Data Description

Parameters

The main parameter in this data set is ice velocity, with associated errors and standard deviations (in m/yr). The variables included in the data file are described in Table 1. All variables are on a two-dimensional grid of 12445 by 12445 cells, except for x and y, which are unidimensional with 12445 cells each.

Note: To convert the VX and VY ice velocity components into magnitude (speed) and direction (angle), as well as their relative errors, use the following equations (Mouginot et al., 2012):

1. speed = sqrt (VX² + VY²)
2. angle = arctan (VY / VX)
3. error = sqrt (ERRX² + ERRY²)
4. error of flow direction = error / (2*speed)

When computing the inverse tangent, users should take note of the function's inherent ambiguities. While the standard arctan function typically does not account for angles that differ by 180°, most modern computer languages and math software packages include the function ATAN2, which uses the signs of both vector components to place the angle in the proper quadrant.

File Information

Format

This data set includes one data file, antarctic_ice_vel_phase_map_v01.nc. The data file is provided in netCDF-4 (.nc) format, following version 1.6 of the Climate and Forecast (CF) metadata conventions.

File Contents

The velocity components for the x- and y-direction, as defined by the polar stereographic grid, are stored in the netCDF variables VX and VY and are recorded in m/yr. Error estimates for the velocity components are provided as variables ERRX and ERRY; however, these values should be used more as an indication of relative quality rather than absolute error. More information about the error estimates is provided in the Quality, Errors, and Limitations section as well as in Mouginot et al. (2019). The data also include the standard deviations for the velocity estimates (STDX, STDY), a count of scenes (CNT) used to estimate the values for each pixel, and an index denoting the SOURCE of the data, i.e., whether the velocity was estimated via tracking-based method, phase-based method, an interpolation of the two, or is missing altogether.

Spatial Information

Coverage

Spatial coverage includes the continent of Antarctica, namely the entire polar region southwards of 60° S.

Resolution

The spatial resolution of the grid is 450 m by 450 m.

Geolocation

The following table provides information for geolocating this data set.

Temporal Information

Coverage

The data were collected between 1996 and 2018. Detailed information is provided in the Data Aquisition and Processing section.

Resolution

Not applicable. The maps are an averaged representation of the period in which data were collected.

Data Acquisition and Processing

Background

This data set provides a new map of Antarctic ice velocity that is ten times more precise than prior maps and reveals ice motion at a high precision over 80% of the continent versus 20% in the past. The ice motion vector map provides novel constraints on interior ice motion and its connection with the glaciers and ice streams that control the stability and mass balance of the Antarctic Ice Sheet.

Instrumentation

The ice velocity map was derived from a variety of satellite radar interferometry data from the sensors listed in Table 3.

Note: Data from Sentinel-1/SAR, Landsat-8/OLI, and TanDEM-X/TerraSAR-X/SAR were only used for the speckle-tracking analysis, while data from ALOS2/PALSAR2 were only used in the phase analysis. Data from all the other sensors were used in both approaches.

Acquisition and Processing

Several analysis techniques using SAR data were used to generate the velocity maps:

1. Combination of interferometric phases of two independent tracks to retrieve the surface flow vector (Mouginot et al., 2019), applied to a vast region of Antarctica
2. Speckle tracking in both along-track (azimuth) and across-track (range) directions, used to augment the phase map in coastal areas (Rignot et al., 2011; Mouginot et al., 2012; Mouginot et al., 2017)
3. Calculation of two-dimensional offsets in amplitude imagery (Mouginot et al., 2012)
4. Combinations of (range) interferometric phases along two independent tracks (Mouginot et al., 2012)

In all cases, surface parallel flow is assumed, a conventional approach for ice sheets. The Landsat-8 data are processed using repeat image feature tracking (see Mouginot et al., 2017).

Phase-derived velocities were mostly acquired between 2007 and 2018, while regions covered by tracking-derived velocity (along the coasts) are mostly representative of the years 2013–2017. Additional data acquired between 1996 and 2018 were used as needed to maximize coverage. SAR acquisitions between 2006 and 2018 were coordinated by the International Polar Year (IPY) Space Task Group and its successor, the Polar Space Task Group (PSTG).

Figure 1 shows the distribution of data tracks from different data sources, as well as a complete map of ice motion in Antarctica.

The unwrapped phase uses C-Band sensors (~5.5 cm) with 1461 tracks from RADARSAT-1 and RADARSAT-2, 76 tracks from ERS-1 and ERS-2, 319 tracks from ENVISAT/ASAR; and L-Band sensors (~23.6 cm) with 556 tracks from ALOS/PALSAR and 3 tracks from ALOS2/PALSAR2 (Figure 1, top row). The combination of phase and tracking mosaics covers 99.8% of Antarctica (Figure 1, main map), with a significantly improved description of ice flow from the ice divides in the deep interior to the periphery. Figure 2 shows the ice velocity mosaics obtained from each method alone. 

The phase mosaic (Figure 2B) itself does not cover all of Antarctica for two reasons: firstly, the interferometric phase is aliased in areas of fast flow due to the long temporal baseline of the data; and secondly, some sectors do not have sufficient multi-track coverage to provide reliable estimates of the vector ice motion. Overall, the phase-based map covers 71% of Antarctica, or 9.9 million square km, while the tracking-based map (Figure 2A) covers 99.6%, or 13.8 million square km (Mouginot et al., 2017). Figure 3 illustrates how the two mosaics are merged.

Quality, Errors, and Limitations

A detailed description of the data and their quality is provided in Mouginot et al. (2019). The precision of ice flow mapping varies with the sensor, the geographic location, the technique of interferometric analysis, the time period of analysis, the repeat cycle, and the amount of data stacking. Overall, there is notable improvement in precision from phase-based mapping relative to tracking-based mapping, which is especially significant along ice divides when examining the error in the flow direction. Figure 4 compares errors estimates for the two methods, for both speed and direction. 

Finally, Table 4 provides a phase error for each of the sensors utilized in the phase-based velocity mapping. Note that phase error is determined for a single unwrapped phase using the nominal repeat cycle and assuming a conservative error of π radians (½ cycle) (Mouginot et al., 2019).

Software and Tools

Unidata at the University Corporation for Atmospheric Research maintains an extensive list of freely available Software for Manipulating or Displaying NetCDF Data.

Related Data Sets

• MEaSUREs InSAR-based Antarctica Ice Velocity Map
• MEaSUREs Antarctic Grounding Line from Differential Satellite Radar Interferometry
• MEaSUREs InSAR-Based Ice Velocity Maps of Central Antarctica: 1997 and 2009
• MeaSUREs InSAR-Based Ice Velocity of the Amundsen Sea Embayment, Antarctica
• MEaSUREs Antarctic Boundaries for IPY 2007-2009 from Satellite Radar
• MEaSUREs Annual Antarctic Ice Velocity Maps 2005-2017

Related Websites

• NASA MEaSUREs Data at NSIDC
• Antarctic Ice Sheet Velocity (AIV) and Mapping Data at NSIDC
• NASA MEaSUREs

Contacts and Acknowledgments

Investigators:

Dr. Jeremie Mouginot^1,2
1University of California, Irvine
Department of Earth System Science
Croul Hall
Irvine, California 92697
USA

²Institut des Géosciences de l’Environnement (IGE)
Université Grenoble Alpes
France

Dr. Eric Rignot
University of California, Irvine
Department of Earth System Science
Croul Hall
Irvine, California 92697
USA

Dr. Bernd Scheuchl
University of California, Irvine
Department of Earth System Science
Croul Hall
Irvine, California 92697
USA

Acknowledgments:

These data were generated through a grant from the NASA MEaSUREs program. Spaceborne Synthetic Aperture Radar (SAR) acquisitions were provided through the following data agencies:

• ALOS PALSAR, ALOS-2/PALSAR-2: Japan Aerospace Exploration Agency (JAXA)
• ENVISAT ASAR, ERS-1, ERS-2: European Space Agency (ESA)
• Sentinel-1: Copernicus/ESA
• RADARSAT-1, RADARSAT-2: Canadian Space Agency (CSA), MDA
• TerraSAR-X, TanDEM-X as well as the TanDEM-X DEM of Antarctica: German Space Agency (DLR)
• Landsat-8 (optical): made available by the United States Geological Survey (USGS).

Data acquisitions made between 2006 and 2018 are courtesy of the International Polar Year (IPY) Space Task Group and its successor, the Polar Space Task Group (PSTG). This data set contains modified Copernicus Sentinel data (2014–2017), acquired by the European Space Agency, distributed through the Alaska Satellite Facility, and processed by E. Rignot, J. Mouginot, and B. Scheuchl.

References

• Rignot, E., Mouginot, J., & Scheuchl, B. (2011). Ice Flow of the Antarctic Ice Sheet. Science, 333(6048), 1427–1430. https://doi.org/10.1126/science.1208336
• Mouginot, J., Scheuchl, B., & Rignot, E. (2012). Mapping of Ice Motion in Antarctica Using Synthetic-Aperture Radar Data. Remote Sensing, 4(9), 2753–2767. https://doi.org/10.3390/rs4092753
• Scheuchl, B., Mouginot, J., & Rignot, E. (2012). Ice velocity changes in the Ross and Ronne sectors observed using satellite radar data from 1997 and 2009. The Cryosphere, 6(5), 1019–1030. https://doi.org/10.5194/tc-6-1019-2012
• Mouginot, J., Rignot, E., & Scheuchl, B. (2014). Sustained increase in ice discharge from the Amundsen Sea Embayment, West Antarctica, from 1973 to 2013. Geophysical Research Letters, 41(5), 1576–1584. https://doi.org/10.1002/2013gl059069
• Mouginot, J., Rignot, E., Scheuchl, B., & Millan, R. (2017). Comprehensive Annual Ice Sheet Velocity Mapping Using Landsat-8, Sentinel-1, and RADARSAT-2 Data. Remote Sensing, 9(4), 364. https://doi.org/10.3390/rs9040364
• Mouginot, J., Rignot, E., & Scheuchl, B. (2019). Continent‐wide, interferometric SAR phase, mapping of Antarctic ice velocity. Geophysical Research Letters. https://doi.org/10.1029/2019gl083826