Johannes Landmann, Matthias Huss, Thomas Knecht, Horst Machguth

Corresponding author: Francisco Navarro

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

GlaThiDa is an internationally collected, standardized dataset of glacier thickness from in-situ and remotely sensed observations, based on data submissions, literature review and airborne data from NASA’s Operation IceBridge. GlaThiDa is a contribution to the Working Group on Glacier Ice Thickness Estimation formed under the auspices of the International Association of Cryospheric Sciences (IACS). The database is hosted by the World Glacier Monitoring Service (WGMS). GlaThiDa is structured in three data tables of different levels of detail, which are linked through a GlaThiDa_ID that is unique for a given glacier and survey. For one glacier or ice cap there can be multiple entries, each referring to a particular survey (e.g. surveys at different dates, or using different radars). The first table (T) is the overview table containing information on the location and area of the glacier, mean and maximum thickness from interpolated observations including accuracies, the survey method and related information, as well as the investigator and the source of the data. The second table (TT) includes ice-thickness data (mean and/or max) averaged over surface-elevation bands by given lower and upper boundaries from ice-thickness maps or digital elevation models. The third table (TTT) contains point data including a point ID, related coordinates, the point elevation and the thickness value. Table TTT reflects the original observations, which are more or less extensive depending on the survey method and the level of detail of the data description in the literature. GlaThiDa was first released in 2014 (version 1.0), and version 2.1 followed in 2016. Version 3.0 has just been launched in 2019. In addition to several technical improvements, nearly 3600 ice-thickness surveys have been added to the previous version, making a total of 5181 ice-thickness surveys. GlaThiDa has great potential as a reference dataset for calibrating and validating regional and global glacier-volume estimates. This is critical, as a comparison between GlaThiDa 1.0 observational ice thicknesses with results from area- and slope-dependent approaches revealed large deviations, with all estimation approaches showing a tendency to overestimation. For glaciers, the median relative absolute deviation was around 30% when analysing the different estimation approaches. The large amount of additional in-situ data made available by GlaThiDa 3.0 will hopefully show a different picture.


Interpreting radar bed-echo power from active subglacial lakes on lower Mercer and Whillans ice streams, West Antarctica

Matthew Siegfried, Dustin Schroeder

Corresponding author: Matthew Siegfried

Corresponding author e-mail: siegfried@mines.edu

Lakes beneath the Antarctic ice sheet have been mapped across the continent primarily through two different observational methods. Airborne radio-echo sounding (RES) has been used to detect hydraulically flat, radiometrically bright reflectors at the ice–bed interface, which have been inferred to be locations of ponded subglacial water. More recently, repeat height observations from airborne and satellite altimetry have been used to identified clusters of spatially coherent surface-height displacements, which are interpreted as the surface expression of subglacial water movement in to and out of individual ‘active’ subglacial lakes. Where both RES and satellite altimetry data are available, the datasets provide conflicting information about the ice–bed interface: with a single exception: active subglacial lakes do not display the typical flat, bright signature of subglacial water in RES data, suggesting that our physical understanding of Antarctic subglacial hydrology, especially beneath fast-moving ice streams and outlet glaciers, remains incomplete. Here, we use data from an airborne RES campaign that surveyed a well characterized set of active subglacial lakes on lower Mercer and Whillans ice streams to explore the inconsistency between these two observational techniques. In particular, we test hypotheses of increased roughness, scattering and englacial attenuation due to the presence of an active subglacial lake system that could generate the observed RES profiles in these locations.


The origin of ice-shelf channels revisited

Reinhard Drews, Daniela Jansen, Steven Franke, Frank Pattyn, Olaf Eisen

Corresponding author: Reinhard Drews

Corresponding author e-mail: reinhard.drews@uni-tuebingen.de

Antarctic ice shelves often contain narrow, curvilinear tracts of thin ice, termed ice-shelf channels, that impact ice-shelf stability. Their surface depressions appear prominently in satellite imagery and form an interesting morphology of unknown origin, the more so because the processes leading to ice-shelf channel formation are unclear. Here we investigate the origin of ice-shelf channels at the Roi Baudouin Ice Shelf, which have previously been attributed to ice overriding sediment ridges formed by long-term deposition in subglacial water conduits. However, due to a limited radar dataset at the time, the shape and upstream extent of the basal obstacles was unclear, making it difficult to pinpoint the exact type of the subglacial landform. In 2019, we revisted this location with an improved airborne ultra-wideband radar and collected a number of across-flow profiles upstream of the grounding line. Consistent with previous suggestions, we find that the basal obstacles shrink in size with increasing distance from the grounding line until they are invisible about 10 km upstream. However, the change in size is gradual in the first 8 km, and then very abrupt in the last 2 km. Variations in basal reflectivity indicate the existence of patches of subglacial water in all profile lines. Once fully evaluated, this rich dataset has the potential to classify the basal landform and hence shed light on the origin of ice-shelf channels and their impact on ice-shelf stability.


Automatic detection of snow layers in radar images

Yuchen Wang, David Crandall, Mingze Xu, Lora Koenig, John Paden, Geoffrey Fox

Corresponding author: Yuchen Wang

Corresponding author e-mail: wang617@iu.edu

Internal layer depths can be used to estimate past accumulation rates and these same depths and their related slopes can provide additional constraints on ice flow that can be used to test and tune ice-sheet models. While ground-penetrating radar is able to collect observations of the layer structure of snow and ice, the process of manually labeling these observations with layer boundaries is slow and laborious. Recent work has developed automatic techniques for finding ice–bed boundaries, but finding internal boundaries is much more challenging because the number of layers is unknown and the layers can disappear, reappear, merge and split. We investigate, develop and apply approaches for automatic internal layer finding using techniques from probabilistic graphical models and deep machine learning. These techniques are state-of-the-art for recognition of consumer images in computer vision, but applying them for layer finding is challenging because of the noise in echogram data and the fine-grained nature of this recognition task. We use 2011 and 2012 Greenland Snow Radar data from Operation IceBridge with human-labeled data as ground truth to evaluate the proposed algorithms.


Assessing the affect of subaerial volcanism on englacial attenuation in West Antarctica

Enrica Quartini, Duncan Young, Manasa Sudunagunta, Dustin Schroeder, Donald Blankenship

Corresponding author: Duncan Young

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

The West Antarctic Ice Sheet (WAIS) lies in a distributed zone of active volcanism. Significant ash deposits related to subaerial volcanism are found in the WAIS near the Executive Committee Range. In this presentation we examine the interaction between volcanism in Marie Byrd Land, West Antarctica and the englacial structure of the WAIS by first mapping out the distribution of englacial ash layers and then comparing this distribution to empirically derived attenuation rates, following on the work of Schroeder et al. (2016). We use data from the 2004/05 AGASEA aerogeophysical campagin and the 2012–14 GIMBLE aerogeophysical campaign, both of which used data from the 60 MHz HiCARS radar sounder. We test the hypothesis that englacial ash layers have a minor impact on englacial attenuation rates.


Interpretation of radar data informed from 3-D modelling

Gwendolyn J.-M.C. Leysinger Vieli, Richard C.A. Hindmarsh

Corresponding author: Gwendolyn J.-M.C. Leysinger Vieli

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

Advances in radar systems and radar-processing technology has led to radio-echo sounding (RES) data of ice-sheets that now allow for highly detailed observations of englacial layer structures even near the ice-sheet base. This has led to recent observations of plume-like internal ice-layer structures, rising from the ice-sheet base to up to half of the ice thickness. The origins of these plumes are not yet fully understood, with multiple plausible explanations existing, such as: basal freeze-on; variability in basal slipperiness; changes in near-basal ice rheology. In radar profiles, these fascinating three-dimensional (3-D) plumes are highly complex layer structures, which contain a wealth of information waiting to be unlocked. However, in our interpretation of radar profiles we often lack awareness of their inherently 3-D nature. Not only do radar images need to be viewed with their obliquity to ice flow in mind, but also with the knowledge that upstream conditions can significantly influence the layer structure downstream. This is especially true for the mechanism of basal freeze-on with internal layers being deformed downstream by the advection of freeze-on bodies. Here we specifically explore plume-like layer structures originating from and influenced by changes in basal conditions with the 3-D numerical model BASISM. We display the modeled isochrones along transects cutting the model domain at different locations and angles. These isochrones solely resulting from changes in transect location and/or angle can result in very different and complex structures. The modelling of isochrones together with the observation of layer structures will support future interpretations of radar profiles and possibly resolve the mechanism responsible in generating any given plume.


Towards an autonomous, mobile radar platform for ground based radar surveys including SAR focusing and polarimetry

Mohammedreza Ershadi, Reinhard Drews, Andreas Zell, Keith W. Nicholls, Todd A. Ehlers

Corresponding author: Reinhard Drews

Corresponding author e-mail: reinhard.drews@uni-tuebingen.de

Significant technical improvement in radioglaciology over the last 10 years has led to a plethora of scientific discoveries on ice sheets such as the extensive stratigraphically disturbed basal layer, or the evolution of the subglacial hydrology. It now becomes more evident that sub-kilometer-scale features such as subglacial water conduits, basal terraces or ice-shelf channeling on ice shelves play an important role in defining the dynamics of the entire ice sheet. However, despite the improvements in radar technology, the way radar profiles are collected in the field has not greatly evolved and is often done on foot or with snowmobiles. While the latter is sufficient for standard common offset profiling, it is often too time-consuming and imprecise for more advanced ground-based survey modes such as radar polarimetry, SAR focusing or 3-D migration. We aim to overcome this limitation by towing an advanced phase-sensitive radar system (ApRES) with an autonomous and mobile robot. Scientific targets for such a system are manifold, and we first focus on collecting data suitable for SAR focusing and radar polarimetry. The former requires autonomously collecting data in start–stop mode along pre-defined tracks, and the latter includes integration of a multiple inputs/multiple outputs (MIMO) system acquiring radar returns in co- and cross-polarization simultaneously. Here, we present first results on the coupled system design of wheeled robot and polarimetric ApRES, including robot simulations, GPS integration and radar forward modeling. First scientific deployment of the system will be near an Antarctic grounding line gearing towards resolving the internal stratigraphy across ice-shelf channels and basal terraces, the area-wide detection of basal melt rates and detection of subglacial water conduits.


Ice-sheet surface processes and properties from NASA IceBridge shallow radar sounding

Brooke Medley

Corresponding author: Brooke Medley

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

NASA’s Operation IceBridge began in 2009 with the goal of bridging the laser-altimetry gap between ICESat (2003–09) and ICESat-2 (2018–present). The polar altimetry community now faces the challenging task of linking two satellite missions with nearly a decade of separation, a period over which the ice sheets have undergone substantial change. Because surface elevation is the function of several processes (i.e. variations in snow accumulation, firn densification, ice dynamics and solid Earth deformation), additional constraints are necessary to untangle their independent contributions. Thus, IceBridge hosts an entire suite of instruments largely designed to improve our understanding of the ice-sheet system as a whole, including radar systems developed by the Center for Remote Sensing of Ice Sheets (CReSIS). For this presentation, we focus on the shallow systems that image the near-surface internal stratigraphy down to depths of ~5–100 m. We provide an overview of the CReSIS ‘snow’ radar system and then discuss several recent advancements to our understanding of the surface and near-surface processes and properties. The first results from the OIB shallow radar focused on improved spatiotemporal measurements of snow-accumulation rates for refinement of atmospheric models and novel measurements of firn-compaction rates for firn-model validation. However, the focus has turned towards more nuanced methods for extraction of small-scale variability in snow accumulation at the ice-sheet scale, firn-densification model calibration, and even more experimental methods to determine the absorption and scattering properties of the firn column by fitting a physically based model to the waveforms for retrieval of firn density over the Greenland Ice Sheet. After an overview of these developments, we provide insight into how these measurements have improved or will improve interpretations of ICESat/ICESat-2 elevation change. Large volumes of the IceBridge shallow radar data have yet to be analyzed, suggesting that significant discoveries with these systems remain to be made.


Estimation of ice-crystal orientation fabric from fully polarimetric ice-sounding radar data

Jørgen Dall

Corresponding author: Jørgen Dall

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

The rheology of an ice sheet depends on the crystal orientation fabric (COF). COF maps can help understanding and modelling of the dynamics of an ice sheet. The COF also provides valuable information on past ice deformation, as the COF develops as a result of such a deformation. An ice crystal is birefringent, i.e. its index of refraction depends on the polarization of the electromagnetic field with respect to the crystal orientation. This enables radar-based COF mapping. The orientation distribution of crystals in ice cores has been studied for decades, and COFs have also been studied with radars – in particular ground-based, single-polarization radars operated with adjustable antenna orientation. More recently the degree of crystal alignment versus depth has been measured with a fully polarimetric, airborne, P-band radar using a technique based on the complex coherence between two simultaneously acquired, orthogonal, co-polarized data sets. In that study the polarizations were selected such as to simplify the birefringence estimation, i.e. the electrical field vectors were parallel/orthogonal to the vertical girdle fabric at the NEEM ice core drilling site in Greenland. In this presentation the polarimetric coherence technique is further developed to also estimate the COF orientation. When fully polarimetric data are available, any transmit and receive polarizations can be synthesized. The COF orientation is estimated by finding the polarization pair for which the extreme values of the coherence phase are obtained. For a simple COF scenario, it is sufficient to synthesize linear, orthogonal, co-polarized data. The proposed technique has been tested with modelled and measured data. The modelled data assume the ice to have two slightly different indices of refraction in two orthogonal directions in the horizontal plane. The measured data were acquired at NEEM with the polarimetric airborne radar ice sounder POLARIS, a P-band radar developed for ESA by the Technical University of Denmark. The results obtained with the modelled data are in good agreement with those of the measured data. The results are validated against ice-core analyses according to which the axis of symmetry of the index ellipsoid are found to be aligned with the ice divide as compared to a vertical girdle fabric parallel to the ice divide. Apart from this validation the focus of the presentation is on the COF estimation – not the glaciological interpretation of the COF.


A comparison of multiple radio-wave attenuation methods applied to high-frequency common-offset radar surveys of the Northeast Greenland Ice Stream

Benjamin Hills, Knut Christianson, Nicholas Holschuh

Corresponding author: Benjamin Hills

Corresponding author e-mail: benjaminhhills@gmail.com

Power interpretations of radio-echo sounding (RES) data are becoming more common in glaciological applications. Bed-echo power is used as a proxy for basal properties (particularly the presence of water) and/or bed roughness, and power losses are used to infer properties of the ice column, such as ice temperature and chemistry. Any interpretation of echo power has an implicit assumption, or underlying calculation, of attenuative losses. While prior studies use either paired thermodynamic and conductivity models or the data themselves to calculate attenuation rates, there is no standard method to do so, and up to now there has been no robust comparison between methods. Here, we use ground-based RES data from the Northeast Greenland Ice Stream to assess the two groups of the methods for calculating attenuation rates commonly used on ice: (1) methods that infer depth-averaged attenuation from a linear regression of echo power for a single reflector as a function of ice thickness; and (2) those that use the diminution of reflector power as a function of depth, using multiple reflectors within one radar trace. Using the multiple-reflector method, we calculate attenuation rates to be approximately 5 dB km–1 higher within the ice-stream shear margins than in neighboring regions of low shear, presumably corresponding to the warmer ice there. However, we find large discrepancies between methods, with some methods yielding non-physical negative attenuation rates. Hence, there is high sensitivity to the inherent assumptions for each method. For this particular dataset, methods that use the internal reflectors (both multiple-reflector and single-reflector methods) achieve attenuation rates that are closest to expected values.


A numerical toolbox to guide radar and seismic field campaign planning

Steven Bernsen, Christopher Gerbi, Seth Campbell, Senthil Vel, Knut Christianson

Corresponding author: Christopher Gerbi

Corresponding author e-mail: steven.bernsen@maine.edu

Describing the three-dimensional thermal and structural architecture of ice streams, ice sheets and other glacial features is critical to understanding the dynamics of those regions, as well as making predictions of how the cryosphere will evolve. For example, in areas of high strain (e.g. a lateral shear margin), the development of preferred crystal orientation fabric or the generation of heat from viscous dissipation may lead to non-linear effects in flow dynamics. Resolving these features helps us to understand the response of the glacial system to changes in mass distribution or longitudinal stresses driven by variability in the ice-shelf front. This project describes the current state of development of a computational framework to calculate the effect of the three-dimensional distribution of ice boundaries, temperature, water content, fabric and other englacial parameters on radar and seismic signals. The final product will be a stand-alone modeling toolbox that will (a) help researchers improve their likelihood of a successful field campaign through testing the impact of the potential distribution of features on geophysical data and (b) provide a theoretical framework that can be used for parameter estimation via inverse modeling. Site-specific questions this toolbox can address include: Can we distinguish changes in the temperature structure or preferred crystal orientation fabric via seismic and/or radar methods? What data acquisition plan minimizes interference from the ice–bed interface? The major components of a radar and seismic modeling toolbox are: (1) to provide a cross-platform open-source toolbox with an intuitive but robust interface for PIs at all levels of computational experience, (2) to have the capability to generate full waveform solutions for isotropic, anisotropic and polythermal ice, (3) to be scalable to problems that range from small (<1 km2) to larger, regional modeling with the ability to implement high-performance computing. A successful toolbox requires input from the broad community that engages in radar and seismic methods to examine the englacial structure and basal properties in glaciers and ice sheets.


Debris thickness variability and internal structure of an Alaskan debris-covered glacier from sled-borne ground-penetrating radar

Tyler Meng, Eric Petersen, John Holt, Chris Larsen

Corresponding author: Tyler Meng

Corresponding author e-mail: tmeng@email.arizona.edu

Debris-covered glaciers (DCGs) are an important part of the glacial–periglacial continuum. Consisting of massive glacial ice overlain by sediments sourced from supraglacial rockfall and ablation lag, observations indicate that DCG morphology, structur, and response to changing climates differ from clean-ice glaciers. DCGs have been documented in regions of high topographic relief on both Earth and Mars. It is hypothesized that the supraglacial debris morphology and internal debris bands record glacial and interglacial periods at varying sensitivities, which can aide in validating models of both terrestrial and Martian climate. DCGs also have implications for water resources as part of the total hydrological budget over time, and they may present hazards to alpine communities as annually integrated temperatures increase. However, DCG morphological and structural evolution in response to changing climates is poorly understood due to a paucity of full-process DCG models and the challenge of imaging bed topography and internal structure of DCGs for model input and comparison. We present new sled-borne GPR data on a DCG in the Wrangell Mountains of Alaska, USA, and the results from two field surveys to test a method for acquiring data while the DCG is snow-covered to improve antenna coupling, penetration and coverage. GPR surveys were conducted on different portions of a DCG at frequencies of 50, 100 and 200 MHz over the span of 12 combined field days in 2018 and 2019. While this method presents challenges of its own, its ability to acquire hundreds of meters of GPR lines per day is unprecedented in terms of efficiency of GPR surveys on DCGs. The resulting dataset allows for the characterization of supraglacial debris thickness variability, the imaging of basal topography down to 40+ m depth, and internal reflections in the toe and trunk of the glacial indicate the presence of englacial debris bands. This GPR method greatly improves our understanding of the Alaskan DCG’s internal structure in relation to its surface morphology and velocity field, which we have mapped with repeat airborne lidar and photogrammetry. Comparing these results with studies of other DCGs on Earth and Mars will allow for a more comprehensive understanding of their response to changing climates with interdisciplinary implications for cryosphere and climate science, hazard assessment and human-resource utilization.


Planetary analogy and radioglacology

Cyril Grima

Corresponding author: Cyril Grima

Corresponding author e-mail: cyril.grima@gmail.com

The understanding of a geophysical environment is limited by the physical boundaries of the natural setting where it is investigated. The behavior, evolution, fate and interactions of its various components are better apprehended and predicted when they have been observed through the widest range of physical conditions. Cryospheres are the most prevalent planetary environments in our solar system. Icy regions and worlds are widespread, and their physical manifestations demonstrate an outstanding variety of physical conditions and settings. The Earth’s cryosphere provides a diversified sample of those processes, and complements what is observed on other bodies, e.g. subglacial oceans, tidal-driven tectonism, water vapor plumes, ice convection, cryovolcanism, possible ice subduction. The prevalence of icy worlds and the growing interest of science agencies in their exploration will combine to make the cryospheres some of the most explored planetary environments in the next few decades. In the coming race to shape a holistic understanding of a cryosphere, radar investigations have a place of choice. But comparing and concatenating worlds in a wider picture should be undertaken with a thoughtful methodology. A key method for comparing observations obtained in various planetary settings and across different physical conditions is analogical reasoning. We recall the fundamental pathways of science inference and highlight how planetary analogs can be inserted in this scheme to augment the science. We outline fundamental misconceptions about the use of analogy in science and provide guidelines for its proper application. Then we illustrate cases where radioglaciology has been used, and could be used, to investigate planetary cryospheres through comparative analogy of various worlds.


Freezing or melting? New insights into the thermodynamic and glaciological setting of the South Pole subglacial lake from recent ice-penetrating radar surveys

Benjamin Hills, Knut Christianson, Nicholas Holschuh, T. J. Fudge, Eric Steig

Corresponding author: Benjamin Hills

Corresponding author e-mail: benjaminhhills@gmail.com

Several hundred subglacial lakes have been identified beneath the Antarctic ice sheet using ice-penetrating radar. The majority of these putative lakes are small water bodies under thick ice in East Antarctica. Most detailed studies have, however, focused on active lakes under West Antarctic ice streams or especially large lakes in East Antarctica. Thus, our understanding of the many smaller lakes remains quite limited; in particular, we lack knowledge of the interaction between these lakes and the adjacent ice, the glacial substrate and the larger subglacial hydraulic system. Here, we present the most extensive geophysical exploration to date of a small subglacial lake near the geographic South Pole. Prior to this study, both airborne radar and seismic reflections had been used to independently confirm the existence of this lake. However, the thermodynamic setting and therefore the ultimate fate of the lake are debated, with some authors arguing that the bed should be cold and the lake is now freezing over time. We use GPS and an ice-temperature model to test this hypothesis and to constrain the glaciological setting in which South Pole Lake exists. In the 2018/19 austral summer, we conducted extensive high-frequency (3 MHz) ice-penetrating radar and kinematic GPS surveys. Based on the spatial extent of elevated bed-echo intensities and the closed hydropotential basin, South Pole Lake is similar in size (~35 km2, smaller than previously stated ~100 km2) and hydropotential setting (bounded on the downstream side by hydraulic gradients of ~20 kPa km–1) to active West Antarctic lakes. Alongside our field observations, we use a temperature model to assess the likelihood of melting/freezing at the ice–water interface over a wide range of plausible geothermal flux and modern and paleo accumulation rates for this area. In contrast to previous hypotheses, our geophysical data and temperature modeling indicate that a temperate bed is realistic at the South Pole.


Multipass SAR processing for radar depth-sounder clutter suppression and tomographic processing and measurement of displacement

Bailey Miller, John Paden, Emily Arnold, Tom Jordan, Carl Leuschen, Keith Nicholls, David Porter, Carl Robinson, Fernando Rodriguez-Morales

Corresponding author: John Paden

Corresponding author e-mail: paden@ku.edu

In radar sounding, multipass measurements can be used to increase the spatial aperture of the radar to refine spatial resolution and to measure surface displacement due to firn compaction and basal melting. Multipass airborne synthetic-aperture radar (SAR) side-looking measurements have a long history of successful topography and displacement estimation of the Earth’s surface. However, airborne multipass SAR tomography has not been demonstrated for radar depth-sounding. Likewise, ground-based multipass measurements have been used to measure firn compaction, but not from an airborne platform. Airborne multipass displacement measurements using incoherent radar images have been made for basal melting, but so far only ground-based radars have made use of phase-coherent measurements to achieve the fine resolution promised by phase-sensitive methods. One of the requirements to achieve clutter suppression and phase-coherent displacement measurements is to properly sample the signal in space in order to both reject the unwanted clutter signal and coherently sum or compare the desired signal. To properly sample the signal, the platform needs to precisely control the flight path. We call deviations from the desired trajectory the flight path error. The other type of error is due to GPS trajectory and attitude errors of the radar antenna phase centers which we call GPS error. We systematically vary flight-path error and GPS trajectory error and produce signal-to-noise ratio performance grids that can be used to understand the level of GPS error and flight-path error that can be tolerated. Under the assumption of a point target model, we find that the system should be very robust to flight-path errors, even in the order of many wavelengths, but GPS errors must be kept much smaller than a wavelength to avoid substantial reductions in performance. We show results from several multipass datasets: 1) a uninhabited aerial vehicle (UAV)-mounted low-frequency (35 MHz) sounder dataset taken over Russell Glacier, Greenland; 2) the same low-frequency radar over Jakobshavn Glacier, mounted on a Twin Otter and with increased transmit power; 3) preliminary results from the accumulation radar installed on a British Antarctic Survey Twin Otter as part of the recent Thwaites MELT campaign; and 4) opportunistic repeat-pass measurements from the Operation IceBridge campaigns.


Mapping the time-evolving firn structure on South Cascade Glacier, Washington, USA, using monopulse ice-penetrating radar

Annika Horlings, Benjamin Hills, John Christian, Erin Whorton, Knut Christianson

Corresponding author: Annika Horlings

Corresponding author e-mail: annikah2@uw.edu

Firn on temperate alpine glaciers is important in near-surface processes, such as surface and near-surface energy balance, meltwater transport and water storage. These processes affect glacier mass balance and dynamics, but they are not well captured in conventional glacier mass-balance measurements. However, ice-penetrating radar measurements can inform mapping of the firn structure and reveal water-transport pathways or evolution of the internal layers of the firn as the snow compacts into glacier ice. Ice-penetrating radar data were collected in April 2017 and 2018 from the accumulation zone of South Cascade Glacier, a temperature alpine glacier in the North Cascades Range of Washington, USA, and one of the most comprehensively studied glaciers in the Western Hemisphere. We identified the accumulation zone through a coherent reflector at the previous summer surface, as well as the presence of internal firn layers below from previous summers. Within the accumulation zone, we mapped successive firn layers, matched firn layers between both years that data were collected, and attempted to determine firn compaction by measuring thickness changes between successive layers through time. We then ran firn-densification models with a meltwater scheme from the Community Firn Model, a Lagrangian framework that includes 13 published firn-densification models. We comparde the firn structure and interpreted compaction rates from the radar profiles to the model-predicted compaction rates. From this, we assess the ability of repeat monopulse ice-penetrating radars to resolve firn-compaction rates on temperature alpine glaciers.


A new, flexible airborne HF radar sounder for temperate ice studies, and application to major Alaskan glaciers

John Holt, Michael Christoffersen, Martin Truffer, Christopher Larsen, Brandon Tober, John Paden

Corresponding author: John Holt

Corresponding author e-mail: jwholt@email.arizona.edu

Radar sounding of temperate glaciers to determine ice thickness is a persistently difficult problem. Strong attenuation and scattering of radio-frequency signals caused by warm, crevassed ice necessitates transmitting high-power, long wavelength signals to achieve the necessary depth penetration. A common challenge for fast flowing glaciers is heavy crevassing, making surface-based radar infeasible. In this scenario airborne radar becomes necessary to quickly and safely acquire thickness measurements over the crevassed areas. Airborne radar presents a unique set of challenges, however. Many temperate glaciers inhabit relatively narrow valleys, which often causes ambiguity when interpreting radar data. Reflections from valley walls or other off-nadir features can very often appear to be the base of a glacier, leading to incorrect interpretation if this issue is not acknowledged and dealt with. We present a newly developed radar sounder suitable for airborne use in such environments. The radar transmits a 2 kw peak power linear chirp with a center frequency of either 2.5 MHz with 2.5 MHz bandwidth, or 5 MHz center with 5 MHz bandwidth. We typically operate the radar with a 5 μs pulse duration and 2 kHz pulse repetition frequency, allowing for significant along-track summation to improve SNR. We are also implementing along-track focusing to improve resolution. This radar has been operated in Alaska, USA, as part of NASA’s Operation IceBridge Alaska (OIB-AK), and we have detected basal reflections up to 900 m deep there. Due to the low flight altitudes required for laser altimetry measurements, we typically do not recover a surface echo from the radar data, and hence rely on laser altimetry for the surface position. Clutter simulation is a vital part of the interpretation process. Data have been acquired over many Alaskan glaciers including the Bering, Tana, Logan, Malaspina and Hubbard, which is the largest tidewater glacier in North America. The radar successfully penetrated to the base of the highly crevassed Hubbard terminal lobe, measuring thicknesses greater than 600 m. These new data are combined with previously acquired thickness measurements and historical bathymetric data to create a raster map of the bed below Hubbard. This bed map shows a large trough underneath the convergence of Hubbard and Valerie Glacier that reaches depths of over 400 m below sea level.


Using internal layers to constrain migration of the Greenland ice divide

Michelle Koutnik, Nicholas Holschuh, Knut Christianson, Andrew Hoffman, Joseph MacGregor, Benoit Lecavalier

Corresponding author: Michelle Koutnik

Corresponding author e-mail: mkoutnik@uw.edu

Volume change of the Greenland ice sheet through the last deglaciation must be constrained in order to understand glacial-to-Holocene sea level, as well as understand aspects of ice-sheet response to climate forcing. Relative sea-level data and exposure-age dates of marginal deposits are available along the Greenland coast and can be used with ice, earth and sea-level models to constrain the timing and magnitude of deglaciation and Holocene ice-sheet evolution, but direct constraints from the interior of Greenland are currently limited. However, dated internal layers are available and we present progress and a perspective toward using the available Greenland radiostratigraphy to provide constraints on ice-sheet evolution. Ice-divide position is primarily controlled by (1) the pattern and rate of accumulation and by (2) marginal ice dynamics. These factors change with the climate, and thus it is likely that the ice divide migrated during the last deglaciation. This idea is reinforced by continent-scale reconstructions of Greenland spanning the last 20 000 years, which exhibit central Summit divide migration of nearly 100 km. We present on the potential to use internal layers to constrain the reconstructed direction, magnitude and timing of divide migration. Radar-detected internal reflections represent past subaerial surfaces that have been buried by surface accumulation, reshaped by basal-ice melting and/or freeze-on, and displaced and strained by ice flow. These reflections preserve information about past spatiotemporal changes in those processes; however, disentangling the effects from multiple parameters can be challenging. Our goal is to determine the theoretical imprint of divide migration more generally, and evaluate the agreement between continent-scale model reconstructions and mapped isochrones spanning the modern divide. We focus on three radar lines from northern, central and southern Greenland in an effort to capture regional differences in divide migration due to regional ice dynamics and accumulation-rate change into the Holocene. This work provides a perspective that may inform future efforts to more rigorously assimilate layers into continent-scale model ice-sheet reconstructions.


Ice internal-layer tracking with deep multiscale neural network

Maryam Rahnemoonfar, Jimmy Johnson, Masoud Yari, John Paden

Corresponding author: Maryam Rahnemoonfar

Corresponding author e-mail: maryam.rahnemoonfar@tamucc.edu

Significant resources have been spent in collecting and storing large and heterogeneous datasets during expensive Arctic and Antarctic fieldwork. Recent large-scale radar surveys of Greenland and Antarctica reveal internal ice layers on a continental scale. While traditional analyses can provide some insight, the complexity, scale and multidisciplinary nature of the data necessitate advanced intelligent solutions. In recent years, the research community has witnessed advances in artificial intelligence (AI). Recent advances in deep neural networks (DNNs) and massive datasets have facilitated progress in AI tasks such as image classification, object detection and semantic segmentation. Despite this progress, these algorithms are limited to electro-optical data with large labeled datasets. There is a critical need to develop more advanced machine learning and deep learning algorithms for both visual and non-visual sensors collecting data for real-world scenarios during various polar ice missions. For polar ice applications, data-driven algorithms that can be utilized by domain experts are increasingly valuable. Detecting and tracing internal layers is a challenging task as these layers may merge together and can be weak or undetectable in some parts. Some layers are even hard for a trained expert to detect. Boundary detection intrinsically is a difficult process that requires the integration of information from various spatial scales and sources in order to balance the local and global information. This issue would be more intense for radar images due to their noisy characteristic. The current techniques for tracking of internal ice layers are mainly semi-automatic, which are computationally expansive and time consuming. They cannot be applied to a large dataset of ice internal layers on the continental scale. In this research we developed a fully automatic multiscale network for tracking ice internal layers. Our multi-scale network is supervised and integrated with wavelet features in the hidden layer. Wavelet decomposition and integration with the deep multi-scale network provides different organic contextual information at several levels. We have trained and tested our algorithm in snow radar data collected by CReSIS. Experimental results in comparison to the ground-truth and state-of-the-art technique demonstrate the effectiveness of our approach.


Greenland firn density from Operation IceBridge radar

Thomas B. Overly, Brooke Medley, Nathan T. Kurtz, H. Jay Zwally

Corresponding author: Thomas B. Overly

Corresponding author e-mail: thomas.overly@nasa.gov

The Greenland ice sheet’s porous firn layer acts as a reservoir for surface melt, inhibiting the melt from reaching the ocean. Understanding the magnitude and evolution of the firn column’s meltwater storage capacity will help constrain measurements of Greenland’s contribution to sea-level rise. We assess firn meltwater-storage capacity over the Greenland ice sheet using two Operation IceBridge (OIB) radars. Fitting a physically based model to the radar waveforms, we determine the absorption and scattering properties of the firn column. Firn density is then inferred through empirical and theoretical relationships between density and the retrieved extinction coefficient. We validate our OIB-derived density measurements using shallow firn cores collected by the Greenland Traverse for Accumulation and Climate Studies (GreenTrACS). We present OIB-derived firn density alongside Regional Climate Model density estimates. Future work seeks to apply our methods to CryoSat-2 waveforms, expanding the spatial coverage of firn density measurements and improving estimates of firn meltwater storage capacity.


Diurnal changes in observed glacier speed: insights into high-frequency glacier slip variability

Andrew Hoffman, John Christian, Annika Horlings, Ben Hills, Nick Holschuh, Knut Christianson

Corresponding author: Andrew Hoffman

Corresponding author e-mail: hoffmaao@uw.edu

Surface meltwater has been well linked to short timescale (seasonal to sub-daily) velocity changes of mountain glaciers; however, the spatial pattern and evolution of summer speed-ups and higher frequency (sub-daily) fluctuations in glacier basal sliding are less well understood, in part, because it so difficult to sample high-frequency, low-amplitude changes in glacier speed over large spatial areas. Using ground-based portable radar interferometer (GPRI), a new radar technology that can resolve sub-hourly changes in distributed glacier surface velocity, we can begin to understand the controls on high-frequency basal-slip variability, and diagnose whether mountain glaciers respond linearly to meltwater input. Over the last three summers, we regularly deployed a GPRI on the north face of Mt Baker, an active glaciated stratovolcano in the North Cascades of Washington, USA, scanning both the Coleman and Roosevelt glaciers to produce differential interferograms that we translate into a time series of line-of-sight glacier velocity. To isolate the spatial expression and variability of daily evolving subglacial drainage, we use frequency-dependent component analysis to identify velocity patterns with the maximum ratio of diurnal frequency to total variance. Our results provide insight into the linkages between surface melt and glacier slip that suggests that locally controlled glacier surface slope, bed conditions and melt episodes can drive very different glacier responses despite essentially identical climate forcing.


A general-purpose surface clutter simulator with multi-planet applications

Michael S. Christoffersen, John W. Holt, Scott D. Kempf

Corresponding author: Michael S. Christoffersen

Corresponding author e-mail: mchristoffersen@utexas.edu

A radar sounder operating from an airplane or satellite over a non-smooth surface will often receive reflections from off-nadir topography. This ‘surface clutter’ can be coincident in time with reflections expected from subsurface nadir interfaces, such as the base of a glacier, causing ambiguity when interpreting the radar data. Sometimes, clutter suppression is possible using an antenna with a focused beam pattern or a phased array of antennas, but in situations requiring very long wavelength signals or simple antenna geometry, these solutions become infeasible and steps must be taken to ensure that the clutter is not interpreted as subsurface data. One approach is to simulate all expected surface clutter using a digital elevation model (DEM) and the radar trajectory. This will allow an interpreter to rule out features in the radar data that occur in the same place as predicted clutter returns. Previously, this has been implemented on a case-by-case basis for a given radar and region. We have developed a facet-based, incoherent surface-clutter simulator that is capable of ingesting all common georeferenced raster DEM formats of any coordinate reference system and arbitrarily formatted navigation data. The parameters of the output simulations, such as sampling frequency and trace length, can be specified to directly match the instrument whose radar data is to be compared. In addition to a ‘cluttergram’ that essentially mimics the radar data format, we can produce left-side-only and right-side-only clutter simulations, left and right defined by the velocity vector of the sounding platform, and echo power maps depicting clutter strength and first-return locations on the surface. To date this software has been used to simulate products that match data from the SHARAD instrument on the Mars Reconnaissance Orbiter, the Lunar Radar Sounder onboard the SELENE lunar orbiter, and multiple airborne radar sounders used on Earth. The radar sounding of valley glaciers on Earth is particularly challenging due to the geometry of valley walls with respect to the radar, especially when flight paths are along the centerline for along-flow altimetry measurements. Once an approximate ice thickness is known, the clutter simulator can be used in a predictive capacity to evaluate flight paths to minimize clutter at the time delay expected for bed echoes. It can also be used to evaluate orbital parameters for mission planning.


Using snow radar to characterize the accumulation zone of South Cascade Glacier, Washington, USA

John E. Christian, Erin Whorton, Benjamin Hills, Joshua Driscol, Knut Christianson

Corresponding author: John E. Christian

Corresponding author e-mail: jemc2@uw.edu

South Cascade Glacier is a small temperate glacier in Washington’s Cascade mountains, and is one of the US Geological Survey’s four benchmark glaciers. South Cascade has a rich history of glaciological research, and the longest mass-balance record in North America (1959–present). The glacier has retreated and thinned since the early 20th century, and point mass-balance measurements indicate that little of the surface has experienced net accumulation in recent years. Ongoing change in the glacier’s geometry has motivated reanalysis of the spatial pattern of mass balance in order to improve glacier-wide mass-balance estimates and to understand South Cascade Glacier’s future in a warming climate. As a part of this effort, we are using ground-penetrating radar to map snow depth and englacial structure. In April of 2017 and 2018, we conducted surveys using a 500 MHz Pulsekko ground-penetrating radar system and differential GPS. In 2017, a centerline and several cross profiles revealed a consistent, strong reflector at 5–12 m depth over the entire glacier. Over most of the glacier, there were no continuous reflecting horizons below this layer, and we interpreted this reflector as the snow–ice interface. In limited portions of the upper glacier, there were multiple reflectors above and below the summer interface, which we interpreted as internal layers in the snow and firn. We identified the area with deeper reflectors as the accumulation zone. In 2018, we conducted a repeat survey that targeted this area with additional radar lines to better characterize the accumulation zone. These surveys suggest that the accumulation area is now limited to the glacier’s southwest margin, which is consistent with the extent of first-year firn visible in late-summer imagery. Winter snowpack is deeper in this region too, indicating that non-conventional accumulation processes are a substantial control on the location and extent of the modern accumulation zone. Based on the surrounding topography and characteristics of internal layers, we hypothesize that avalanches and wind deposition supply the extra snow necessary to maintain this small accumulation zone. Notably, the highest mass-balance stake falls outside the region of enhanced accumulation and typically experiences negative annual balances. These surveys and continued radar work will provide distributed maps of snow depth and will improve assumptions about elevation dependence in existing mass-balance models.


A comparison of radar-inferred temperature characterization techniques to investigate thermal regime changes in Antarctica

Eliza Dawson, Dustin Schroeder, Alex Miltenberger, Winnie Chu, Helene Seroussi

Corresponding author: Eliza Dawson

Corresponding author e-mail: ejdawson@stanford.edu

Temperature is a key control on the flow field of ice sheets – both ice viscosity and the onset of sliding are dependent on temperature – yet direct measurements of the thermal state are rare. Currently, the only direct temperature profiles available are from expensive ice-core drill sites. Away from these sites there is little to no information about the temperature field. Radar sounding provides a powerful technique for understanding thermal processes since the attenuation of radio waves through ice is temperature-dependent. Using Operation Ice Bridge radar-sounding flight lines and existing empirical relationships between radar attenuation, temperature and chemistry, we calculate depth-averaged englacial attenuation rates. We apply several attenuation correction approaches, including use of bed reflectivity, layer tracing and a new technique using a Bayesian framework, to reconstruct temperature profiles and provide a comparison of techniques for parts of Antarctica with unique thermal regimes. We use these methods to study parts of interior Antarctica characterized by a frozen bed, sections of the East Antarctic coast characterized by fast-flowing outlet glaciers that are thawed near the terminus but might have originated in frozen regimes, and parts of the Weddell Sea sector characterized by streaming flow. Providing a comparison of radar-inferred temperature structure aids in constraining the thermomechanics and dynamics in ice-sheet models.


Multi-instrument synthesis of radar-sounding observations of the Thwaites Glacier and Pine Island Glacier catchments, West Antarctica

Eliza Dawson, Dustin Schroeder, Andrew Hilger, Davide Castelletti, Winnie Chu, Thomas Jordan, Helene Seroussi, Duncan Young, David Vaughan

Corresponding author: Eliza Dawson

Corresponding author e-mail: ejdawson@stanford.edu

Recent observational and modeling studies have shown that the behavior and stability of both Thwaites Glacier and Pine Island Glacier in the Amundsen Sea Embayment of the West Antarctic Ice Sheet are modulated by a combination of ocean forcing, bed topography and basal conditions. Despite this, little research has focused on characterizing the basal condition context for modeling current and potential interaction across their boundary. This is due, in part, to the fact that modern radar data in this region were collected by three different radar systems and much of the Thwaites/Pine Island boundary lies at the boundary of these data sets. These include the 2004 survey of Thwaites Glacier by the University of Texas Institute for Geophysics HiCARS system and the 2004 campaign over Pine Island Glacier by the British Antarctic Survey PASIN system. To address this, we process and present profiles from both surveys. We also present estimates of bed reflectivity and attenuation rates spanning both glacier catchments. These estimates also include cross-calibration to account for the radar-sounding system differences in power and center frequency. This provides the first cross-survey map of basal reflectivity, abruptness and attenuation spanning the entire Amundsen Sea Embayment. This 2004 map also serves as baseline for comparing more recent and upcoming radar profiles across the region to understand its subsurface evolution. Most notably we find that (contrary to water-routing models on existing bed and surface data sets) subglacial water from the Bentley Trench flows into the Pine Island catchment, redrawing the subglacial hydrologic catchment to match the surface ice-flow catchment.


The ultrawideband airborne radar survey of Hiawatha Glacier and its implications for future investigations of ice-sheet margins

Joseph MacGregor, Mark Fahnestock, Tobias Binder, Olaf Eisen, Veit Helm, Kurt Kjær, Nicolaj Larsen, Mathieu Morlighem, John Paden

Corresponding author: Joseph MacGregor

Corresponding author e-mail: joseph.a.macgregor@nasa.gov

The hypothesis that an impact crater underlies Hiawatha Glacier in northwest Greenland was motivated by serendipitous NASA airborne radar sounding, but geologic field mapping and a focused survey using a new ultrawideband (UWB; 150&ndasg 520 MHz) radar system onboard the Alfred Wegener Institute’s Polar 6 Basler DC-3T aircraft were essential to confirm this hypothesis. Here we describe the multiple anomalous subsurface features observed by this survey and discuss the broader potential value of UWB airborne radar sounding of ice sheets. In the near-surface of Hiawatha Glacier’s ablation zone, discrete cross-bedded units up to tens of meters thick are likely superimposed ice associated with advected former supraglacial lakes. Elsewhere, numerous emerging reflections intersect the surface where satellite-observed outcrops of units with known visual characteristics from the Holocene epoch and the Last Glacial Period (LGP), and radar-reflectivity and surface-stratigraphy patterns are self-consistent. The Holocene–LGP interface is conforming above the northeast half of the crater but undulates dramatically at sub-kilometer scales above its southwestern half, likely recording a significant change in ice flow of unknown origin. Within the basal ice above the crater, numerous bed-originating reflectors emanate mostly from the central uplift, and point scatterers are common, which we interpret as evidence of basal freeze-on over a bedrock obstacle and vigorous subglacial erosion of the crater, including quarrying of boulder-sized clasts. Finally, a remarkably flat reflection typically 15 m below the ice–bed interface is sometimes detected at the downstream end of the crater, which likely represents the first detection of a subglacial groundwater table below presumably well drained impact breccia. As with previous generational advances in airborne ice-penetrating systems, we should expect that UWB surveys of ice sheets will detect previously unobserved subsurface features and identify new directions in radioglaciology. While some of the reported features may be unique to Hiawatha Glacier, our observations demonstrate the value of this new UWB airborne radar system to advance our understanding of processes at ice-sheet margins, which have historically been a challenging radioglaciological target.


Mapping firn aquifers on antarctic ice shelves using satellite, airborne and field-based radars and microwave radiometers

Julie Miller, Ted Scambos, David Long, Bruce Wallin, Clement Miege, Lynn Montgomery, Mary Jo Brodzik, Molly Hardman

Corresponding author: Julie Miller

Corresponding author e-mail: jzmiller.research@gmail.com

We present satellite- and airborne-derived indications of firn aquifers on ice shelves in Antarctica. Our observations are based on time-series analysis of microwave data collected from multiple active (ASCAT, ERS) and passive (SMAP, AMSR-E, SSM/I, SMMR) satellite instruments, and radio-echo sounding data collected as part of British Antarctic Survey and NASA Operation IceBridge airborne campaigns. We also present field validation of a perennial firn aquifer on the Wilkins Ice Shelf, the first such confirmation for the continent. Recent work in Greenland has demonstrated that both C- and L-band radar backscatter and microwave brightness temperature time series exhibit distinct signatures indicative of subsurface meltwater storage, and radio-echo sounding data exhibit reflections from the upper surface of meltwater stored at depth as well as an absence of reflections from the base. These indicators are also observed on several ice shelves in Antarctica, and are used to map firn aquifer extent. Two field sites in the Antarctic Peninsula selected on the basis of the satellite and airborne analysis were visited by a science team in December 2018: (1) the southern Wilkins Ice Shelf (70.80° S, 71.72° W), and (2) the southeastern George VI Ice Shelf (72.89° S, 68.43° W). A perennial firn aquifer was identified at 12–15 m depth at the Wilkins site, but only ice layers (up to 20 cm thickness) were seen at the George VI site. Snow and firn characteristics were derived from shallow firn cores (~35 m) and snow pits (~1.5 m), and local ground-based GPR profiles. Weather conditions at both sites are being monitored from two Automated Meteorology-Ice-Geophysics Observing System (AMIGOS) stations installed as part of the field effort. Previous studies have suggested that the critical component of hydrofracture-driven ice-shelf disintegration is meltwater ponding during the summer months, which provides a reservoir to feed propagating fractures until they reach the ice-shelf base. However, perennial firn aquifers, such as the one observed on the Wilkins Ice Shelf, represent alternate reservoirs capable of storing significant volumes of mobile meltwater year-round, within porous recrystallized snow. Past disintegration events on the Wilkins Ice Shelf may be linked to this expansive reservoir. Thus, the formation of perennial firn aquifers — rather than ponds – may rapidly induce ice-shelf instability in other high-melt, high-accumulation areas of Antarctica.


The subglacial roughness of Antarctica in a global context

Duncan Young, Jamin Greenbaum, Lucas Beem, Jason Roberts, Donald Blankenship

Corresponding author: Duncan Young

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

Over the last 25 years, extensive ice-penetrating radar (IPR) coverage of Antarctica has been obtained, at line spacings down to 1 km in some cases. However, many glacial processes occur at finer scales, so inferring likely landscape parameters is required for a useful interpolation between lines. Here we present a compilation of IPR-derived profile roughness data covering three great basins of Antarctica: the Byrd Subglacial Basin in West Antarctica, and the Wilkes Subglacial Basin and Aurora Subglacial Basins in East Antarctica; and treat these data using root-mean-square deviation (RMSD). Coverage is provided by a range of IPR systems with varying vintages with differing instrument and processing parameters; we present approaches to account for the differences between these systems. We use RMSD to investigate the self-affine behavior of the bed at kilometer scales and extract fractal parameters from the data to predict roughness and uncertainties in ice-thickness measurement and compare to a simulation of a global vegetation-free subaerial topography derived from the SRTM mission. This approach allows us to place Antarctica’s topography in the context of Earth’s global topography. We systematically compare Antarctica’s roughness distribution to both recently glaciated landmasses and tectonically isolated continents.


Radar reflections from basal ice may be misinterpreted as frozen bed

Neil Foley, Slawek Tulaczyk, Jake Walter

Corresponding author: Neil Foley

Corresponding author e-mail: ntfoley@ucsc.edu

Radio-echo sounding (RES) is the principal tool for investigating beds of glaciers and ice sheets due to its ease of use and the wealth of information it delivers about ice thickness, structure and bed conditions. Especially in polar ice, dielectric properties are such that RES signal experiences limited attenuation and penetrates to the bed. However, conditions at the beds of glaciers are often very different from those within the ice column, and these conditions can modify the electrical properties of ice and complicate interpretation of RES data. In particular, the addition of ions to frozen-on basal ice may significantly modify dielectric permittivity and electrical conductivity to the point that radar attenuation greatly increases, preventing RES from imaging the bottom of the ice. This has been previously observed in accreted marine ice of ice shelves. Here, we show evidence of a similar scenario in Taylor Glacier (TG), a land-terminating outlet glacier draining the East Antarctic ice sheet. Past analyses of relative radar bed reflectivity indicated that the lowermost 4–5 km of TG has a frozen bed with low radar reflectivity. However, EM resistivity measurements indicate that liquid brine is widespread beneath this part of the glacier. TG has thick basal ice visible at its terminus. We used collocated transects of 100 MHz RES and active-source seismic reflections surveys to image the bottom of the TG. Our results show that the RES consistently reflects ~15 m higher than the seismic measurements, which we interpret as indicating the thickness of the basal ice. Seismic waves are sensitive to elastic properties of ice, which do not change much between meteoric ice and basal ice, so they pass through and reflect off the actual glacier base. Radio waves, however, reflect off the strong dielectric contrast between pure meteoric ice and salty basal ice, with the transmitted energy rapidly attenuated in the basal ice so no actual basal reflector is observed. Our observations at TG highlight the need to carefully interpret bottom radar reflectors in places where basal conditions are not known. Dim basal radar reflections are often interpreted as indication of frozen conditions at the base. Our findings show that basal ice may have sufficiently different electrical properties from meteoric ice to cause similar dim radar reflections despite overlying brine-saturated sediments.


Spatial variability of snow depth observed with ground-penetrating radar and airborne lidar measurement on Hardangervidda, southern Norway

Kjetil Melvold, Thomas Skaugen

Corresponding author: Kjetil Melvold

Corresponding author e-mail: kjme@nve.no

In the mountains of Norway, the snow depth is highly variable due to strong winds and open terrain. This has a significant influence on the duration of the melting season, and thus on flood and runoff estimation. To investigate snow conditions on one of Europe’s largest mountain plateaus, Hardangervidda, we conducted snow-measurement campaigns in spring 2008 using ground-penetrating radar (GPR), global navigation satellite system (GNSS) and airborne lidar scanning (AL) at the approximate time of annual snow maximum (mid-April). We used a commercial GPR system with a 350 MHz and 1000 MHz and GNSS receiver that was mounted on a sledge behind a snowmobile in order to map an approximate 80 km long transect. The same transect was mapped by AL 5 days after the GPR survey and in snow-free conditions in autumn 2008. The radar data show highly variable snow depths ranging from 0 m to 10 m and detailed snow stratigraphy down to the ground surface. We present the distribution of snow depth as derived from the AL data and GPR data. The analysis shows that the spatial distribution of snow depth obtained from the AL is similar to that that obtained by GPR. At the local scale some variation exists between AL snow depth and GPR-derived snow depth using a constant radar-wave velocity. We will therefore investigate whether the AL and GNSS snow-depth data can be used to investigate the variability of radar-wave velocity along the radar profile by relating this snow depth to two-wave-travel times obtained from the radar data.


Four decades of radio-echo sounding of temperate glaciers in Iceland

Helgi Björnsson, Finnur Pálsson, Eyjólfur Magnússon

Corresponding author: Helgi Björnsson

Corresponding author e-mail: hb@hi.is

In the mid 1970s, low-MHz-frequency radar was developed in Iceland for the radio-echo sounding of temperate glaciers. From that time, we have applied this technology in extensive studies of most glaciers in Iceland. We have quantified ice mass volumes, provided detailed topographic maps of subglacial landscapes, demarcated glacial drainage basins, located subglacial lakes and volcanoes, and mapped internal tephra layers. With the addition of high-frequency sounding, the boundaries between cold and warm ice were mapped in polythermal glaciers in Svalbard and the subglacial surface was determined. In 1983, the same low-frequency radar was used to locate WWI aeroplanes at 80 m depth in the temperate ice sheet of southeast Greenland. The resulting data have been used for modelling past and future glacier evolution. The have found many practical applications in glacier hydrology, including hydropower operation, planning roads and bridges, and predictions of the paths of catastrophic floods accompanying subglacial eruptions, as well as in predictions of the emerging landscapes of the ~4000 km2 of Iceland that may soon become glacier-free.


The ice flux to the Lambert Glacier–Amery Ice Shelf system from the east side, East Antarctica

Xiangbin Cui, Wenjia Du, Huan Xie, Bo Sun

Corresponding author: Xiangbin Cui

Corresponding author e-mail: cuixiangbin@pric.org.cn

Study of the mass balance of the Antarctic Ice Sheet (AIS) is critical to estimate its potential contribution to global sea-level rise in the future. As the largest drainage system, the Lambert Glacier–Amery Ice Shelf drainage basin plays an important role in the mass balance of the Antarctic Ice Shelf. In this study, the ice thickness measured by airborne ice-penetrating radar and the ice velocity measured by in-situ GPS stations along the route of the Chinese inland traverse from Zhongshan Station to Dome A, passing through the east side of the drainage system, were used to calculate the ice flux with unprecedented accuracy. The results show that the total ice flux across the traverse is about 24.7 ± 2.8 Gt a–1. Two drainage basins crossed by the Chinese National Antarctic Expedition (CHINARE) show ice flux values of 20.9 ± 1.9 Gt a–1 (drainage basin B–C) and 3.8 ± 0.4 Gt a–1 (drainage basin C–Cp), respectively. The ice flux values in both regions are coincident with the mass balance calculated from ICESat. Meanwhile, the C–Cp basin shows an ice flux value of 6.6 ± 0.8 Gt a–1 across the grounding line, contributing to a 0.018 ± 0.002 mm a–1 sea level equivalent.


CReSIS airborne radars and platforms for ice and snow sounding

Emily Arnold, Carl Leuschen, Richard Hale, Shawn Keshmiri, Jilu Li, John Paden, Fernando Rodriguez-Morales

Corresponding author: Emily Arnold

Corresponding author e-mail: earnold@ku.edu

Airborne remote sensing using active radars has become an effective tool in geoscience fields for conducting Earth observations. As compared to ground-based and satellite-based methods, airborne remote sensing offers significantly larger spatial coverage than ground-based methods and the ability to conduct finer-grid measurements than satellite-based methods. Over the last 30 years, researchers at the University of Kansas have developed a series of airborne radars for ice- and snow-sounding measurements, and over the last 15 years they have developed several unmanned aerial systems (UAS) for the purpose of polar remote sensing. These research partners and others joined efforts to form the Center for Remote Sensing of Ice Sheets (CReSIS) at the University of Kansas 14 years ago. Current CReSIS radar systems operate over a frequency range of 14 MHz to 38 GHz. The various designs of the CReSIS radar depth sounders – designed to measure bed topography and ice thickness – operate over 14 MHz to 600 MHz. The accumulation radar, used to measure bed topography and internal layering, operates over 600–900 MHz. The snow radar, which operates over 2–18 GHz, measures snow cover and shallow snow layers. Finally, the Ku-band and Ka-band radars operate over 12–18 GHz and 32–38 GHz, respectively, to measure surface topography and near-surface layering. Specific frequency bands of operation for each radar system are driven by the required depth of signal penetration, operating frequencies of the radar antennas (which are often influenced by aircraft integration), and frequency spectra allocated. This paper will provide an overview of the current CReSIS sensors and platforms as well as representative results from these systems. We will highlight the recent system advancements, including: 1) increasing system bandwidth necessary for higher-resolution ultra-wideband (UWB) measurements, 2) miniaturizing radar hardware for integration onto smaller, less expensive UAS platforms; and 3) modifying systems for high-altitude platform operations. For UAS platform development, we have focused on developing smaller, easier to operate and less expensive platforms. CReSIS is now focusing on next-generation platforms with vertical takeoff and landing (VTOL) capabilities to further expand their accessibility to scientists and to increase deployment flexibility by eliminating the need for a runway.


AntArchitecture: an international project to use Antarctic englacial layering to interrogate stability of the Antarctic Ice Sheets

Robert G. Bingham, Olaf Eisen, Nanna B. Karlsson, Joseph A. MacGregor, Neil Ross, Duncan Young

Corresponding author: Robert G. Bingham

Corresponding author e-mail: r.bingham@ed.ac.uk

‘AntArchitecture’ is a new Action Group of the Scientific Committee for Antarctic Research that aims for the first time to determine the stability of the Antarctic ice sheets over past glacial cycles directly from the internal architecture of the ice. Internal architecture describes the 3‐D internal structure of the ice imaged by multiple radar‐sounding surveys undertaken across Antarctica over the last five decades. AntArchitecture aspires to bring together the key datasets on Antarctic Ice Sheet internal layering from the principal institutions and scientists who have been responsible for acquiring, processing and storing them over the last four decades. Key activities include coordinating data-transfer and data-lodging exercises between institutions/countries that will allow datasets acquired by different radar systems to be combined for pan‐continental analysis, and the development of an optimized processing flow for analysis of past data and advice on where future data acquisition needs to be targeted. An expanded outline of AntArchitecture and its timeline of activities can be accessed here: https://www.scar.org/science/antarchitecture/about/ This presentation provides a status report of activities and achievement of AntArchitecture to July 2019.


Comparison of 2-D and 3-D ice-bottom tracking algorithms

Victor Berger, John Paden, Mohanad Al-Ibadi, David Crandall, Geoffrey Fox, Anjali Pare, Maryam Rahnemoonfar, Mingze Xu, Masoud Yari

Corresponding author: John Paden

Corresponding author e-mail: paden@ku.edu

Ice-bottom topography measurements are a critical boundary condition for modeling ice flow and predicting the behavior of an ice sheet in a warming climate. For decades, low-frequency radars have been used to detect the ice bottom because of the low dielectric loss of ice in the HF and VHF frequency bands. Manually tracking the ice bottom in these images is laborious for large airborne campaigns, but the task is feasible. Simple semi-automated active contour (snake) tracking techniques have helped reduce the time commitment as well. Since 2005, multichannel antenna arrays deployed by the University of Kansas and now other institutions have made it possible to generate 3-D swath imagery in a single pass using the multiple signal classification (MUSIC) array-processing algorithm. The volume of data and the difficulty of visualizing 3-D imagery effectively necessitate automated tracking methods. In the past few years, methods have been developed and demonstrated to automatically track the ice bottom in 2-D and 3-D radar depth-sounder imagery. However, these various methods were applied to different datasets and evaluated using different tracking-accuracy metrics, which makes it difficult to properly compare the performance of the algorithms. In this work, we use the same training and test datasets for each algorithm and compare their performance directly against manually-tracked ground-truth data, using uniform performance metrics. We also discuss and compare the performance of a new 3-D tracker which is able to adaptively work with the more optimal maximum likelihood estimation (MLE) array-processing algorithm used for estimating the direction of arrival.


Thwaites MELT accumulation radar

John Paden, Fernando Rodriguez-Morales, Carl Robinson, Alejandra Escalera, Richard Hale, Krishna Karidi, Cameron Lewis, Bradley Schroeder, Jiaxuan Shang

Corresponding author: John Paden

Corresponding author e-mail: paden@ku.edu

We have developed and deployed a compact two-channel accumulation radar for phase-sensitive repeat-pass displacement measurements of basal melt and firn compaction. The radar operates from 600–900 MHz with a 400 W transmitter. This frequency range has sufficient penetration for thin ice and enables the use of a small antenna array that can be installed in the limited space underneath the floor of a Twin Otter aircraft. We detail the system electronics and antenna including the various calibration mechanisms built in to monitor phase stability and system health. The system includes two channels to simultaneously capture a low-gain and high-gain version of the signal so that the strong surface and shallow-layer scattering as well as weak ice-bottom and deep-layer returns can be captured. A high-speed transmit–receive (T/R) switch is used to capture an undistorted surface return at a flight altitude of 1500 ft (450 m) AGL. Zero-pi modulation is used to reduce coherent noise from the system. We added a directional coupler at the output of the T/R switch to monitor the signal as it travels to and from the antenna. This signal is attenuated and then supplied to the receiver via a calibration switch in the receiver to continuously monitor the phase response of the system. The radar system was deployed in 2019 on a British Antarctic Survey (BAS) Twin Otter as part of the Thwaites NSF–NERC MELT project. To integrate to the BAS aircraft, we redesigned the camera bay plate to fit both the BAS Riegl lidar and the radar antenna array. We performed modal analysis on the camera bay plate to verify that the operational blade-passage frequencies of the aircraft would not cause vibrational failure. Structural analysis was performed to ensure the maximum displacement was less than 0.1 mm relative to the original plate with only the lidar installed. Stiffeners were added around the plate to achieve this low displacement, and stiffeners were also added to the antenna housing to achieve desirable modal responses. The antenna array comprises four Vivaldi antennas in a metal/fiberglass composite housing lined with absorbent material. The antenna housing is designed to reduce reflections off the aircraft camera bay walls, to shield the antennas from aircraft EMI, and to reduce interference with the lidar. We present results from the survey, evaluate the radar’s noise and radiometric performance, and demonstrate the antennas-in-the-loop system impulse response.


Snow variability on Antarctic sea ice evaluated by a drone-mounted snow radar

Wolfgang Rack, Adrian Tan, Ian Platt, Greg Leonard, Gemma Brett, Pat Langhorne, Dan Price, Josh McCulloch

Corresponding author: Wolfgang Rack

Corresponding author e-mail: wolfgang.rack@canterbury.ac.nz

The snow cover on sea ice is of primary interest for a number of geophysical parameters such as ocean–atmosphere heat flux, surface albedo, radiative transfer and sea-ice thickness. These quantities are directly related to atmospheric circulation and high-salinity shelf water production. Despite the global significance of these processes, snow on sea ice remains insufficiently explored because of its complex morphology and dynamic nature, and the logistical and technical constraints on measuring even simple characteristics such as representative snow depth. Significant progress has been made using airborne snow radar, but the quality of these measurements depends on the sea ice morphology and validation is normally very difficult. To bridge the gap between surface and conventional airborne measurements, and in order to validate any kind of remotely sensed snow measurements or snow depth retrieval, we have developed and validated a drone-operated snow depth radar. The design goal of the radar system is to measure snow depth from 5 m height at 5 cm resolution over a radar footprint of 1 m. Reference measurements were conducted over Antarctic land-fast sea ice in 2016 using a horn antenna to identify key radar parameters. With these insights, the stepped-frequency radar was operated at a centre frequency of 3 GHz with 5 GHz bandwidth and was first flown in 2017. Here we give an overview of our radar design and its measurement platform, and we present the observed snow distribution on fast ice in the McMurdo Sound area of the western Ross Sea in 2017. Snow-depth statistics are compared to conventional, in situ measurements; the limitations in terms of snow depth and for operation over flooded sea ice are also discussed. Two-way travel time and properties of the reflected radar signal are related to the snow density and simple estimations of snow stratigraphy. The typical snow depth in our measurement area is 30 cm. We show that the drone-operated snow radar system is capable of characterizing snow depth up to about 50 cm depth and identifying internal layers over spatial scales in excess of a kilometre.


Radar reflection characteristics and melt–freeze variability at the Ross Ice Shelf, Antarctica, studied by phase-sensitive radar surveys

Wolfgang Rack, Adrian McDonald, Christina Hulbe, Craig Stevens, Oliver Marsh, Dan Price, Michelle Ryan, Kelly Gragg, Joe Snodgrass

Corresponding author: Wolfgang Rack

Corresponding author e-mail: wolfgang.rack@canterbury.ac.nz

The Ross Ice Shelf modifies the discharge of ice from both the West and East Antarctic ice sheets. As ice shelves are vulnerable to climate warming and their fate is controlled by oceanic processes in the ice-shelf cavity, a better understanding of melting and freezing patterns at the ice-shelf base is imperative. Obtaining direct access to the ice-shelf cavity for measurements is still difficult and costly. For the Ross Ice Shelf it has only been achieved twice over the past 40 years – the last time in 2017 at the Hot Water Drilling Site 2 (HWD-2) about 380 km southeast of Ross Island. The shortage of measurements has been improved by the development of autonomous phase-ensitive radio-echo sounders (ApRES) and a number of recent studies have been conducted to accurately quantify the magnitude and variability of basal melting. Here we present data from a suite of ApRES measurements across the Ross Ice Shelf over the period 2015–18. Measurement sites along the South Pole traverse (SPOT) route were re-occupied year after year. At HWD-2 an ApRES operated throughout the winter, and additional points were established in its closer vicinity. In 2018 new ApRES points were established across the whole ice shelf along the SPOT route, close to and within the grounding zone of Kamb Ice Stream and in the vicinity of a new drill site (HWD-1). We have analysed the radar waveforms for changes in reflection properties at the base, and for the pattern of internal layers and morphology. The results show a variability of waveforms between the ice shelf margin in the northwest and HWD-2 near the centre of the ice shelf. This relates to ice dynamics and the difference between basal melting and freezing. At some sites a fuzzy appearance of the basal reflections indicates either a very small amount of freezing, debris in the deepest section of the ice column, or a combination of the two. Visual observation at HWD-2 offers a direct validation of the wave form. Other sites show a clear melting pattern, sometimes small enough that it can only be revealed after a 1-year re-occupation of the measurement site. For the area between HWD-1 and HWD-2 we have so far only obtained measurements from one season, but the waveforms already give an indication as to where we may be able to record basal melting after a re-measurement in later years. We will present a map of basal conditions of the ice shelf and show a comparison to radio-echo-sounding measurements from earlier years.


Bed roughness as a control on the drainage of subglacial water

Timothy Creyts, Dustin Schroeder, Cyril Grima, Winnie Chu, Thomas Marcel Jordan, John Paden, Riley Culberg

Corresponding author: Timothy Creyts

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

Drainage through subglacial water systems depends on supply, storage and routing. Most deterministic models of drainage make a priori assumptions about the morphology, shape and extent of the water system. These assumptions determine how the competition between opening and closing of the water system leads to changes in basal water pressure. Estimates of ice-sheet discharge to the ocean is linked critically to sliding rules that are controlled by basal water pressure. For these deterministic water models, pressure is calculated on small scales relative to ice drainage, leading to a potential mismatch at leading scales. The deterministic models of the water system are implicit in ice models that include sliding coupled to drainage. Here, we depart from these assumptions and use scales associated with the radar. In particular, we are interested in using roughness as measured by radar. These measurements have a footprint much larger than the small scales associated with the deterministic models, and we modify existing drainage theory to accommodate this broader-scale information. Bed roughness can be determined from airborne geophysical surveys and allows for extensive measures of bed characteristics. We examine how these opening and closing rules can be modified and how the existing formulations can be succinctly changed to incorporate data. Time permitting, we will show how the spatial structure of bed information measured from ice sheets can contribute to drainage-system behaviors. We conclude with a forward-looking discussion of how drainage-system models can be developed for ice-sheet models using radar data.


DASHER (deployable Antarctic Sheet exploration rovers)

Mark Haynes, Alex Gardner, Kalind Carpenter, Samuel Prager, David Hawkins, Tushar Thrivikraman

Corresponding author: Mark Haynes

Corresponding author e-mail: mark.s.haynes@jpl.nasa.gov

The DASHER (deployable Antarctic Sheet exploration rovers) concept is focused around developing a herd of low-cost autonomous robots for polar science applications. To date, the DASHER concept has been optimized to map ice-shelf subaqueous melt-channel evolution using multi-input multi-output ground-penetrating synthetic aperture radar (MIMO GPSAR). We report progress on technology demonstrations related to the DASHER mobility platform and GPSAR. The mobility platform is a four-wheeled 1 m2 robot that is solar- rechargeable and self-navigating with lidar and real-time kinemetric GPS localization. The GPSAR instrument consists of ultra-wideband software-defined radars built on the Ettus USRP E312 with frequency stacking and wireless MIMO timing synchronization. It uses a cavity-backed bow-tie patch antenna. Successful field demonstrations of the mobility platform and radar in deep snow have been conducted. Future planned tests consist of 1) multi-agent autonomy and 2) bed detection on ice sheets.


Coherent simulators for active/passive radar sounding and subsurface synthetic aperture radar processing

Mark Haynes, Yang Lei, Sean Peters, Andrew Romero-Wolf, Dustin Schroeder, David Hawkins, Charles Elachi

Corresponding author: Mark Haynes

Corresponding author e-mail: mark.s.haynes@jpl.nasa.gov

We report ongoing work in radar-sounding simulation and synthetic aperture radar (SAR) processing for low-frequency planetary and terrestrial radar sounders. High-fidelity simulation of complex surface and subsurface scattering scenes is needed for broad-based system engineering, subsurface SAR algorithm development, and validating/refining observable signatures of scientific hypotheses. Sounding domains are electrically large and computationally expensive and therefore require custom solutions that address specific aspects of the problem. Three classes of radar sounder simulator are presented: 1) a frequency-domain coherent multi-layer facet model based on a new partial-aggregate scattering-matrix method for producing radargrams of arbitrary pulse-repetition frequency from a single simulation, 2) full-wave 2-D pseudo-spectral time domain for heterogenous dielectrics with features to handle oblique incidence, and 3) passive sounding over faceted surfaces. We demonstrate SAR focusing for each method, with emphasis on findings related to signal-to-noise ratio, resolution and the point-target response of passive sounding SAR.


Retrieval of firn aquifer thickness and englacial water volume using ice-penetrating radar sounding

Winnie Chu, Dustin Schroeder, Matthew Siegfried

Corresponding author: Winnie Chu

Corresponding author e-mail: wchu28@stanford.edu

Along the coast of southwestern Greenland an extensive perennial firn aquifer was recently discovered using firn cores and ground-based and airborne snow-accumulation radar. Until now, studies were only able to map the location of firn aquifers but were unable to constrain aquifer thickness, a key attribute for assessing the ultimate impact of englacial storage on ice-sheet dynamics. We have designed a novel approach that combines NASA IceBridge radar-sounding data and numerical ice-sheet models to provide the first catchment-wide constraints on firn-aquifer thickness for Helheim Glacier. Our results reveal two dynamic aquifer bodies with varying thickness from 4 m to 25 m. Through repeated radar sounding and laser altimetry analysis, we also discovered that one of the aquifers has thinned up to 15 m, with the ice surface above lowered by up to 2 m in response to a reduction in surface melting between 2012 and 2014. Together, these new thickness estimates indicate that the Helheim Glacier firn-aquifer system can only store ~2.3 Gt of water, significantly less than the 4.8 Gt reported by earlier studies based on localized measurements from boreholes and passive seismic data.


Megadunes and wind glaze on Earth and Mars: variations with climate and slope

Ted Scambos, Jan Lenaerts, Mark Fahnestock, Christopher Shuman

Corresponding author: Ted Scambos

Corresponding author e-mail: teds@nsidc.org

Megadunes are large linear snow antidunes of kilometer-scale wavelength produced by prolonged air–snow interaction in a persistent katabatic wind environment. These are interspersed with wind-glaze bands, having near-zero accumulation and intense upper firn recrystallization. Continent-wide remote sensing analyses shows that terrestrial megadunes are found almost exclusively in mid to upper elevations of the East Antarctic ice sheet, in areas of near-uniform regional slope. However, in other areas, having higher wind regimes and slightly steeper slopes, glaze and megadune areas become more chaotic. Radar profiles from several past studies and field work by the authors show that, with increasing mean surface wind speed (from RACMO models and inferred from steeper regional slope), megadune layering shifts from continuous between dune crests (with thinner layer spacing between crests) to discontinuous as wind-glaze characteristics begin to develop between crests. With greater wind speed and ablation (regionally) megadune accumulation become isolated in small reverse-slope regions with wide areas of intense wind glaze and wind scour surrounding the dune. Martian analogs have been proposed based on orbiting radar systems there. Antarctic field work and in situ meteorologic data together with satellite measurements, new digital elevation models and climate-model results enable characterization of the full range of megadune forms as well as the relationship of megadune morphology to climate and topograhic setting. Fundamental characteristics of megadunes include wind-transverse crests, very low height-to-width ratio (~1 : 200) and higher windward-face accumulation relative to lower to near-zero (e.g. wind glaze) or even slightly negative (wind scour) lee-side surface mass balance. This leeward-side range produces a range of radar profile forms scaling with mean slope in the wind direction and regional surface-snow input. Megadunes and related wind-scour features represent an additional facies of Antarctica’s ice sheet firn, a distinct sub-facies of the dry snow zone in which air–snow interactions dominate and melt or near-melt processes do not occur. We present a conceptual model for their formation related to atmospheric standing-wave events in the near-surface inversion layer.


Airborne measurements of a Ka-band radar altimeter over Greenland land ice and Arctic sea ice

Jilu Li, Bruno Camps-Raga, Fernando Rodriguez-Morales, Carl Leuschen, John Paden, Inès Otosaka, Andrew Shepherd

Corresponding author: Jilu Li

Corresponding author e-mail: jiluli@ku.edu

Satellite radar altimeters are used to monitor the rapidly changing ice sheets and sea ice in polar regions. They provide spatial coverage and continuous temporal monitoring at a relatively low operation cost. However, the data interpretation is not straightforward because of the relatively coarse range resolution and large footprint of the altimeters, and the effects of the complex surface and snow cover on the signal penetration and the shape of the waveforms. On the other hand, airborne radar altimeters can achieve much smaller footprints and finer range resolution, being able to resolve internal reflections and thus provide a valuable means for the calibration and validation of satellite radar altimetry data. The penetration of radio waves in snowpack is significantly affected by frequency. Compared to Ku-band signals, a Ka-band signal attenuates quickly as the snow grain size increases due to volume scattering. Therefore, the Ka-band peak backscattering horizon will locate closer to the air–snow interface than the snow–ice interface when compared to the Ku-band peak. This difference in snow penetration presents a potential to retrieve snow depth by combining the observations of Ka- and Ku-band satellite radar altimeters. Additionally, ice-grain size could be derived directly from Ku-band scattering because the volume scattering dominates the response at the smaller Ka-band wavelength, which is not possible at the longer Ku-band wavelength because surface scattering dominates in this case. During the 2015 NASA Operation IceBridge Arctic campaign, the Center for Remote Sensing of Ice Sheets simultaneously collected airborne Ka-band and Ku-band altimeter data. The Ka-band altimeter operated at a center frequency of 35 GHz with a 6 GHz bandwidth. In this work, we describe the system and antenna design and the installation on the NASA C-130. We present sample results over land ice and sea ice, and discuss the observed backscattering characteristics and penetration depth based on comparisons with the Ku-band (12–18 GHz) measurements and theoretical backscattering models. We also compare the airborne altimeter waveforms with those from the SARAL/Altika satellite Ka-band altimeter at crossovers to illustrate that this data set can be used for the calibration and validation of the satellite observations.


Autonomous monitoring of englacial hydrology using stationary impulse radar systems

Gwenn Flowers, Laurent Mingo, David Bigelow

Corresponding author: Gwenn Flowers

Corresponding author e-mail: gflowers@sfu.ca

Englacial hydrological properties and processes are difficult to observe. Borehole sensors permit high temporal resolution but are limited to the point-scale, while geophysical surveys can provide a broad spatial picture but are limited to a snapshot in time. To characterize the role of the englacial hydrological system at high temporal resolution and over a sample volume during the filling and drainage of an ice-marginal lake dammed by the Kaskawulsh Glacier (sub-Arctic Yukon, Canada), we have deployed autonomous low-frequency impulse radar systems in three summer melt seasons from 2014–17. Equipped with a timing engine, each system is programmed to wake up every 3–4 h to make a measurement. Low power requirements for both the transmitter and receiver (7–10 Ah batteries and 10–20 W solar panels) allow autonomous operation for months at a time. The time-variable radiostratigraphy revealed by the radar systems indicates dynamic changes in the englacial hydrological system associated with the filling and drainage of the adjacent ice-dammed lake. Within 500 m of the lake, the sudden appearance of englacial reflectors is coincident with abrupt changes in water pressure measured in nearby shallow boreholes, consistent with rapid injection of water into the englacial system. Subsequent changes in two-way traveltime to persistent englacial reflectors closely mirrors the recorded lake level, suggesting a change in radar-wave velocity arising from a change in englacial water content that accompanies lake filling and drainage. These traveltime changes can be explained by a change in saturated englacial porosity of up to ~10% between lake-full and lake-empty conditions, within the range of englacial porosities required by an independent catchment-scale water balance. Calculations of radar internal reflection power at discrete depth intervals reveal both gradual and abrupt increases and decreases in englacial brightness within a kilometer of the ice front that we interpret as changes in englacial storage. This interpretation is corroborated by the results of spatially extensive common-offset radar surveys that document differences in storage before and after lake drainage. Autonomous monitoring with stationary-impulse radar systems has provided direct evidence of shallow and deep storage within a fractured englacial aquifer, which, at peak lake level, accommodates over 30% of all water in the catchment.


I wonder what’s under: radio-echo sounding glacier flow bands in the Ross Ice Shelf

Martin Forbes, Christina Hulbe, Kelly Gragg, Gregory Leonard, Wolfgang Rack, Craig Stevens

Corresponding author: Martin Forbes

Corresponding author e-mail: martin8forbes@gmail.com

In the austral summer of 2017, the Aotearoa New Zealand Ross Ice Shelf programme drilled two boreholes through the ice shelf at ~80°39′9″ S, 174°28′2″ E, approximately 450 km downstream of the grounding zone. Video footage of the borehole revealed that basal ~60 m of the ~360 m thick ice shelf was bubble-poor and contained irregularly distributed sediment. The boundary is sharp and appears to involve fractures propagated into the bubbly glacier ice. The two boreholes lie in ice that originated from the glacier-left margin of the Liv Glacier. We conducted subsurface imaging using low-frequency radio-echo sounding (RES) to provide glaciological context for the borehole site. Our radar profiles traverse a number of glaciers’ flow bands and suture zones from Mercer Ice Stream to Beardmore Glacier. The present aim is to explain the basal accreted ice observed at the borehole site and its implications for interpretation of RES imaging of glacier boundaries within ice shelves. Ice shelves are composites of ice originating from various glaciers and ice streams and, while the ice is continuous, it is not homogenous. The sutured margins between glacier flow bands have been of particular interest because those associated with rift-tip arrest may deform at different rates from adjacent ice. Marine ice has been observed, using RES, to underplate suture zones (the ice is relatively thin, making it a trap for rising fresh, cold meltwater). The characteristic signature is a bright reflector, an apparent basal reflector, that rises upward without an accompanying depression at the surface. The implication is that ice with different dielectric properties occupies the deep part of the ice column. Radio-echo sounding profiles across the Liv Glacier flow band exhibit characteristics attributed elsewhere to suture zones with basal accreted marine ice. Our direct observation of this ice, and the sediments it contains, demonstrate that it is not marine ice. Rather, it has a terrestrial origin. This drives us to more closely examine other glacier boundaries near the borehole site. We find that suture zones are themselves composites of ice with different characteristics such as thickness, apparent basal roughness and underplating by accreted ice.
Hot-water drilling was undertaken by the Victoria University of Wellington Hot Water Drilling initiative. We thank the hot-water drilling team led by A. Pyne and D. Mendeno. Logistics support was provided by Antarctica New Zealand.


Empirical noise estimation in time-domain back-projected bistatic SAR images

F. Scott Turner, G. Wesley Patterson, J. Robert Jensen, Davide Castelletti, Dustin Schroeder

Corresponding author: F. Scott Turner

Corresponding author e-mail: Scott.Turner@jhuapl.edu

NASA’s Mini-RF instrument on the Lunar Reconnaissance Orbiter (LRO) is currently operating in concert with the Arecibo Observatory (AO) and the Goldstone deep space communications complex 34-meter antenna DSS-13 to collect bistatic radar data of the Moon. These data provide a means to characterize the scattering properties of the upper meter(s) of the lunar surface, as a function of bistatic angle, at S-band (12.6 cm) and X-Band (4.2 cm) wavelengths. Mini- RF bistatic observations of the lunar polar regions include surface areas that are easily identified as being in complete RADAR shadow. The values of pixels on the processing grids within these shadow regions should be almost entirely noise. Unfortunately, the transmitter geometry makes using digital elevation models (DEMs) difficult to use for precise assessment of shadow region boundaries. Examining the statistics of pixels clearly interior to these regions, however, leads to a technique for noise estimation in the time-domain back-projected Mini-RF bistatic SAR images. It could be expected that local populations of pixels in shadow regions are truly noise, and thus would be independent and drawn from a normal distribution with fixed variance and mean. This would then lead to the conclusion that individual complex sample values are chi-squared distributed. While this is not precisely the case with the Mini-RF bistatic data, Monte Carlo simulations of processed noise can be used to derive collect-specific distributions. Further, a simple transformation applied to these distributions approximates the standard chi-squared one, which is then used to isolate noise pixels from signal. The resultant noise identification technique compares well qualitatively with detailed topographic illumination simulations.


Internal ice-penetrating radar stratigraphy at the Little Dome C Oldest Ice site

Marie Cavitte, Duncan Young, Catherine Ritz, Frederic Parrenin, Rob Mulvaney, Massimo Frezzotti, Julius Rix, Jason Roberts, Donald Blankenship

Corresponding author: Marie Cavitte

Corresponding author e-mail: mariecavitte@gmail.com

Little Dome C, just south of Dome C, is one of the primary target areas for the Beyond EPICA Oldest Ice (BE-OI) European project to recover an ice-core climate record that goes back 1.5 million years. Several ice-penetrating radar surveys have been collected over this region over the last decade. The older UTIG and OIA surveys consist of airborne radar data with a wider line spacing (≥1 km) and cover the entire Dome C region. Newer GPR surveys were then collected during the 2016/17 and 2017/18 field seasons and focussed on the Little Dome C area at a finer spatial resolution. We trace and date the internal stratigraphy to provide boundary conditions to an ice-flow model in order to constrain ice-flow history in the region and identify the best Oldest Ice drilling site. We have extended the 18 internal radar reflections mapped in the UTIG and OIA radar surveys into the new GPR surveys in order to obtain the most detailed internal stratigraphy at Little Dome C. The dense spatial coverage of the GPR surveys allows the tracing of additional deeper radar reflections. Most internal reflections are dated at the EPICA Dome C ice core using the AICC2012 age-depth chronology. The gridded and dense geometry of the new and old radar surveys implies a high number of crossovers, which allow for regular checks on tracing accuracy and allow us to obtain rigorous depth uncertainties. Finally, we use the Parrenin et al. (2017) 1-D inverse model to reconstruct a whole wealth of information on Little Dome C to assess its suitability for Oldest Ice: basal ages, basal age resolution, basal melting rates, steady-state accumulation rates, p exponent in the Lliboutry flow formulation, depth of the 1.5 Ma isochrone, etc., by inverting the internal dated stratigraphy.


Mini-RF S- and X-band bistatic observations of South Polar craters on the Moon

Wes Patterson, Parvathy Prem, Angela Stickle, Joshua Cahill

Corresponding author: Wes Patterson

Corresponding author e-mail: gerald.patterson@jhuapl.edu

Mini-RF aboard NASA’s Lunar Reconnaissance Orbiter is a hybrid dual-polarized synthetic aperture radar. Its receiver operates in concert with transmitters at either the Arecibo Observatory or the Goldstone deep space communications complex 34 mantenna DSS-13 to collect S- and X-band bistatic radar data of the Moon, respectively. Bistatic radar data provide a means to probe the near subsurface for the presence of water ice, which exhibits a strong response in the form of a coherent backscatter opposition effect. Mini-RF S-band bistatic observations of Cabeus Crater floor materials display a clear opposition response consistent with the presence of water ice. Initial X-band bistatic observations of Cabeus and Amundsen crater floor materials do not show a similar response. This could indicate that, if water ice is present in Cabeus Crater floor materials, it is buried beneath ~0.5 m of regolith that does not include radar-detectible deposits of water ice.


Temporal and spatial changes in reflectivity over 9 years beneath the Rutford Ice Stream, West Antarctica

Rebecca Schlegel, Tavi Murray, Edward King, Andy Smith, Alex Brisbourne, Adam Booth, Roger Clark, Stephen Cornford

Corresponding author: Wes Patterson

Corresponding author e-mail: gerald.patterson@jhuapl.edu

The Rutford Ice Stream, a fast-flowing ice stream in West Antarctica, has been subject of ongoing research over the past three decades. The main aim of the past and present research is to understand bed properties, glacier dynamics and the resulting bedforms. Both drumlins and mega-scale glacier lineations (MSGL) have been mapped beneath the Rutford Ice Stream. Topography changes over time have been identified in previous studies using seismic and radio-echo sounding (RES) techniques, showing that the basal sediments are highly mobile. These studies have also identified a transition from upstream deforming sediment to downstream basal sliding. We used RES data collected in 2007/08 and 2016/17 to calculate the bed reflection power over an 18 × 18 km grid. The resolution is comparable to that of typical marine swath bathymetry data. Data in both seasons were acquired over the same area with the same RES system, so direct comparison is possible. We combine the bed reflectivity with the locations of bedforms as well as temporal topographical changes to investigate the link between the processes of erosion and deposition. Patterns in the bed reflection power are elongated in the flow direction and are often correlated to bedforms. The transition zone between deformation and sliding correlates with spatial changes in reflectivity. Temporal changes in reflectivity around bedforms are identified and may contribute to the understanding of the formation of subglacial bedforms. The speed of ice streams is known to be strongly modulated by the interactions of ice and the underlying sediment, hence the shape of the bed and bedforms. Direct observations of the bed conditions, and the link between fast ice-flow speed and the shape of the bed, could improve the basal boundary conditions in ice-dynamic models.


Glacier topography and hydrology of the Indus Basin in the Himalayan region using remote sensing and GIS

Swati Tak, Ashok K. Keshari

Corresponding author: Swati Tak

Corresponding author e-mail:takswati7@gmail.com

The worldwide focus on glacier studies is increasing as they react to climate change and influence the hydrology and environment of the region. The current concern is mainly hydrological, climatic and the water resources of the Himalayan region. Various streams in the Indus basin have their origin in the glaciated mountain area. It contributes a significant part of the streamflow and is the main water source for the local economy. The highly varying complex topography, the uneven distribution of the glaciers and climatic variability lead to melting of ice. Therefore, a glaciers inventory has been created to read the topography and understand the hydrology of the study region. This study also brings out the climate change aspect very clearly through the advance and retreat stages of the glaciers. The retreat of some glaciers, particularly Parbati glacier, is a major concern in the Indus basin as the climate and topography of this region affects the hydrology of the entire basin. This study will be useful for planning, designing and managing of water systems in the Indus basin, especially for sustainability and environmental protection.