With the successful launches of ICESat-2 and CRYOSAT-2 the scientific community has a unique opportunity to improve and refine estimates of mass balance changes over Greenland. However, the high quality of these data sets must be matched by high quality retrieval and modelling schemes for density and accumulation and by studies that focus on understanding the interaction between the electromagnetic signal and snow properties. An important source of uncertainty for combining radar and lidar altimeter datasets over the Greenland ice sheet is associated with the penetration depth of the electromagnetic signal into the firn or snow. This depends on the vertical distribution of the snow/firn properties such as density and the presence of ice layers. Errors associated with penetration depth can result in a meter-scale spurious variations in surface height that are not simply related to changes in ice-sheet mass but rather associated with the interaction between the snowpack and the electromagnetic signal.
Within this context, the synergic use of regional climate and electromagnetic models in combination with airborne and spaceborne remote sensing datasets represents a high quality framework for producing improved estimates of mass change and attribution, to inter-calibrate data from past, current and future radar and lidar altimetry missions and to reduce the uncertainty associated with the elevation changes detected from current and planned spaceborne altimetry missions. In this talk, we present our results concerning two major aspects of the problems mentioned above. The fist aspect concerns the development of an offline module of the MAR regional climate model aimed at understanding the impact of the atmospheric forcing on the simulation of the snow vertical properties. We first show and discuss how the errors on atmospheric forcing propagate through the simulation of snow/firn vertical quantities and assess the performance of MAR by comparing its outputs with firn core measurements available from the SumUp community database and other datasets. Then, we use the outputs of the MAR model as inputs to a multi-layer electromagnetic model to compute the penetration depth of the electromagnetic signal and study how errors in the atmospheric forcing and the snow model’s physics propagate to the penetration depth and, therefore, how they impact estimates of elevation changes and their comparison with elevation changes obtained from lidar measurements

Area D - Board 433

Over the last decade, Synthetic Aperture Radar (SAR) altimetry has transformed our ability to monitor Earth’s ice sheets, and to track their response to changes in the climate. However, the steep and complex topography of mountainous regions means that this technique has seldom been used to monitor alpine glaciers, despite its relatively high spatial and temporal resolution. It is important to assess the utility of new techniques and instruments for systematically monitoring these data sparse areas because of the strong reliance on glaciers as water resources in mountainous communities and the potential hazard of glacial lake outburst flood events as ice recedes. SAR altimetry has proven particularly valuable in recent decades in resolving the detailed pattern of mass loss across Greenland and Antarctica, but with few applications to alpine glaciers. Following the recent launch of the Sentinel-3 satellites into 27-day repeat orbits, there is now an opportunity to assess the ability of these new SAR altimeter measurements to study changes in mountain glaciers and ice caps. This poster will present a preliminary assessment of Sentinel-3 altimetry data over such challenging terrain, and outline our broader project, which aims to use a range of new remote sensing platforms to quantify ice mass loss in these regions. We will focus on a first evaluation of the data over several glaciated regions, including analysis of Sentinel-3 tracks that cross the Cordillera Vilcanota, Peru, the Southern Patagonian ice field, and the Himalaya, and discuss the quality of the SAR altimeter data acquired to date. Our evaluation will include an examination of novel approaches to elevation retrievals from SAR altimeter waveforms, and comparison of Sentinel-3 altimetry data with existing digital elevation models. In the long-term, we will also evaluate the potential for SAR altimetry to provide regular, reliable estimates of glacier mass loss to allow local authorities, businesses, and homes to plan for changes in water availability.

Area D - Board 434

High PRF altimeters transmit pulses at a high pulse repetition frequency thus making the received echoes suitable for coherent processing on-ground.

Conventional delay-Doppler processing (DDP, commonly called SAR or High Resolution) coherently integrates echoes in a burst-by-burst basis to provide single look waveforms referred to a specific ground location, which after being correctly aligned (compensating for the slant-range migration, among others) can be incoherently averaged, increasing the performance in terms of the speckle reduction and the along-track resolution compared with the traditional Low Resolution Mode and in turns in terms of geophysical retrieval.

The Fully Focused delay-Doppler processing (FF-DDP, also known as Fully Focused SAR) moves one step ahead and intends to coherently integrate the echoes over a time longer than a burst to get an even higher along-track resolution with an improved speckle reduction with respect to DDP.

The Open Burst (or interleaved) transmission mode to be implemented in Sentinel 6 and the Polar Ice Topography (PICE) missions makes more suitable the exploitation of the FF-DDP thanks to the uniform along track sampling of the scene. However, in the conventional Closed Burst mode (like in CryoSat-2), replicas induced by the non-uniform sampling of the Doppler spectrum will be mixed with the main echo and, in most cases, won't be able to be filtered out.

Swath mode processing has been used to monitor elevation of areas with complex topography such as over ice sheet margins, ice caps and mountain glaciers, improving upon the resolution of conventional radar altimetry. Swath mode relies on an accurate angle of arrival of the measured echo, this is obtained from the SAR Interferometric mode of CryoSat-2 and PICE and post-processing strategies resolving the ambiguous nature of the phase measurement.

The PICE Mission will include Open Burst and Interferometric capabilities. It will be the first altimeter to be able to combine both methodologies to increase both the along and across-track resolutions improving the current performances of CryoSat-2 over small glaciers that can't be observed properly.

We will use real CryoSat-2 Closed Burst SARIn data to assess the current performances and limitations in the CR-2 mission. On the other hand, simulated Open Burst SARIn data will be generated to be then processed with FF-DDP and Swath mode to analyse the potential improvements that the PICE mission can bring when operating over sloped terrain and both methods are combined properly.

Area D - Board 435

The Sentinel-3 (S3) mission is today flying in constellation following the successful launch of the S3B satellite (April 2018) and associated commissioning phase. The first satellite, S3A, was launched in February 2016 and is nominally operated in routine phase since October 2017 along now with S3B, this latter being 140˚ ahead of S3A on the same orbit plane.
The S3 Surface Topography Mission (STM), built on the heritage of ESA’s ERS, Envisat and CryoSat missions, is primarily an ocean mission, however, the STM mission is intended to provide continuity with the ESA CryoSat High-Resolution sea-ice and land ice surface measurements used to derive floating ice thickness and ice sheet topography over the two poles.
The S3 Payload Data Ground Segment (PDGS) is operated in partnership between ESA for the Land products and EUMETSAT for the Marine products. But this common PDGS infrastructure prevents today a full flexibility to implement specific STM level 1 processing recommended by the user community for the Cryosphere Altimetry data processing. The expected evolutions of the ground processing cover in particular the inclusion of the zero-padding (i.e. over-sampling of the waveforms), the implementation of the Hamming weighting function for the sea-ice processing and the implementation of the waveform centering for the processing of the ice margins, where the slope is larger than 1%.
This paper aims at presenting the performance of the operational S3 STM sea-ice and land ice products following the last achievements implemented in the S3 STM Processing Baselines for both S3A and S3B, as well as the availability of new beta-products generated from ad-hoc ESA prototypes and the intended plans for dedicated end-to-end “L0 to L2” processing chains per surface types as an intended evolution of the S3 LAND PDGS.

Area D - Board 437

Sentinel-3B was launched in April 2018, joining Sentinel-3A (launched in February 2016) in a constellation aimed at providing optimal coverage and revisit times of the global oceans and the Cryosphere. Sentinel-3 STM (Surface Topography Missions) operate a KU/C-band SAR Radar Altimeter, continuing a 26 year heritage of ESA altimetry missions from ERS-1 in 1991, ERS-2 in 1995, ENVISAT's RA2 in 2002 and Cryosat-2 SIRAL in 2010. Although Sentinel-3 STM missions are primarily aimed at global oceans they provide good coverage of the Cryosphere up to 81.5° of latitude. This allows measurement of Arctic sea ice freeboard (from which thickness can be derived) and the elevation change of the ice sheets and ice shelves of Antarctica and Greenland (from which mass balance can be calculated) including the critical glacial margins which are a major contributor to global sea level rise.
Sentinel-3B is currently nearing the end of its 9 month commissioning phase during which it operated for 4-months in a tandem orbit, 30 seconds apart from Sentinel-3A, before being drifted to the same orbital plane, but 140° ahead. The tandem phase provides the opportunity to cross-calibrate both missions and instrument operating modes (LRM and SAR) over sea ice and land ice surfaces.
As an Expert Support Laboratory (ESL) for ice surfaces to the ESA Sentinel-3 STM Mission Performance Centre (MPC), UCL's Mullard Space Science Laboratory (MSSL) is currently leading an in-depth study in to the performance of Sentinel-3B, cross-calibration of Sentinel-3B with Sentinel-3A over sea ice and land ice surfaces and comparison with contemporaneous Cryosat-2 LRM, SAR and SARin measurements. Here we present the results from this commissioning study providing scientists and users of operational Copernicus Sentinel-3A and 3B data with detailed information on the respective inter-mission biases and instrument performance issues specific to sea ice and land ice parameters in the latest ESA Level-2 product baselines. We also present information on any planned Sentinel-3 L2 product evolutions relevant to ice surfaces which will be recommended for implementation as a result of this study and the on-going work of the MPC.

Area D - Board 438

Long-term Earth observation data is critically important to the global effort to combat climate change. Surface elevation change data from altimetry is one of a suite of parameters used to characterise ice sheets, whose variations affect sea level and climate. Sentinel-3 is the lastest radar altimetry satellite whose data can be added to the time series currently being compiled.

We present an investigation into how well Sentinel-3A data continues the known trends in surface elevation change observations spanning Antarctica for the past 25 years. The results show the relationship between Sentinel-3 observations and previous timeseries of observations from CryoSat-2, which is still operative, and continues a series of overlapping land ice radar altimetry missions starting with ERS1, ERS2 and Envisat [1].

Only approximately 2 years of Sentinel-3A data are available with the satellite in a SAR mode configuration supporting long-term land ice surface elevation studies, so this poster is necessarily a first look. To properly investigate surface elevation trends regions are split into a set of grid cells and a time-varying quadratic surface model is fitted to each [2]. This must take account of the observing satellite's heading and corrections for glacial isostatic adjustments in the land mass. It must cover a long enough time period to remove seasonal effects. Further, corrections must be made for variable backscatter power in the radar altimeter measurements, using another time- and heading-varying model, and inter-mission calibration also requires a time period sufficient to establish an elevation change trend [3]. That is why it is only now that these investigations can be started.

This work is a further contribution to an evolving body of results made at first for the ESA Climate Change Initiative, and contributing to the World Climate Research Program Global Sea Level Budget Group [4]. Those results have been used to close the sea level budget to within 0.3mm/yr, while acknowledging that further details of the land water storage component are needed, which this work is intended to aid.

References:
[1] The IMBIE team, Mass Balance of the Antarctic Ice Sheet from 1992 to 2017, Nature 558, 219-222 (2018), https://doi.org/10.1038/s41586-018-0179-y
[2] Malcolm McMillan et al, A high‐resolution record of Greenland mass balance, GRL, 16th June 2016, https://doi.org/10.1002/2016GL069666
[3] Lin Gilbert et al, Surface Elevation Change over Ice Sheets – combining 25 yrs of Radar Altimeter Observations, poster at "25 Years of Progress in Radar Altimetry" Symposium 24th-29th September 2018, proceedings in press.
[4] https://www.wcrp-climate.org/news/science-highlights/1363-gsb-1993-present, as of 6th October 2018.

Area D - Board 439

The Ice, Cloud, and land Elevation Satellite – 2 (ICESat-2) observatory was launched on 15 September 2018 to measure ice sheet and glacier elevation change, sea ice freeboard, and enable the determination of the heights of Earth’s forests. ICESat-2 current orbit inclination allows ATLAS to collect data between 88 degrees north latitude and 88 degrees south latitude from nominally 500km elevation above Earth's surface. ICESat-2’s laser altimeter, the Advanced Topographic Laser Altimetry System (ATLAS) uses green (532 nm) laser light and single-photon sensitive detection to measure elevation along each of its six beams ten thousand times per second. In this presentation, we describe the major components of the observatory and the ATLAS instrument. We summarize the first six months of on orbit data collection and present the status of the observatory and the ATLAS instrument. We present on the status of the lower-level data products including the Level-2A data product (ATL03), which provides the geodetic location (i.e. the latitude, longitude and elevation) of the ground bounce point of photons detected by ATLAS. The ATL03 data product is the primary product used for higher-level (Level 3A) surface-specific data products such as glacier and ice sheet elevation, sea ice freeboard, vegetation canopy height, ocean surface topography, and inland water body elevation. This presentation will also present the plans for future data collection, the geolocation uncertainty of the ATL03 global geolocated photon data product, the status of data product availability, and plans for data reprocessing.

Area D - Board 440

The Arctic Ocean is a challenging region, because of its complex and not well-documented bathymetry, together combined with the intermittent presence of sea ice and the fact that the in situ tidal observations are sparse at such high latitudes. The existing bathymetry map of the Arctic is derived from combination of sparse ship sounding data, particularly on latitudes higher than 81° (e.g. IBCAOv3) as well as contoured pre 1990 soviet data. The gaps (or data holes) are filled by extrapolation and is consequently somehow inaccurate.
Smith&Sandwell combined the ship sounding data with bathymetry predicted from gravity up to latitude 81° to fill the gaps. However, on even higher latitudes, the gravity to bathymetry prediction was never attempted due to the lack of high resolution (e.g. altimetry derived) quality marine gravity maps.
However with the availability of particularly SAR altimetry from Cryosat-2 altimetry the progress in accuracy of Artic marine gravity has been huge. This is largely due to the SAR altimetry data which enables gravity field prediction in leads to much higher accuracy than possible with conventional altimetry. As part of the ESA CryoSat Plus for Oceans (CP4O) project, the Arctic bathymetry will be predicted from the DTU17 marine gravity, which showed massive improvement compared to the previous models (e.g. DTU15) in the Arctic. Significant adaptation of the method by Smith & Sandwell has to be made in order to account for particularly inaccuracies at longer wavelength.
The aim of this work is to demonstrate the contribution of marine gravity for the bathymetry prediction in the Arctic. With this presentation, the methodology to develop the gravity to bathymetry prediction will be presented. Due to the presence of thick sediments on the seafloor, the gravity and seafloor topography correlation is not significant in some areas. Such areas will be flagged and the accuracy of the predicted bathymetry will be evaluated.

Area D - Board 441

Abstract: In recent years, there has been a large focus of the Arctic due the rapid changes of the region. The sea level of the Arctic Ocean is an important climate indicator. The Arctic sea ice is decreasing and has since 1997 experienced a steepening in the decrease. The Arctic sea level determination is challenging due to the seasonal to permanent sea ice cover, the lack of regional coverage of satellites, the satellite instruments ability to measure ice, insufficient geophysical models, residual orbit errors and challenging retracking of satellite altimeter data.
We present the DTU/TUM 25-year sea level record based on satellite radar altimetry data in the Arctic Ocean. We are using satellite data from the European Space Agency: ERS-1, ERS-2, Envisat and CryoSat-2 (present) satellites. This sea level product is contributing to the ESA CCI Sea level budget closure, and is of version 2.
We use updated geophysical corrections and a combination of altimeter data: REAPER (ERS-1), ALES+ retracker (ERS-2, Envisat), combined Rads and DTUs in-house retracker LARS (CryoSat-2). REAPER is a reprocessed ERS-1 product from ESA. The ALES+ is an upgraded version of the Adaptive Leading Edge Subwaveform Retracker that has been developed to improve data quality and quantity in the coastal ocean, without degrading the results in the open ocean. ALES+ aims at retracking peaky waveforms typical of lead reflections without modifying the fitting model used in the open ocean. From LARS we are using a dedicated lead retracking system.
There are very few tide gauges in the Arctic Ocean, and no long term in-situ measurements in the central Arctic or in the sea ice covered areas to validate the altimeter data against. The ESA CCI DTU/TUM sea level record is compared with tide gauges available and show good results when correcting for glacial isostatic adjustments. Finally, we discuss the seasonal and regional variations over the past 25 years. The sea level trend in the Arctic Ocean is increasing, but not as fast as the mid latitude sea level.

Area D - Board 442

Satellite-borne radar and laser altimetry have been used to record surface elevation over the contemporary ice sheets. Previously, conventional radar altimeters have performed poorly over high sloping terrain with heterogenous topography. However, the novel synthetic aperture radar interferometric (SARIn) mode of CryoSat-2, launched in 2010, has improved capability in these regions, principally over areas experiencing the largest changes around the ice-sheet margins and outlet regions. But CryoSat-2 employs its low resolution mode (LRM) over the ice-sheet interior, which fails to detect small-scale changes both spatially and temporally, such as seasonal and local accumulation and melt events.

Sentinel-3A, launched in February 2016, is the first satellite to use synthetic aperture radar (SAR) across the interior of the GrIS. This has improved the along-track resolution (sampling) to approximately 300 m, compared to CryoSat-2’s LRM footprint which has a diameter approximately equal to 1.65 km. Although the ground track separation is smaller for CryoSat-2 due to its 369-day repeat cycle, Sentinel-3’s 27-day repeat cycle provides high temporal resolution allowing for elevation changes to be calculated at monthly time intervals.

Here, we investigate the performance of the Sentinel-3A SAR altimetry mission across the GrIS. Using repeat-track and crossover analysis for observed elevation changes, we produce a monthly time series of the elevation change, describing seasonal (intra-annual) change over the interior of the ice-sheet (above the equilibrium line altitude) from June 2016 to December 2018. We also investigate the sensitivity of our results to different spatio-temporal statistical interpolation methods for the presented time-series and compare results with CryoSat-2 LRM over the ice-sheet interior.

Additionally, we explore a preliminary time-series incorporating data from Sentinel-3B (launched in April 2018). Combining data from Sentinels-3A and -3B reduces the ground track separation distance by half compared to using only Sentinel-3A elevation products. We quantify the improvement in precision of the elevation change time series by including Sentinel 3A and 3B data.

Area D - Board 443

The Sentinel-3A and Sentinel-3B satellites were launched in 2016 and 2018, respectively, each carrying onboard a Ku-band Synthetic Aperture Radar (SAR) altimeter. With these satellites reaching a latitude of ~81.5 degrees and operating with a repeat period of 27 days, they offer the opportunity to provide continuous measurements of surface topography and volume change across most of the Antarctic and Greenland Ice Sheets, thereby extending the existing 25 year altimeter record for decades to come. The global operation of Sentinel-3 in SAR mode differs from all past Ku-band instruments; for the first time SAR measurements are routinely acquired across the interiors of the ice sheets; however unlike CryoSat-2 there is no interferometer to aid signal retrieval in regions of complex coastal terrain. In view of these differences and the novel characteristics of the Sentinel-3 system, early assessments of the performance of the instrument are required, to evaluate the satellite’s utility for monitoring Earth’s Polar regions.

Here, we present analysis of data acquired by Sentinel-3A during the two and a half years of routine operations, to assess the performance of the SAR altimeter to date. We focus both on inland ice sheet regions, where Sentinel-3 provides the first operational SAR altimeter measurements, and also on coastal areas with more complex topography. We firstly implement our own Level-2 processing to investigate the performance of several empirical retrackers. We then assess the performance of the mission for determining ice sheet elevation by (1) evaluating instrument shot-to-shot precision over the Lake Vostok validation site in Antarctica, (2) assessing absolute accuracy through comparisons with airborne reference data from Operation IceBridge campaigns, and investigating the dependency upon ice sheet characteristics such as surface slope, and (3) evaluating Antarctic-wide measurement repeatability using a single-cycle cross-over approach. We conclude by assessing the capability of Sentinel-3, after its first two years of routine operations, to measure ice sheet elevation change, and to resolve signals of glaciological imbalance.

Area D - Board 444

For the past 25 years, polar-orbiting satellite radar altimeters have provided a valuable record of ice sheet elevation change and mass balance. One of the principle challenges associated with radar altimetry comes from the relatively large ground footprint of conventional pulse-limited radars, which reduces their capacity to make measurements in areas of complex topographic terrain. In recent years, progress has been made towards improving ground resolution, through the implementation of Synthetic Aperture Radar (SAR), or Delay-Doppler, techniques. In 2010, the launch of CryoSat-2 heralded the start of a new era of SAR Interferometric (SARIn) altimetry. However, because the satellite operated in a combination of SARIn and Low Resolution Mode (LRM) over the ice sheets, many of the non-interferometric SAR altimeter processing techniques have so far been optimized for water and sea ice surfaces only. The launch of Sentinel-3, which provides full non-interferometric SAR coverage of the ice sheets, therefore presents the opportunity to further develop these SAR processing methodologies over ice sheets, for the purpose of scientific exploitation of this operational mission.

Here we present results from SPICE (Sentinel-3 Performance Improvement for Ice Sheets), a 2 year study that has been funded by ESA’s SEOM (Scientific Exploitation of Operational Missions) programme. The aim of SPICE is to contribute to the development and understanding of SAR altimetry processing over ice sheets, through the emulation of Sentinel-3 data from dedicated CryoSat SAR acquisitions made at the Lake Vostok, Dome C and Spirit sites in East Antarctica, and from reprocessed SARIn data in Greenland. More specifically, SPICE has focused on (1) developing Sentinel-3 SAR altimetry processing methodologies over the Polar ice sheets, and (2) investigating radar wave penetration through comparisons of Ku- and Ka-band satellite measurements. In Phase 1 of the project we have evaluated existing processing methodologies, and in Phase 2 we have investigated new evolutions to the L1b Delay-Doppler Processing and L2 retracking chains. In this presentation we focus on (1) the novel L2 processing developments used to improve retracking over complex topographic surfaces, such as those found across the margins of the Antarctic Ice Sheet, and (2) the impact of volume scattering on SAR retrievals, through a comparison of SAR waveforms to equivalent Ku- and Ka-band LRM acquisitions.

Area D - Board 445

This abstract presents the last advances of Thales Alenia Space to support ESA in the definition of the radar altimeter for the P-ICE (PolarIce) mission (ongoing phase A/B1) in the frame of the Copernicus Expansion program. The main objective of this work is to allow the continuity of the Cryosat-2/Siral-2 mission which is currently operated in orbit and to add significant enhancements for ice topography measurements.

The radar instrument so-called IRIS (Interferometric Radar altimeter for Ice and Snow) includes not only Ku-Band as the Siral-2 instrument but also Ka-Band with the aim to retrieve the snow depth on sea ice surfaces to take account of the snow loading effect. Interferometry will be provided in Ku-Band as for Siral-2 whereas Ka-Band is operated with a single receive channel.

For this new instrument, the LRM mode is dropped being substituted by the SAR Closed-Burst (CB) mode over open ocean and coastal zones to improve range noise performances. An enhanced version of the SAR Interferometric (SARIn) closed burst mode, targeted for Land Ice and Glaciers surfaces, is also included by providing 4 times as many measurement bursts as Cryosat-2 to improve height and phase accuracies over these surfaces. At last, the SARIn Open-Burst mode operated at a PRF of 18 kHz will be used over sea ice surfaces in order to improve the azimuth resolution down to the fully-focused SAR performance and to provide the angle of arrival of sea ice leads. Ka-Band pulses are added on top of the Ku-Band chronograms for all the modes without affecting the original Ku-Band performances. Ku- and Ka-Band pulses are transmitted and received simultaneously in order to maximize the number of transmitted pulses and to have co-located Ku/Ka echoes.

Another major evolution of IRIS wrt Siral-2 is the use of Ku/Ka chirps with 500-MHz bandwidth in order to improve the accuracy in retrieving sea ice freeboards. In addition, the instrument will provide improved tracking performances by tracking Points of Closest Approach (POCAs) on slopes of up to 5° by a smart combination of Closed-Loop mode over small slopes and Open-Loop mode over higher slopes. At last, the antenna stability will be improved wrt Siral-2 in order to obtain a better accuracy in the retrieval of the across-track angle of arrival (20” vs. 35 arcsec) when interferometry in Ku-Band is enabled. For this, the optimisation of the CFRP structure of the antenna baseplate supporting the two reflectors is being studied to provide maximum stability of the antenna sub-system between two external calibrations .

Regarding the central electronics, a compact design based on the Poseidon-4 altimeter heritage and new highly integrated architectures used for telecom satellites is studied for the IRIS instrument. In particular, the GaN technology is being analysed for the Ku-Band SSPA as well as RF on PCB technology for Ku/Ka RF sections.

As a conclusion, the IRIS instrument achieves significant scientific progress wrt Siral-2 while it includes the most up to date technologies releasing thus an optimum design.

Area D - Board 446

Ice sheet height changes derived from satellite repeat altimetry by local least-squares fitting methods depend on the parameters introduced. In addition, incomplete spatial sampling by satellite altimetry requires interpolation and possibly smoothing.

In Greenland the biggest height changes are distributed along the coast, where steep ice sheet slopes and rock outcrop shape the surface. CryoSat-2 operates in SARIn mode in these areas, providing high spatial resolution. As individual height measurements in a repeat altimetry box have to be topographically corrected to the box center, the local topography model needs to be able to capture the complex surface, which cannot always be solved successfully. These challenges emphasize the importance of carefully applied interpolation.

By simulating CryoSat-2 measurements in SARIn mode over the Northeast Greenland Icestream we assess the influence of choices in the process of height change derivation. With real data available for comparison, we can make additional statistical assessments and identify shortcomings of the different approaches.

Previous studies investigate the influence of modeling seasonal changes or signal parameters. We focus on the repeat altimetry box size and local topography modeling in the boxes on the height change results. Bigger box sizes include many observations and allow higher complexity in topography modelling, while they may overlook individual height change features. We use up to nine parameters for topography modeling as well as different DEMs in combination with box diameters up to several kilometers.

The resulting linear height changes are interpolated on a regular grid. Besides Ordinary Kriging and Inverse Distance Weighting, we apply Heterogeneous measurement-error Filtered Kriging. This method uses individual error variances of the input data to determine the weights in the process of interpolation.

We highlight the differences of the resulting height changes, including their uncertainty estimates. Our research restrains the parametrization of suitable repeat altimetry approaches and shows improvement of the final result by including the a posteriori errors of the repeat altimetry method into the interpolation.

Area D - Board 447

Climate warming has caused drastic changes in the cryosphere, which is impacting rates of sea level change. Repeat observations by satellite altimetry over the last 25 years has allowed to quantify changes to the mass of ice sheets and to understand the processes that link the cryosphere, atmosphere and the ocean. Interactions at the ice-ocean boundary is particularly important, due to suggestions that ocean variations directly impact ice discharge. At the ice-ocean boundary, ice shelves play a key role by buttressing inland ice; ocean forcing has caused many of them to melt, thin or collapse, leading to an increase in the flow of grounded ice into the ocean.

Recent large scale altimetry studies have been powerful in resolving ice shelf thinning at spatial scale of 10’s of kilometres. Whilst this is sufficient to resolve bulk thinning, it does not resolve the spatial details of elevation change which relates to ocean circulation in the cavity and can be critical for the stability of ice shelves and of grounded ice. Observations of elevation and change are often diverse in nature, accuracy and have varying spatial and temporal coverage. Therefore, building a comprehensive record of elevation and ice shelf melt requires development of methods in conjunction with applying strategies to account for measurement diversity.

Here we propose, design and apply new strategies for deriving robust high-resolution elevation changes from a combination of observations from diverse imaging modes, accounting for data bias, noise and resolution differences. We will present preliminary results of time-dependant elevations over the Dotson Ice shelf taking into account variability in surface interactions and errors in space and time. Additionally, we intend to quantify elevation change over key regions and link this change to external forcing with the intention of improving our understanding of the current state of the cryosphere and to deduce these regions melting rates; hence understanding their contribution to sea level change.

Area D - Board 448

The Copernicus Space Component (CSC) includes a series of seven space missions called 'Sentinels’. The so called “CSC Expansion” programme includes the new missions that have been identified by the European Commission (EC) as priorities for implementation in the coming years by providing additional capabilities in support of current emerging user needs. Six candidate missions have been identified to fulfill those needs. Polar Ice and Snow Topography Mission is one of these missions to answer the polar regions monitoring needs.
The objectives of Polar Ice Mission are to provide enhanced land ice elevation and sea ice thickness measurements. It will implement high spatial resolution measurements for improved lead detection and additional capability to determine snow loading on sea ice. Phase A/B1 studies are on-going to analyze and design the most appropriate mission concept to meet the observation and performance requirements.
In order to assess expected Altimeter and radiometer System Performance, an Instrument and System Performance Simulator (SIRS), and a L0/L1b/L2 Ground Processor Prototype (GPP), are being designed and developed under a OHB/ESTEC contract.
Polar Ice mission is being designed to take advantage from the heritage of Cryosat-2, Sentinel-3, SARAL and Sentinel-6 missions and to implement new capabilities by exploiting synergies between different measurements: Simultaneous Ku and Ka band, interferometry over sea ice and land ice, interleaved mode and interferometry …
We propose in this talk to present the End-to-End simulator and the ground prototype processor functions including the new algorithms to account for the new advances of this missions in comparison of CryoSat-2, Sentinel-3, AltiKa, Sentinel-6.

Area D - Board 449

The main difficulties to retrieve sea ice thickness (SIT) in the southern ocean come from the lack of in-situ observation and knowledge related to sea ice. For instance, whereas polar expeditions and in-situ observations over the Arctic have enabled to construct snow depth climatologies (e.g the Warren climatology), there are no equivalent data available over Antarctica. By consequence, except for a few studies such as Zwally et al, 2008 or Kurtz et al, 2012, based on ICESat data, sea ice thickness estimations over Antarctica nearly remain nonexistent and no valid sea ice volume estimations have yet been drawn (SI-CCI-2015 report). The objective of this presentation is to review our recent developments leading towards sea ice thickness estimations over Antarctica.
First, we detail the methodology used to derive sea ice freeboard from altimetric power echo measurements and we present a 2002-2017 Envisat/Cryosat-2 sea ice radar freeboard time serie. The continuity between these 2 satellites is ensured by a re-calibration of the Envisat Low Resolution Mode (LRM) on the Cryosat Synthetic Aperture radar (SAR) mode we describe in this presentation. Thereafter, the computation of sea ice thickness from sea ice freeboard require snow depth fields. Recently, the ESA CryoSeaNICE project demonstrated that snow depth is one of the most important obstacle to compute sea ice thickness. Meanwhile, Guerreiro et al, 2016, has showed the ability to retrieve snow depth from CryoSat-2 Ku and Saral/AltiKa Ka radar frequencies. Based on this approach, we present in a second part the first bi-frequency altimetric snow depth product over Antarctica computed from the Cryosat-2 Pseudo-LRM ESA product. The relevancy of the solution will be assessed by comparison with ICESat and Operation Ice Bridge (OIB) laser measurements. From these results, a 2013-2018 sea ice thickness time serie will be presented and assessed. Finally, we briefly explain how these sea ice products will be used to better understand recent sea ice variations and to derive sea ice volume estimations.


References
Guerreiro, K., Fleury, S., Zakharova, E., Rémy, F., & Kouraev, A. (2016). Potential for estimation of snow depth on Arctic sea ice from CryoSat-2 and SARAL/AltiKa missions. Remote Sensing of Environment, 186, 339-349.

Kurtz, N. T., & Markus, T. (2012). Satellite observations of Antarctic sea ice thickness and volume. Journal of Geophysical Research: Oceans, 117(C8).

Wally, H. J., Yi, D., Kwok, R., & Zhao, Y. (2008). ICESat measurements of sea ice freeboard and estimates of sea ice thickness in the Weddell Sea. Journal of Geophysical Research: Oceans, 113(C2)

Area D - Board 450

The grounding zone is the transition region between grounded and floating ice. It is the boundary at which the ocean has the most significant influence on the ice sheet. Its position, dynamics and structure are critical for mass budget calculations and as an input into numerical models. Moreover, monitoring the location of the grounding line in areas where there has been dynamic thinning and grounding line retreat (e.g. within Amundsen Sea Sector of West Antarctica) is critical in assessing the stability of the ice sheet.

The most reliable approach to map the grounding zone remotely is to measure the limit of tidal flexure of the ice shelf (point F) using differential synthetic aperture radar interferometry (DInSAR) or ICESat repeat-track measurements. These methods can also be used to characterise stress gradients and ice rheology by measuring the tidal flexure across the grounding zone. However, these methods are yet to provide satisfactory spatial and temporal coverage of the whole of the Antarctic grounding zone. We can also detect the grounding line by mapping the break-in-slope between the relatively flat ice shelf and grounded ice. This method improves coverage but can be unreliable in low sloping regions where the break-in-slope is difficult to detect.

CryoSat-2 provides a new opportunity to map the grounding zone. The SAR-interferometric (SARIn) mode, used over the margins of the Antarctic Ice sheet, provides improved performance compared to conventional altimetry, particularly over breaks-in-slope commonly associated with the grounding zone. Here, we use a combination of CryoSat-2 standard and swath elevation data to map point F and the limit of hydrostatic equilibrium (point H) using a new method based on detecting the tidal signal in pseudo crossovers.

We first discuss the results of applying this method to the Antarctic grounding zone and show that the mapped grounding line is in good agreement with previous observations from DInSAR and ICESat measurements. We then present the results of the spatial distribution of the grounding zone width (the distance between points F and H). This can provide insights about its physical structure, and we investigate how the grounding zone width is affected by several factors that are known to contribute to its overall deformation.

Area D - Board 451

Pine lsland Glacier (PIG) and Thwaites Glacier (THW) in the Amundsen Sea Embayment are among the fastest changing outlet glaciers in West Antarctica. Over the last decades both glaciers have accelerated, retreated and thinned rapidly and if these changes continue, a sufficient weakening or loss of their ice shelves may result in marine ice sheet instability with substantial grounding line retreat and large consequences for global sea level. Yet, assessing how much, how fast the ice in the Amundsen Sea Embayment will change, remains a major uncertainty.

In this study, we use time series of multi-source satellite imagery (i.e. Landsat, Sentinel-1, Sentinel-2, Aster, Cryosat) to show the rapid development of damage areas on the PIG and THW ice shelves that appear crucial to the future of both glaciers, but are not included in current models of their retreat. These damage areas consist of highly crevassed areas near the grounding line and of open fractures that contain dense ice melange within the shear zones of both ice shelves. Satellite observations over the past two decades show the evolution from lack of crevasses in 1997 to rapidly growing crevasse-damaged areas near the grounding line and in shear zones on both ice shelves in 2018 (Figure 1).

The observed damage compromises the integrity of both PIG and THW ice shelves, similar to the preconditioning of the shear zones of Larsen B prior to its collapse. An example of this preconditioning can be seen in the new rift that developed across PIG ice shelf in September 2018 and that resulted in unprecedented retreat of the ice shelf front in October 2018. Moreover, it shows that the weakening of the ice shelf margins, which was previously shown in a model to be key for THW's collapse, is already underway for THW's ice shelf.

To assess the importance of the damage in the shear zones, a continuum damage model (CDM) was coupled to the BISICLES ice sheet model. The model was applied on an idealized marine ice sheet geometry with retrograde bed slopes and strong lateral stresses, following the MISMIP+ setup which has strong similarities to PIG conditions. Several time-dependent simulations were carried using the CDM-BISICLES model with different ocean-induced melting and damage parameterizations to assess the importance of damage in the shear zones relative to ocean melting. In this framework, several damage scenarios were implemented by locally enhancing damage in the shear zone at the grounding line.

The model results highlight the importance of damage in the shear zones as driver for ice shelf instability as they precondition the ice shelves with structural weak zones that start at the shear zone near the grounding line and are advected downstream. The model results also indicate enhanced thinning at the grounding line and in the shear zones. Damage in shear zones reduces ice-shelf buttressing, leading to accelerated ice transport to the ocean, and consequent thinning of grounded ice upstream. Finally, the model shows a speedup of the glacier tongue as a result of the damage, which results in an increase in strain rate and in velocity gradients across the shear zone similar to the increasing velocity gradients across PIG and THW observed by satellites since 1992.

The damage also has an important impact on the modeled grounding line retreat as the enhanced damage scenarios in the model initiate a grounding line retreat equivalent to quadrupling the oceanic melting near the grounding line from 25 to 100 m/year. This illustrates that weakening these glaciers at their most vulnerable locations is a very effective way of introducing grounding line retreat, increased ice flux and hence mass loss.

The combination of these satellite observations and model results suggest that damage processes are key to ice shelf stability, grounding line retreat, and future sea level contributions from Antarctica. Moreover, they underline the need for incorporating damage processes in models to improve sea level rise projections.

Area D - Board 452

The majority of meteoric ice that forms in Antarctica leaves the ice sheet through floating ice shelves, many of which have been thinning substantially over the last 25 years. A significant proportion of ice-shelf thinning has been driven by submarine melting facilitated by increased access of relatively warm modified Circumpolar Deep Water to sub-shelf cavities. This mass reduction is significant due to the role ice shelves play in buttressing ice sheets against unstable retreat, and it is important that the role of future ocean warming in Antarctic stability be understood. Under some ice shelves, concentrated melting is found near the grounding line and along inverted channels. These channels guide buoyant melt-laden outflow, which can lead to localised melting of the sea ice cover. The channels may also potentially lead to heightened crevassing, which in turn affects ice-shelf stability. Meanwhile, numerical studies suggest that buttressing loss is sensitive to the location of ice removal within an ice-shelf. Thus it is important that we observe spatial patterns, as well as magnitudes, of ice-shelf thinning, in order to improve understanding of the ocean drivers of thinning and of their impacts on ice-shelf stability.

Here we present high-resolution altimetry measurements acquired between 2010 and 2018 under a number of ice shelves in West and East Antarctica. We show that a significant proportion of these ice shelves undergoes thinning along narrow regions of a few kilometre-wide and distributed across the ice shelves. We show that thinning is driven by basal melting and discuss the role of the ocean, of sub-glacial discharge, and the impact of localised thinning on ice shelf stability. We also discuss the abilities and challenges of ocean models in representing sub-ice shelf melt rates. We find that a number of factors influence the accuracy of modelled melt, with bathymetry being the leading factor; but this accuracy is only important in isolated regions under the ice shelves, such as shear margins and grounding lines. Thus representations of ocean physics, as well as ice physics, must be improved in these critical regions.

Area D - Board 453

In this study, satellite-derived observations of sea ice concentration, drift, and thickness are combined to provide a climatology and inter-annual monthly variability of the sea ice volume budget over the growth season (November-April) for the CryoSat-2 period Octobre 2010 to April 2018. This allows the first wintertime observational decomposition of the dynamic (advection and divergence) and thermodynamic drivers of ice volume change.

Dynamic and thermodynamic processes will be separated by applying similar methods to Holland and Kwok (2012), which decomposed the governing equation of ice concentration, C:

[1] ∂C/∂t + ∇.(uC) = f_C - r

where u is ice drift motion, f_C is thermodynamic freezing or melting, and r is the concentration change from mass-conserving mechanical ice redistribution processes which convert ice area to ice thickness, such as ridging and rafting. ∂C/∂t represents ice intensification and ∇.(uC) represents ice flux divergence, the dynamic contribution to ice concentration.

Equation 1 can be rearranged and the dynamic contribution, ∇.(uC), expanded to show the contributions from advection, (-u.∇C) and divergence, (-C∇.u), determining four terms:

[2] ∂C/∂t = -u.∇C - C∇.u + f - r

In this sign convention, the positive of each term corresponds to an increase in sea ice intensification.

The governing equation of the volume budget is of the same form but combines the thickness and concentration data:

[3]
∂Ch/∂t = -u.∇Ch - Ch∇.u + f_Ch

where h is the thickness of the ice and the resulting product of the two datasets, Ch, is the effective thickness. If Ch were to be multiplied by the grid cell area this would give the volume, V. It is not necessary to take this step because the area remains constant and does not influence the relative values of the terms. However, when deriving an Arctic-wide climatology spatial integration across the grid cell areas is required. Ridging is taken into account by effective thickness change, therefore, it is not included in the calculations.

The method is replicated using model simulations from the Centre for Climate Observation and Modelling (CPOM)-modified Los Alamos sea-ice model (CICE), providing a test of the model’s ability to calculate the volume budgets but also identifying unrealistic growth regimes in the CryoSat-2 observational datasets. Sensitivity to several observational datasets is performed to provide an estimated uncertainty of the budget calculations.

The observational results show ice gain in the central Arctic is dominated by ice freezing with contributions from convergence. Divergence at the coastlines of the Arctic form an ice sink where freezing generates new ice. Advection is shown to drive ice equatorward and induce melting at the ice edge where ice becomes thermodynamically unstable. The dynamic components are found to grow in influence throughout the growth season.

The 2016/17 winter growth season budget shows reduced thermodynamic intensification and stronger dynamic tendencies which may be in response to thin initial ice and an exceptionally warm winter. Compared to the observed volume budget, the CICE model displays similar patterns of thermodynamic freezing, however, dynamic components in the central Arctic are significantly reduced whilst they are over-amplified at the ice edge.

Area D - Board 454

In July 2012 an unprecedented proportion (98.6%) of the surface area of the Greenland ice sheet experienced melt as increased transport of warm air from the south, associated with a negative North Atlantic Oscillation pattern, created anomalously high near surface temperatures. For high altitude regions of the ice sheet interior (approximately > 2800 m.a.s.l.), this was the first time temperatures had risen high enough to force surface melt since satellite records began in 1979. Surface melt on the Greenland ice sheet induces poorly understood changes in snowpack scattering characteristics, which complicate elevation retrievals from radar altimetry and introduce considerable uncertainty into estimates of ice sheet elevation change and mass balance. Here, we use a waveform deconvolution model which retrieves the depth-distribution of radar backscatter from low-resolution mode CryoSat-2 satellite altimeter waveforms. Using this technique, we map spatial and temporal variability in Ku band radar backscatter and penetration depth across the interior of the Greenland ice sheet between 2011 and 2017. From these waveform inversions we identify the extreme surface melt in the summer of 2012 as a unique event in the altimeter record, which abruptly resets the radar scattering horizon to close to the air-snow interface. As a result, we estimate a reduction in penetration depth of 2.23 ± 1.41 m on average at surface elevations greater than 2000 m. Subsequently, due to the accumulation of snowfall on the ice sheet surface, we observe a continuous increase in radar penetration depth until the end of 2017, where we find the scattering horizon to be at a depth of 3.28 ± 1.13 m, approximately 50 cm above that recorded before July 2012 (3.79 ± 1.12 m). We investigate the ability of direct knowledge of radar penetration depth to inform ice sheet elevation change derived from radar altimetry and find that for shorter time intervals, accounting for correlated changes in elevation and penetration depth allows us to better reconcile trends from CryoSat-2 with independent measurements from NASA Operation IceBridge airborne laser altimetry.

Area D - Board 456

By measuring directly the sea ice freeboard, satellite altimetry has proven to be a powerful tool to estimate sea ice thickness from space for the last 25 years.
Since 2010, the ESA’s CryoSat-2 (CS2) radar altimetry mission has made a significant contribution to this field. Compared to previously launched missions, CS2 enabled the observation of Arctic sea ice at latitudes up to 88N at a higher along-track ground resolution, when operating in SAR and SAR Interferometric (SARIn) modes. Previous studies showed how the phase information available in the SARIn mode can be used to reduce the uncertainty of the sea surface height in ice-covered regions (Armitage and Davidson, 2014) and, consequently, the uncertainty of freeboard retrievals (Di Bella et al., 2018).
This work investigates how retracking more than one peak in level 1b SARIn waveforms – in combination with the respective phase information – could increase the amount of valid sea ice as well as sea surface height measurements, so as to ultimately reduce even further the uncertainty of freeboard estimates. The challenges and limitations of this method will be identified with the support of Sentinel-1 SAR images and airborne validation data from the ESA CryoVEx and NASA Operation IceBridge campaigns.


References:
Armitage, T. W. K., & Davidson, M. W. J. (2014). Using the interferometric capabilities of the ESA CryoSat-2 mission to improve the accuracy of sea ice freeboard retrievals. IEEE Transactions on Geoscience and Remote Sensing, 52(1), 529–536. http://doi.org/10.1109/TGRS.2013.2242082

Di Bella, A., Skourup, H., Bouffard, J., & Parrinello, T. (2018). Uncertainty reduction of Arctic sea ice freeboard from CryoSat-2 interferometric mode. Advances in Space Research, 62(6), 1251–1264. http://doi.org/10.1016/j.asr.2018.03.018

Area D - Board 457

Satellite altimetry is one of the most important methods to monitor global changes in sea ice thickness. It can provide extensive spatial and temporal measurements of sea ice thickness through converting ice freeboard measurements to thickness by assuming hydrostatic Equilibrium. As a bistatic radar altimeter concept, Global Navigation Satellite System Reflectometry (GNSS-R) or PAsive Reflectometry and Interferometry System (PARIS) can perform altimetry along several tracks simultaneously and provide high density of surface height measurements. This passive altimetry concept can be also extended to sea ice, which could also benefit long term, global sea ice monitoring.
The presence of sea ice at the water surface significantly shifts the diffuse reflection limit and improves the phase coherence of L-band observations. Spaceborne GNSS-R data collected by the UK-DMC mission and TechDemoSat-1 (TDS-1) mission have shown that GNSS signals reflected from sea ice exhibit strong coherent characteristics. This makes it possible to derive more precise sea ice altimetry measurements from the GNSS-R carrier phase measurements. By processing the raw data streams collected by the TDS-1 mission over the Hudson bay (see Fig. 1a), high-precision sea ice altimetry measurements have been derived. The altimetric results show good consistency with the DTU15 mean sea surface model with a root-mean-square difference of a few centimeters (see Fig. 1b).
More important, the difference observed between the GNSS-R altimetric results and the MSS shows good correlation with the co-located sea ice thickness data from the Soil Moisture and Ocean Salinity (SMOS) mission, with lower altimetric height as the thickness increases (i.e., opposite sense than in laser and Ku band radar altimeters) (see Fig 1c and 1d). These features are consistent with the draft variations associated to the thickness variations and therefore consistent with the hypothesis that the dominant reflecting layer is the interface between the bottom of the ice and sea water underneath. This finding would open the possibility to measure ice thickness from space in a way complementary to the current approaches. An advantage of measuring the draft is that its uncertainty propagates into thickness estimates with a smaller factor than the freeboard uncertainties.
By processing the other 8 sets of raw data acquired with the UK TDS-1 GNSS-R receiver across the Arctic and Southern Oceans, the feasibility of GNSS-R phase altimetry technique over sea ice has been further tested with different specular reflection conditions, e.g. over sea ice with different thickness and at different elevation angles. We will present the results of the analysis of these data sets in this talk.

Area D - Board 458

Cryosat-2 has provided measurements of pan-Arctic sea ice thickness since 2010 with unprecedented spatial coverage and frequency. However, it remains uncertain how the Ku-band radar interacts with the vast range of scatterers that can be present within the satellite footprint, including sea ice with varying physical properties and multi-scale roughness, snow cover, and leads. Here, we present a numerical model designed to simulate delay-Doppler SAR (Synthetic Aperture Radar) altimeter echoes from snow-covered sea ice, such as those detected by Cryosat-2. Backscattered echoes are simulated directly from triangular facet-based models of actual sea ice topography generated from Operation IceBridge Airborne Topographic Mapper (ATM) data obtained during Arctic campaigns, as well as virtual statistical models simulated artificially. We use these waveform simulations to investigate the sensitivity of SAR altimeter echoes to variations in satellite parameters (height, pitch, roll) and sea ice properties (physical properties, roughness, presence of water etc.). We show that the conventional Gaussian assumption for sea ice surface roughness may be introducing significant error into the Cryosat-2 waveform retracking process. Compared to a more representative lognormal surface, an echo simulated from a Gaussian surface with rms roughness height of 0.2 m underestimates the ice freeboard by 5 cm – potentially underestimating sea ice thickness by around 50 cm. This model will ultimately be used to improve retrievals of key sea ice properties, including freeboard, surface roughness and snow depth, from SAR altimeter observations.


Figure Caption: Example of a Cryosat-2 echo modelled from real sea ice surface topography, including (a) ATM observations of the sea ice surface (with the extent of the pulse-limited footprint in black and lead locations highlighted by the yellow bars); (b) a lognormal fit to the surface height PDF; (c) the true Cryosat-2 SIRAL echo, and (d) the modelled echo (using the inset SIRAL antenna parameters). By separating the backscattering contributions of surface and volume scattering components, we are able to model the sub-component backscattered echoes from the air-snow interface, snow volume, snow-ice interface, and leads, shown by the different colours in (d).

Area D - Board 459

The detection of sea ice leads is a prerequisite for the estimation of ice freeboard and thickness from altimeter data. The classification of altimeter waveforms is generally performed using statistical parameters on the echo power or Machine Learning approaches directly on the waveforms. The validation of lead detector is often performed using in-situ or airborne campaign such as Operation Ice Bridge or CryoVEx. This classification process can be validated using other satellite measurements such as SAR images, for example those provided by S-1. This enables to cover different ice types, different seasons and larger regions in the Arctic compared to in-situ or airborne measurements. Better generalization capabilities can be expected with a validation/optimization using spaceborne SAR images.

In this presentation/paper, the following aspects will be detailed:
- A dedicated S-1 lead detector has been implemented. We adopted an approach based on the morphological gray-scale reconstruction which is adequate for the detection of local extrema corresponding to leads. The algorithm has been tested and optimized on a dataset of about 300 S-1 images in EW mode.
- A robust and consistent methodology for the joint assessment of SAR and Altimeter leads detector is provided. Compared with the current state of the art, it is proposed to fully account for the 2-D geometric problem when comparing the 1-D altimeter track and 2-D SAR image. The distance from nadir and the surface of SAR-detected leads within the altimeter footprint are considered.
- A set of 3 standard quantitative criteria is chosen (True Positive Rate, False Positive Rate, and Overall Accuracy). The population of sea ice samples is generally higher than that of leads for consolidated sea ice. An interpretation of the Overall Accuracy values may lead to false analysis, as it is strongly biased towards the most highly-populated class. In addition, finding a trade-off between false alarms and good detection is generally left to subjective interpretation with the TPR/FPR criteria. A fourth criterion is thus added to address these two issues: the Matthews correlation coefficient.
- Some analyses have been performed on a collocated dataset between S1 and SRAL S-3 with a time lag less than 8 minutes, equivalent to about 300 000 altimeter samples (over 160 SAR images). The overall accuracy between the altimeter and SAR detectors is good with correct agreement in about 90 % of cases. The False Positive Rate is very low with less than 3-6%. The Matthews coefficient is introduced and provides new potential parametrization for the altimeter detectors. A trade-off between false alarms and correct detection is thus found numerically. It is proposed to revise the values of the internal threshold used by the SRAL S-3 lead detectors (from 10 dB to 23 dB).
- In addition, the proposed methodology provides some insights into the potential returns from off-nadir. The best agreement is obtained when the distance to nadir considered is not more than around 800 meters. This means that beyond this distance, the returns from off-nadir scatterers become more and more negligible.

A comparison between S1 and SARAL AltiKa is also provided, with a collocated dataset of about 100 SAR images. The intrinsic performance of AltiKA versus S3 SRAL for lead detection is finally discussed.

Area D - Board 460

The Arctic is both most sensitive and a key player of the ongoing climate warming. In the Arctic, sea ice extent and thickness have shrunk substantially during the past decades, and most recent observations indicate that these changes are accelerating, outpacing the most pessimistic climate models predictions.
Models indeed need observations to refine their predictions. These are necessary for a better determination of the current state from which future changes are forecasted, but also to improve the parameterization of energy exchanges at the ocean-ice-atmosphere boundaries. Satellite observations provide a particularly relevant information both in terms of space and time coverage. But while sea ice extent is routinely monitored from space, remote sensing of ice-thickness, although promising, is still in its early stage. Dedicated missions such as CryoSat-2 require in situ observations to calibrate and validate sea ice thickness estimates.
Here we present observations gathered from the “ Ice-T ” (Ice-Thickness) buoy deployed every spring at the North Pole since 2011 as part as the North Pole Environmental Observatory together with complementary instruments, that drift in the direction of Fram Straight.
The Ice-T buoy, developed at LOCEAN since 2005, is an instrument for automated measurement of sea-ice thickness together with the terms entering the thermodynamical sea-ice mass balance. The instrument is intended both for thin and thick ice conditions. It is made of two bodies: a surface buoy trapped in the ice, and a subsurface weighting float (or “fish”), hanging below the ice layer, connected with a cable for data and energy transmission. The instrument currently measures the ice thickness, thermal profiles within the ice and snow layers, and allows as well to estimate the snow load. In addition we can accurately estimate ocean currents at the base of the ice, which ,combined with salinity measurements, makes it possible to estimate the ocean-ice heat flux. Additional sensors include a GPS and provide also basic meteorological parameters. All data are real time transmitted through the iridium communication system. A prototype meteorological mast (BEAR, Budget of Energy for Arctic Research) can be mounted on the top of the buoy to estimate turbulent and radiative heat fluxes, major components of the surface energy budget, which are needed to close the sea ice mass balance. Recent developments include an inertial motion unit which enables to estimate sea state in the presence of ice. Current developpment focus in particular on direct estimates of snow thichness embedding a millimeter-wave radar sensor.

Area D - Board 461

Several studies showed that altimeters are powerful tools to detect, characterize small (<10km²) icebergs but also to estimate the topography and thus volume of larger icebergs (>100km²).
In both conventional pulse limited (LRM) and more recent Delay Doppler altimeters (DDA) waveform data, small icebergs have detectable signature that can be used to detect and to estimate the iceberg's area (Tournadre et al 2007, 2012). Since 2015, the ALTIBERG project (funded by CNES) processes the archives of all past and present altimeters to create and maintain a data base of small icebergs distributed through the CERSAT website (http://cersat.ifremer.fr/user). The database contains the location, time, backscatter, area and range of the individual detected icebergs as well as monthly gridded fields of the iceberg probability of presence, mean area and volume of ice. Up to now this database was limited to the Southern Ocean. Taking advantage of a complete reprocessing (version-2) of the ALTIBERG database, designed primaryly to improve the sea ice flagging, the geographical coverage has been extended to the Northern Hemisphere especially around Greenland and within the Baffin Sea to produce a Northern Hemisphere small iceberg database.
Compared to the Southern Ocean the Greenland region poses specific problems that had to be solved. First while most of the region covered by icebergs in the Southern Ocean is located north of 66°S and is thus sampled by the Topex-Jason altimeter series, most of the Northern Hemisphere icebergs are located mainly north of 66°N. The Topex-Jason series that operates continuously on the same same ground sampling) since 1992 can thus not be used as the backbone for the inter-calibration of the detection and volume estimates from the different altimeters. The number of altimeters available at a given time is thus also more limited than that in the Southern Hemisphere.
In the Northern Hemisphere, icebergs tend to remain close (~some tens of km) to coasts that are characterized by an extremely large numbers of small islands and rocks that have altimeter signature similar to icebergs and can thus skew the results. These small land features are not well represented even in high resolution land masks. A tedious work of detection and control of those rocks and islands had to be conducted to create a rock-island database used to flag the detected targets to eliminate possible land contamination. Eleven altimeters have been processed to create the Northern Hemisphere data base.
As the Topex-Jason series could not be used for inter-calibration as it was done in the Southern Hemisphere, the inter-calibration was conducted using the ERS-ENVISAT-ALTIKA series to ensure the homogeneity and continuity of the time series. As for the Southern Hemisphere, the monthly mean probability of presence, the mean area of icebergs and the mean volume of ice have been computed on a regular polar grid.
The distribution of icebergs reflects the ocean circulation around Greenland. It mainly follows the East and West Greenland Currents along the coast of Greenland and the Baffin Current within the Baffin Bay. The analysis of the total volume of ice shows a constant increase of the volume of ice of icebergs from 1992 to 2017 that clearly reflects the increase of the ice loss of the Greenland Ice sheet.
For larger icebergs it has been demonstrated since Seasat (McIntyre & Cudlip, 1987) that altimeter can measure the oceberg elevation profile. Using large iceberg track data from the National Ice Center and the Bringham Young University it has been possible to collocate the altimeters track over iceberg ad to create a database of elevation profile covering the 2002-2014 period. This database of location, time and freeboard has been used to estimate the distribution of the volume of ice ad its temporal and spatial evolution around Antarctica (Tournadre et al, 2015). More recently, four altimeters (Cryosa, Altika, HY2A and Sentinel3) have been used to precisely map the topography of the giant A68 iceberg that calved in Jul2017 from the Larsen C iceshelf. This high resolution (1km) topography is comparable to the 1km resolution topography obtain by stereo pairs of images (REMA).

For both small and large icebergs, Cryosat-2 SARin mode data represents a new and promising tools. For small iceberga, the interferometric coherence and phase difference allows a better detection and the estimate of the area and surface topography of the iceberg at an hectometric scale. The processing of the Cryosat-2 archive around Greenland provides an unprecedented data set of small iceberg freeboard, area, and backscatter. For large iceberg, the swath processing of Cryosat SARin data allows to analyzed small scale (<100m) details of the surface topography that up to now could only be measured by stereo pair analysis.


McIntyre, N. F. & Cudlip, W. Observation of a giant Antarctic tabular iceberg by satellite radar altimetry Polar Rec., 1987, 145, 458-462.
Tournadre, J. Signature of Lighthouses, Ships, and Small Islands in Altimeter Waveforms J. Atmos. Oceanic Tech., 2007, 24, 1143-1149
Tournadre, J.; Girard-Ardhuin, F. & Legresy, B. Antarctic icebergs distributions, 2002-2010 J. Geophys. Res., 2012, 117, C05004o
Tournadre, J.; Bouhier, N.; Girard-Ardhuin, F. & Rémy, F. Large icebergs characteristics from altimeter waveforms analysis J. Geophys. Res., 2015, 120, 1954-1974
Tournadre Jean, Bouhier Nicolas, Boy F., Dinardo S. (2018). Detection of iceberg using Delay Doppler and interferometric Cryosat-2 altimeter data . Remote Sensing Of Environment , 212, 134-147 .
Figure : cryosat-2 SARin mode over LarsenC (future A68) on April 17th 2016. Top/left Surface elevation (m), top/right surface Backscatter (corrected for nadir angle), botton/left Landsat8 image, bottom/right elevation horizontal gradient (*10-²)

Area D - Board 462

For both small and large icebergs, Cryosat-2 SARin mode data represents a new and promising tools. Small icebergs (or target emerging from the sea surface) have a signature in the thermal noise part of the waveforms. As for SAR data these signature are bright spots that can be detected using classical image processing algorithms. The use of interferometric coherence allows to improve the detection. This is specially true for small icebergs whose backscatter signature is close to the noise level but whose coherence is close to 1. The phase difference over the iceberg allows to precisely positioned the iceberg and to estimate the iceberg elevation. The processing of the interferomtric data allows an estimate of the iceberg area, backscatter and elevation at hectometric scales (Tournadre et al 2018). The processing of the Cryosat-2 archive around Greenland provides an unprecedented data set of small iceberg freeboard, area, and backscatter. The analysis of this data set reveals the ouitflow from the major Greeland icebergs and the associated difference of freeboard and size.
Large Southern Ocean icebergs are mostly tabular and characterized by small variation of elevation, combining classical along track altimeter processing with Cryosat SARin Interferometric swath processing it is possible to estimate an iceberg surface topography at ~200m resolution. This allows to analyzed small scale details of the surface topography that up to now could only be measured by stereo pair analysis. These small scale change of elevation are related to fractures, pressure ridges, flow direction or rifts. SARin data alone or in combination with other altimeter data will allows to follow the evolution of large icebergs topography as they travel into the open ocean and gives thus a better insight into the melting and fracturing of large icebergs. This can not be easily done by images stereo pair analysis because of the iceberg displacement.


Tournadre Jean, Bouhier Nicolas, Boy F., Dinardo S. (2018). Detection of iceberg using Delay Doppler and interferometric Cryosat-2 altimeter data . Remote Sensing Of Environment , 212, 134-147 . https://doi.org/10.1016/j.rse.2018.04.037

Area D - Board 463

The Arctic region is an important component of the climate system and its exact influence is still not clearly understood today. For many years, radar altimeter measurements have been exploited in the Arctic ocean to retrieve the freeboard height from which the sea ice mass can be deduced. Sea Ice regions are very complex for radar altimeters because different surface types (water, sea ice, etc ..) are mixed in the altimeter footprint. Each altimetry mission has its own instrumental characteristics (different radar techniques, different radar frequencies, etc…) and specific processing algorithms have been developed to extract and monitor sea ice extent, sea level and freeboard height of the ice covered ocean.
When using empirical retracking approaches, sea ice processing must be adapted to each mission, which makes the combination of two or more satellites very complex. In the frame of different CNES and ESA projects, CLS has developed new sea ice lead identification methods, new retracking algorithms and innovative freeboard computation strategy based on a physical modelling of the signal to derive multi-mission (including multi-frequency) sea ice freeboard and SLA maps.. We present in our talk this homogeneous approach and its performances over sea ice regions.

Area D - Board 464

Sea ice thickness is one key indicator to understand the causes and consequences of Arctic Change. Climate data records (CDR’s) of Arctic sea ice thickness (SIT) with sufficient length are therefore a high priority data set for both climate research as well as operational activities and forecasts. While the CryoSat-2 mission was pivotal for establishing routine sea-ice thickness retrieval from satellite altimetry since 2010, its current data record is still too short to separate climate trends from inter-annual thickness variability. Significant research and developments efforts have therefore been made in the ESA Climate Change Initiative (CCI) to extent the SIT CDR by using CryoSat’s radar altimeter predecessor Envisat (2002-2012). The evolution from the pulse-limited radar altimeter RA-2 on Envisat to the SAR altimeter SIRAL onboard CryoSat-2 and the subsequent improvement in footprint size however poses a significant challenge for maintaining stability of the 15 year long SIT CDR. One issue linked to the radar altimeter type is preferential sampling for mixed surface types which are more often encountered in the larger Envisat footprint. A separate challenge is the required stability of auxiliary parameters such as snow depth on sea ice and ice density that are required for the conversion of the freeboard measurement of the altimeter into sea ice thickness.
We discuss results and error characterization for both generations of radar altimeter systems as well as current and future strategies to mitigate intermission biases and improve data record stability. These activities are essential to further extent the SIT data record into the past with radar altimeter data from the ERS-1/2 satellites and to ensure continuity with data from the Copernicus Sentinel-3 constellation. For the future, improvements on sea-ice thickness retrieval uncertainty are expected from the joint CryoSat-2/ICESat-2 data acquisition and the PICE Copernicus mission candidate, a concept for a next-generation polar radar altimeter.

Area D - Board 465

Sea level change in the Arctic is important in understanding how a warming climate is changing the dynamics of the Arctic hydro and cryosphere. Altimetry measurements by satellites have been available in the Arctic for more than 25 years. However seasonal and permanent sea ice is contaminating sea surface measurements and thus are sea level estimates often based on a poor data density or seasonal biased.

In order to compare ocean bottom pressure (OBP) from GRACE, which is unaffected by sea ice, with altimetric SSH, we need to add the steric sea level change to OBP. Studies in recent years (Morison et al 2012, Peralta-Ferriz et al 2014 & 2016 and Armitage 2016 & 2018) indicate that especially the halosteric sea level is a larger contribution to both temporal and spatial arctic sea level variations than the actual ocean mass change seen by GRACE.

Often are estimates of steric height variations in the Arctic model-based constrained by a few in-situ measurements (e.g. ECCO v4) or by simply subtracting OBP from altimetric SSH, assuming that the SSH is correct.

In this study, we collect all available hydrographic data north of the 65th latitude by adding recent data from 2016-2017 to the UDASH-database (Behrendt et al, 2017). All in-situ profiles are quality checked and finally interpolated into a 4D-hydrographic grid covering the Arctic region for the altimetry era, thereby making the contributions to the sea level budget (SSH=mass+steric) independent of each other.

In a comparison with oceanographic models, both the models and the in-situ dataset show that freshwater variations are dominating variations in Arctic sea level change, but where the models seem to have little non-seasonal variation, the in-situ data show a clear correlation with the Arctic Oscillation index. The same correlation is seen in SSH altimetry, confirming that freshwater variations dominates the Arctic sea level change.

A GRACE RL06-solution corrected for leakage, glacial isostatic adjustment and atmospheric loading is used together with the hydrographic dataset to consolidate the improved altimetry-based DTUArcticSeaLevel-product in both the spatial and temporal dimension.

Area D - Board 466

Satellite altimetry was initially designed for observing the open ocean, however efforts to retrieve sea surface topography in ice covered areas and sea ice parameters in the Arctic Ocean have a long history (e.g. Peacock, 2004).
The majority of current Arctic Ocean sea surface processings rely on using different retracking algorithms for open ocean and ice covered areas, thus empirical bias corrections between open ocean and ice-covered areas are needed, and are still an open issue. However recent advances in waveform processing allow for a processing continuity from open ocean to ice covered areas, and the retrieval of ocean and sea-ice topography from the same retracking algorithm, which uses a physical echo shape model able to fit ocean waveforms as well as lead waveforms (Poisson et al., 2018). Using this retracking algorithm removes the need for estimating biases between the open and ice covered ocean.
This new physical retracking approach, combined with a neural classification of echoes, was applied on conventionnal altimeter high rate data from SARAL/AltiKa and Envisat missions to produce a sea surface height and sea ice freeboard dataset. Results of the validation performed in the framework of the CNES PEACHI and ESA Sea Level CCI projects suggest that this new method performs well compared to existing products and in-situ data, and that large scale ocean circulation and sea-ice features are consistently observed.
In this talk, we will present the processing scheme used to derive SARAL/AltiKa and Envisat datasets, from waveform retracking to the final gridded product, as well as validation results.
New developments include the processing of delay-doppler altimetry, and the design of a true multi-mission combination system, tailored to the area, to resolve small-scale features. This approach will be used to produce future CMEMS Arctic sea level products.

Area D - Board 468

Along with global climate change, last decades have seen a strong intensification of large circulation systems in the Arctic Ocean, such as the Beaufort Gyre (BG) and Transpolar Drift (TPD). This strengthening impacts heat and transport, potentially influencing the global ocean. For instance, changes in the TPD determine variations in sea ice and freshwater volume export from the Arctic through – mostly – Fram Strait.

Evidence of positive trend in surface motion, particularly enhanced in the summer season, has been found in sea ice drift observations. However, the scarcity of observations in the upper Arctic Ocean, due mostly to ice-coverage, has hindered furthering our understanding. Firstly, it is still uncertain which mechanisms contribute to these changes, and to what extent. For instance, the small increase in wind stress observed at the surface of the ocean cannot fully explain the increase in ice drift. In addition, it is not clear whether the enhanced strengthening observed in the last decade is the result of a steady trend or part of long-term oscillations. Recently, observations of near surface ocean flow field have been greatly improved by reprocessing of satellite altimetry data over the ice-covered oceans. New algorithms allow to retrieve sea surface height (SSH) of water emerging from leads.

Eventually, in this work we aim to contribute to the understanding of long-term variability of near-surface currents in the Arctic Ocean by analysing changes in both the strength and motion patterns of the TPD seasonal to decadal variations. SSH from Cryosat2 and Envisat satellites in the period 2002-2017 will be used to retrieve Arctic surface geostrophic circulation. Over the ice-covered Arctic Ocean we will use SSH observations processed with a leads-detecting algorithm (Ricker et al., 2014). Particular attention will be given to the regions of formation and outflow of the TPD, respectively the Siberian Shelves and Fram Strait.

Here, we present preliminary results of our SSH analysis for months representative of winter and summer seasons. From our first analysis, two challenges emerged: aliasing of sub-monthly variability and consistency of the SSH field across the ice-edge.

• SSH exhibits rapid fluctuations due to barotropic waves, which are unresolved by altimetric satellites. In the eastern shelves, the high amplitude of these waves, together with the spatial aliasing, cover the weaker underlying longer term changes in SSH and surface currents. This effect is still not well evaluated at high latitudes. We present here an assessment of the variance reduction in SSH following the application of a de-aliasing product over the Arctic Ocean north of 65 degrees N.
• In Fram Strait, both ice-covered and ice-free areas are present all year round. Since different processing are used to derive SSH from altimetry over these two types of surface, the question is raised whether it is possible to retrieve coherent circulation patterns across the ice edge. Long term time series of high quality in situ data on circulation are available in this region. These observations are compared against SSH field to test the reliability of altimetry in terms of variability.

Area D - Board 467