GLAS/ICESat Laser Altimetry Digital Elevation Models (DEMs) of Antarctica and Greenland

Summary

This document pertains to the following two digital elevation model (DEM) data sets.

Data Set ID Data Set Title
NSIDC-0304 GLAS/ICESat 500 m Laser Altimetry Digital Elevation Model of Antarctica
NSIDC-0305 GLAS/ICESat 1 km Laser Altimetry Digital Elevation Model of Greenland

The Geoscience Laser Altimeter System (GLAS) instrument on the Ice, Cloud, and land Elevation Satellite (ICESat) provides global measurements of elevation, and repeats measurements along nearly-identical tracks; its primary mission is to measure changes in ice volume (mass balance) over time. These digital elevation models (DEMs) of Antarctica and Greenland are derived from GLAS/ICESat laser altimetry profile data and provide new surface elevation grids of the ice sheets and coastal areas, with greater latitudinal extent and fewer slope-related effects than radar altimetry.

These DEMs are generated from the first seven operational periods (from February 2003 through June 2005) of the GLAS instrument. They are provided on polar stereographic grids at 500 m grid spacing for Antarctica and 1 km grid spacing for Greenland. The grids cover all of Antarctica north of 86° S and all of Greenland south of 83° N. Elevations for both ice sheets are reported as centimeters above the datums, relative to both the WGS 84 Ellipsoid and the EGM96 Geoid, in two separate elevation data files. A data quality map of the interpolation distance is distributed for each ice sheet. ENVI header files are also provided.

The data are in 4-byte (long) signed integer binary files (big endian byte order) and are available via FTP.

Citing These Data

We kindly request that you cite the use of this data set in a publication using the following citation example. For more information, see our Use and Copyright Web page.

The following examples show how to cite the use of these data sets in a publication. The citation to use depends on which data set you use. List the principal investigators, year of data set release, data set title and version (if applicable), publisher information (NSIDC), and indicate digital media distribution.

DiMarzio, J., A. Brenner, R. Schutz, C. A. Shuman, and H. J. Zwally. 2007. GLAS/ICESat 500 m laser altimetry digital elevation model of Antarctica. Boulder, Colorado USA: National Snow and Ice Data Center. Digital media.

DiMarzio, J., A. Brenner, R. Schutz, C. A. Shuman, and H. J. Zwally. 2007. GLAS/ICESat 1 km laser altimetry digital elevation model of Greenland. Boulder, Colorado USA: National Snow and Ice Data Center. Digital media.

Overview Table

Category Description
Data format Binary, 4-byte (long) signed integer, big endian byte order.

Note: The data format information in this document represents the data in its native format as it is archived at NSIDC. If you have downloaded the data using Polaris, please consult the 00README file located in the tar file for information on the data format operations that were performed on this data set.

Spatial coverage and resolution
DEM Data Set
Spatial Coverage
Grid Resolution
Antarctica
Southernmost Latitude:
86.0° S
Northernmost Latitude:
63.0° S
Westernmost Longitude:
180.0° W
Easternmost Longitude:
180.0° E
500 m 
Greenland
Southernmost Latitude:
60.0° N
Northernmost Latitude:
83.0° N
Westernmost Longitude:
 73.0° W
Easternmost Longitude:
11.0° W
1000 m 
Temporal coverage and resolution Average of data from February, 2003 to June, 2005
Grid type and size Polar stereographic grids; size varies by DEM data set:

DEM Data Set Columns Rows
Antarctica
11352
9368
Greenland
2611
2782
Tools for accessing data IDL, ENVI, or other commercial off-the-shelf software for image processing
File naming convention NSIDC_rrrssss_ddddd_vvvvv_uu.dat
File size
DEM Data Set
Size Per Data File
Antarctica
425,382,144 bytes
Greenland
29,055,208 bytes

The three data files for each DEM are distributed in compressed form.
Parameter(s) Elevation above datum (WGS 84 Ellipsoid or EGM96 Geoid), in centimeters
Procedures for obtaining data Data are available via FTP.

Table of Contents

1. Contacts and Acknowledgments
2. Detailed Data Description
3. Data Access and Tools
4. Data Acquisition and Processing
5. References and Related Publications
6. Document Information

1. Contacts and Acknowledgments

Investigator(s)

John P. DiMarzio
SGT Inc.
Cryospheric Sciences Branch
Code 614.1
NASA/Goddard Space Flight Center
Greenbelt, Maryland USA

Anita C. Brenner
SSAI
Code 614.5
NASA/Goddard Space Flight Center
Greenbelt, Maryland USA

Helen A. Fricker
Institute of Geophysics and Planetary Physics
Scripps Institution of Oceanography
University of California, San Diego
La Jolla, California USA

Robert E. Schutz
Center for Space Research
University of Texas, Austin
Austin, Texas USA

Christopher A. Shuman
Cryospheric Sciences Branch
Code 614.1
NASA/Goddard Space Flight Center
Greenbelt, Maryland USA

H. Jay Zwally
Cryospheric Sciences Branch
Code 614.1
NASA/Goddard Space Flight Center
Greenbelt, Maryland USA

Technical Contact

NSIDC User Services
National Snow and Ice Data Center
CIRES, 449 UCB
University of Colorado
Boulder, CO 80309-0449  USA
phone: +1 303.492.6199
fax: +1 303.492.2468
form: Contact NSIDC User Services
e-mail: nsidc@nsidc.org

2. Detailed Data Description

Format

Note: The data format information in this document represents the data in its native format as it is archived at NSIDC. If you have downloaded the data using Polaris, please consult the 00README file located in the tar file for information on the data format operations that were performed on this data set.

The data are in 4-byte (long) signed integer binary files (big endian byte order).

File and Directory Structure

The DEM files are located on the FTP server at the following location.

ftp://sidads.colorado.edu/pub/DATASETS/DEM/

Within the DEM directory, there is a separate subdirectory for each data set, as illustrated below.

Directory structure

File Naming Convention

Data files are named according to the following convention.

NSIDC_Rrrssss_ddddd_vvvvv_uu.dat

where NSIDC indicates the grid was created at NSIDC (derived from data provided by the authors) and

Rrr
=
Region of coverage for the DEM data set (Ant = Antarctica; Grn = Greenland)
ssss
=
Grid spacing (1km or 500m)
ddddd
=
Datum (wgs84 = WGS 84 Ellipsoid datum; egm96 = EGM96 Geoid datum)
vvvv
=
Type of values in grid (elev = elevation; dist = mean distance)
uu
=
Unit of measurement (cm = elevation values reported as integer values in centimeters; mm = distance values reported as integer values in millimeters)

The data files are described in the following table.

Antarctica DEM Data Set
File Name Description
NSIDC_Ant500m_wgs84_elev_cm.dat Elevation (centimeters) relative to WGS 84 ellipsoid datum
NSIDC_Ant500m_egm96_elev_cm.dat Elevation (centimeters) relative to EGM96 Geoid datum
NSIDC_Ant500m_dist_mm.dat Mean distance (millimeters) from contributing GLAS data for each grid cell
Greenland DEM Data Set
File Name Description
NSIDC_Grn1km_wgs84_elev_cm.dat Elevation (centimeters) relative to WGS 84 Ellipsoid datum
NSIDC_Grn1km_egm96_elev_cm.dat Elevation (centimeters) relative to EGM96 Geoid datum
NDISC_Grn1km_dist_mm.dat Mean distance (millimeters) from contributing GLAS data for each grid cell

ENVI header files are also provided for each data file. These header files are named the same as the data files, but with an additional .hdr extension appended to the file name; for example, NSIDC_Ant500m_wgs84_elev_cm.dat.hdr.

The three data files for each DEM are distributed in compressed (zipped) form. Each compressed file has a .gz extension appended to the end of the file name; for example, NSIDC_Ant500m_wgs84_elev_cm.dat.gz.

File Size

File sizes for each DEM data set are shown in the following table. The three data files (two elevation and one mean distance) are all the same size for each DEM (they all contain the same number of grid cells). The sizes shown are uncompressed. The three data files for each DEM are distributed in compressed (zipped) form; the ENVI header files are not compressed.

DEM Data Set
Size Per Data File
ENVI Header File Size
Antarctica
425,382,144 bytes
481 bytes
Greenland
29,055,208 bytes
504 bytes

Spatial Coverage

The following table shows the spatial coverage and the grid resolution for each DEM data set.

DEM Data Set
Spatial Coverage
Grid Resolution
Antarctica
Southernmost Latitude:
86.0° S
Northernmost Latitude:
63.0° S
Westernmost Longitude:
180.0° W
Easternmost Longitude:
180.0° E
500 m 
Greenland
Southernmost Latitude:
60.0° N
Northernmost Latitude:
83.0° N
Westernmost Longitude:
 73.0° W
Easternmost Longitude:
11.0° W
1000 m 

Projection and Grid Description

The data grid projections are polar stereographic, projected from the TOPEX/Poseidon Ellipsoid. However, the horizontal location differences between TOPEX/Poseidon and WGS 84 postings are so small (less than one meter) that the projection ellipsoid can be considered to be WGS 84. (See also: How does the GLAS ellipsoid compare with WGS 84?) The latitude of true scale is 70° S (Antarctica) or 71° N (Greenland). (Note that some other data sets covering Antarctica are projected to 71° S, not 70° S; the 70° S latitude of true scale is the choice of the data providers.) For the EGM96-referenced grids, the ellipsoid-projected data is modified by a geoid correction for each grid cell (the geoid is the approximate local sea level elevation relative to the ellipsoid). The central longitude of the projection (parallel to the image y-axis) is 0° extending up from the South Pole for Antarctica and 45° W extending down from the North Pole for Greenland. The center of the Antarctic grid is not the South Pole. The grid spacing is 500 m for Antarctica and 1 km for Greenland. The full projection information is provided in the ENVI header files associated with each of the data files.

The data are provided in image-formatted grids of binary numbers. Data image arrays have upper-left, row-by-row organization. Image grid size is shown in the following table.

DEM Data Set
Columns
Rows
Antarctica
11352
9368
Greenland
2611
2782

Projection plane corner points are shown in the following table.

Antarctica DEM Data Set
Center Point of Corner Grid Cell
  Column Row x y Latitude Longitude
Upper Left
0.0
0.0
-2812000.0
2299500.0
-57.3452815
-50.7255753
Upper Right
11351.0
0.0
2863500.0
2299500.0
-57.0043684
51.2342036
Lower Left
0.0
9367.0
-2812000.0
-2384000.0
-56.8847122
-130.2911169
Lower Right
11351.0
9367.0
2863500.0
-2384000.0
-56.5495152
129.7789915
Outer Corner of Corner Grid Cell
  Column Row x y Latitude Longitude
Upper Left
-0.5
-0.5
-2812250.0
2299750.0
-57.3422816
-50.7250190
Upper Right
11351.5
-0.5
2863750.0
2299750.0
-57.0013764
51.2336047
Lower Left
-0.5
9367.5
-2812250.0
-2384250.0
-56.8817144
-130.2915680
Lower Right
11351.5
9367.5
2863750.0
-2384250.0
-56.5465248
129.7794862
Greenland DEM Data Set
Center Point of Corner Grid Cell
  Column Row x y Latitude Longitude
Upper Left
0.0
0.0
-890000.0
-629000.0
79.9641229
-99.7495626
Upper Right
2610.0
0.0
1720000.0
-629000.0
73.2101234
24.9126514
Lower Left
0.0
2781.0
-890000.0
-3410000.0
58.2706251
-59.6277136
Lower Right
2610.0
2781.0
1720000.0
-3410000.0
55.7592932
-18.2336764
Outer Corner of Corner Grid Cell
  Column Row x y Latitude Longitude
Upper Left
-0.5
-0.5
-890500.0
-628500.0
79.9630236
-99.7861964
Upper Right
2610.5
-0.5
1720500.0
-628500.0
73.2074291
24.9327117
Lower Left
-0.5
2781.5
-890500.0
-3410500.0
58.2653993
-59.6335252
Lower Right
2610.5
2781.5
1720500.0
-3410500.0
55.7536121
-18.2303578

Temporal Coverage

Data used to produce the DEMs were collected from the GLAS instrument from February, 2003 to June, 2005.

The DEM data providers (DiMarzio et al.) used data from the following GLAS/ICESat lasers and data acquisition periods to generate these DEMs. The DEM data are an average of those data.

Parameter or Variable

The parameter is elevation above the datum (WGS 84 Ellipsoid or EGM96 Geoid), measured in centimeters. The supporting distance files contain the mean distance, in millimeters, to the contributing data for each grid cell.

Parameter Range

The minimum and maximum data values for each data file are shown in the following table.

Antarctica DEM Data Set
File Name Minimum Maximum
NSIDC_Ant500m_wgs84_elev_cm.dat
1
422547
NSIDC_Ant500m_egm96_elev_cm.dat
1
426194
NSIDC_Ant500m_dist_mm.dat
4
996087
Greenland DEM Data Set
File Name Minimum Maximum
NSIDC_Grn1km_wgs84_elev_cm.dat -89838 400459
NSIDC_Grn1km_egm96_elev_cm.dat 1 395636
NDISC_Grn1km_dist_mm.dat 6 994056

Note: The very large negative values in the Greenland DEM data are artifacts of the re-projection and re-gridding process and are limited to the grid corner areas. Users should ignore values that are outside the coast of Greenland.

Sample Images

The following images were created using hill-shading in ENVI from the EGM96 Geoid elevation data.

Antarctica DEM Data Set
Antarctica
  Greenland DEM Data Set
Greenland

Sample Data Record

The following data sample was created from the NSIDC_Grn1km_egm96_elev_cm.dat file using the Unix command "od -t d4". This command interprets each 4-bytes as signed integers and displays the values. In this sample, only a few data values from the file are illustrated, and each row of four data values is preceded by its byte-offset in the file (such as 3221540). The asterisks indicate that all the data in that part of the file are zeroes.

0000000       0       0       0       0
*
3221540       0       0       0    8236
3221560       0       0       0       0
*
3221640       0       0       0    3562
3221660       0       0       0       0
*
3246040       0       0       0   55906
3246060   47321   34042   14507       0
3246100       0       0       0       0
*
3246160       0    1265    2615    2619
3246200    1896       0       0       0
3246220       0       0       0       0
*
3272340       0       0       0   46945
3272360   48231   53159   51417   48451
3272400   34656   16470    8105       0
3272420       0       0       0       0
*
3272460       0       0       0    2336
3272500    2026    3224    2043    1092
3272520       0       0       0       0
*

Error Sources

Several localized elevation artifacts associated with cloud effects, track mislocations, and gridding algorithms are present in the grids. DEM accuracy decreases as track spacing between ICESat profiles increases equatorward.

Quality Assessment

Slope and elevation accuracy of the DEMs is best near the latitudinal limit of satellite coverage (±86°), and provides a good representation of the ice sheet and mountain areas (based on comparison with satellite images). DEM accuracy decreases as track spacing between ICESat profiles increases equatorward.

As a quality and accuracy indicator, the mean distance of all contributing GLAS surface spot data to the grid cell center was recorded in a separate identical grid for each ice sheet. These mean distance files are provided with the elevation data.

3. Data Access and Tools

Data Access

Data are available via FTP. Access to the data is unrestricted, but users are encouraged to register for the data. Registered users will receive e-mail notification about any product changes.

Volume

The total distribution volume for each DEM data set is listed below. The three data files for each data set are distributed in compressed (zipped) form; the ENVI header files are not compressed for distribution.

DEM Data Set Total Volume
Uncompressed
Total Volume with
Compressed Data
Antarctica 1,276,147,875 bytes 433,997,915 bytes
Greenland 87,167,136 bytes 25,141,042 bytes

Software and Tools

View the data files using IDL, ENVI, or other commercial off-the-shelf software for image processing. ENVI header files are distributed with the data files.

Related Data Collections

4. Data Acquisition and Processing

Theory of Measurements

Satellite laser altimetry is an active measurement technique using on-board laser light along with a detector telescope and timing system to measure range to a reflecting spot on the Earth’s surface from the satellite. Precision orbit determination (POD) and precision attitude determination (PAD) are used to calculate a ground location of the laser spot.

Several additional corrections are required, including atmospheric refraction, return waveform analysis for complex or multi-level returns, and detector saturation correction. The measurement is susceptible to obscuration by thick clouds, and thin clouds can affect the reported surface elevation (in general making it appear too low). Clouds also reduce the amount of returned (reflected) energy, and an automatic gain adjustment is applied to the detectors after a few low-energy shots are received. Gain settings range from a minimum of 13 (maximum returned energy) to a maximum of 250 (weak energy return). Characteristics of the surface, such as slope or roughness, can also modify the return waveform. 

Sensor or Instrument Description

Please visit the official NASA Goddard Space Flight Center (GSFC) ICESat/GLAS Web site for details about the ICESat platform and GLAS instrument.

Also see ICESat Reference Orbit Ground Tracks for a summary of the orbits for each laser operational period.

Data Acquisition Methods

Please see the GLAS/ICESat L1B Global Elevation Data (GLA06) documentation for a description of the acquisition of individual elevation points from the ICESat/GLAS sensor.

Data Source

These DEMs were generated using data from the GLAS/ICESat L1B Global Elevation Data (GLA06) data set. This source data set includes the removal of the solid tides and the application of atmospheric corrections (Brenner et al., 2003).

Derivation Techniques and Algorithms

Several criteria and filters were applied to the total data set to minimize the number of low-quality elevation points in the data used for gridding. Off-nadir data acquired during "round the world scans" (a calibrating technique for testing pointing accuracy) were not used. Other non-surface return data (for example, cloud-impacted, fog-impacted, or blowing-snow-impacted elevations) were filtered out by only selecting data points with relatively low gain (that is, high surface return power; gain was limited to <50). No saturation correction was applied to the selected data because the saturation correction was still under development at the time of DEM generation. Saturation can cause errors (elevation underestimates) of up to 50 cm during high power laser operations periods, such as Laser 2A and 3A (Abshire et al., 2005).

Individual elevation point data were gridded onto a polar stereographic projection, combining many elevation points for each grid cell. Satellite altimetry data are unevenly spaced due to obscuration and variations in track spacing. For each grid node, a bi-quadratic surface was fitted to all elevations within a circular region surrounding the node (hereafter referred to as the "cap"). This was accomplished via a singular value decomposition (SVD) algorithm (Zwally et al., 1990). In the fitting routine, a weight (w) is assigned to each elevation within the cap according to

w = 1/(d2σ0)

Where:

d
=
the distance from the spot elevation to the grid node center
σ0
=
the expected error of each elevation

The expected error value (σ0) was set to 20 cm for all observations based on the standard deviation of crossover differences (DiMarzio et al., 2004; see also Shuman et al., 2006). When the fit quality estimate from the SVD algorithm indicates insufficient data within the cap for a valid bi-quadratic surface fit, the cap size was increased to include more data. Once a valid fit was found, an interactive procedure was used in which elevations differing by more than three standard deviations from the fitted surface are eliminated, and the SVD algorithm was re-applied.

Error estimates associated with each node (σg) are determined by the output of the SVD algorithm and the weighted root-mean-squared differences of the remaining elevations with the fitted surface. If σg was greater than 30 m, the estimate was declared invalid and the cap-fitting and error-estimation process was redone with a larger cap size. If no cap size (below 20 km) yields a valid elevation, then a bi-linear fit to the nearest data was applied. Cap sizes ranged from 2 km to 20 km for Antarctica and from 5.5 km to 20 km for Greenland.

The grid spacing is 500 m for Antarctica and 1 km for Greenland. The reduced grid spacing for Antarctica fully exploits the greater data density available in the areas just north of 86° S.

Validation and Accuracy

The DEM data providers (DiMarzio et al.) evaluated the GLAS/ICESat laser altimetry Greenland DEM data set by comparing it to airborne laser altimetry data. Airborne Topographic Mapper (ATM) data over Greenland have a reported accuracy of ±10 cm (Krabill et al., 2002). Using all ATM data within 600 m of the GLAS/ICESat 1 km gridded DEM nodes, a comparison of elevation data over northwestern Greenland (the region of densest ATM tracks) indicates ATM elevations are generally lower than the GLAS/ICESat Greenland DEM elevation data. The overall mean difference between the ATM and GLAS/ICESat DEM data is -41 cm (ATM elevations below GLAS/ICESat DEM elevations), with an error range of ±44 cm. The mean difference varies with slope: for regions with <0.1° slopes, the mean difference is -32 ±43 cm; for regions with 0.1° to 1.0° slopes, the mean difference is -66 ±61 cm.

A second evaluation compared the GLAS/ICESat laser altimetry Greenland DEM to a 5 km gridded radar altimetry DEM data set (Zwally and Brenner, 2001). This comparison indicates a mean difference of -2 ±9 m. The mean difference varies with slope: for regions with <0.1°slopes, the mean difference is 1 ±5 m; for regions with 0.1° to 1.0° slopes, the mean difference is -24 ±20 m.

NSIDC conducted a separate evaluation of the GLAS/ICESat laser altimetry Antarctic and Greenland DEMs, using filtering, shaded-relief processing, and comparison to satellite image mosaics of the ice sheets. Horizontal spatial resolution of these DEMs is approximately 7.5 km, despite the finer gridding scale for the data. This is likely a result of the large "cap" region and the simple surface-fitting routine used to evaluate the grid elevations. In several areas, large spike and divot errors (features not present in the image mosaics) deviated from the likely surface elevation by several tens of meters. A somewhat smaller dimple pattern due to gridding algorithm limitations is also present throughout both data sets. NSIDC provides these DEMs as unfiltered; however, for many regional mapping applications, the user may find that a 7.5 km low-pass-filtered version of the DEMs will give a better representation of the true surface (smoothed to that same scale), with subdued dimple, spike, and divot features, relative to the distributed DEM data.

As a quality and accuracy indicator, the mean distance of all contributing GLAS surface spot data to the grid cell center was recorded in a separate identical grid for each ice sheet. This interpolation distance file is distributed with the elevation data.

5. References and Related Publications

Abshire, J. et al. 2005. Geoscience Laser Altimeter System (GLAS) on the ICESat mission: on-orbit measurement performance. Geophysical Research Letters v. 32, L21S02, doi:10.1029/2005GL1024028.

Brenner, A. et al. 2003. Derivation of range and range distributions from laser pulse waveform analysis for surface elevations, roughness, slope, and vegetations heights, Algorithm Theoretical Basis Document (ATBD) v.4.1, available at http://www.csr.utexas.edu/glas/atbd.html.

DiMarzio, J. P., A. C. Brenner, and H. J. Zwally. 2004. Comparison of Envisat and ERS radar altimetry for ice –sheet elevation change studies. ESA Envisat Sympsium, Salzburg, Austria, September 9, 2004.

Krabill, W. et al. 2002. Aircraft laser altimetry measurement of elevation changes of the Greenland Ice Sheet: technique and accuracy assessment. J. Geodynamics 34, 357-376.

Shuman, C., H. J. Zwally, B. Schutz, A. Brenner, J. DiMarzio, V. Schuedo, and H. Fricker. 2006. ICESat Antarctic elevation data: preliminary precision and accuracy assessment. Geophysical Research Letters 33, L07501, doi:10.1029/2005GL025227.

Zwally, H. J., A. C. Brenner, J. A. Major, T. V. Martin, and R. A. Bindschadler. 1990. Satellite radar altimetry over ice, Vol. 1. Processing and corrections of Seasat data over Greenland. NASA Reference Publication 1233.

Zwally, H. J. and A. C. Brenner. 2001. "Ice sheet dynamics and mass balance." In: Satellite Altimetry and Earth Science, L-L. Fu and A. Cazenave, Eds., Academic Press, Ch. 9, 351-369.

The following related documents are available on NSIDC's Web site.

Document
Description
URL
ICESat Reference Orbit Ground Tracks Summary of ICESat orbits http://nsidc.org/data/icesat/orbit.html
GLAS/ICESat L1B Global Elevation Data (GLA06) Source data set documentation http://nsidc.org/data/gla06.html

6. Document Information

Acronyms and Abbreviations

The following acronyms and abbreviations are used in this document.

ATM Airborne Topographic Mapper
DEM Digital elevation model
EGM96 Earth Gravitational Model 1996
ENVI Environment for Visualizing Images (software from ITT Visual Information Solutions)
FTP File Transfer Protocol
GLAS Geoscience Laser Altimeter System
GSFC Goddard Space Flight Center
NASA National Aeronautics and Space Administration
ICESat Ice, Cloud, and land Elevation Satellite
IDL Interactive Data Language (software from ITT Visual Information Solutions)
NSIDC National Snow and Ice Data Center
PAD Precision attitude determination
POD Precision orbit determination
SVD Singular value decomposition
URL Uniform Resource Locator
WGS 84 World Geodetic System 1984

Document Creation Date

June 2007

Document URL

http://nsidc.org/data/docs/daac/nsidc0304_0305_glas_dems.gd.html