Documentation for AGASEA Ice Thickness Profile Data from the Amundsen Sea Embayment, Antarctica, Version 1

Back to Top

Collapse All / Open All

Detailed Data Description

This data set contains line-based radar-derived ice thickness and bed elevation data, collected as part of the Airborne Geophysical Survey of the Amundsen Embayment (AGASEA) expedition, which took place over Thwaites Glacier in West Antarctica from 2004 to 2005. The data set includes ice thickness, ice sheet bed elevation, and ice sheet surface elevation, derived from ice-penetrating radar and aircraft GPS positions. The data are spaced on a 15 km by 15 km grid over the entire catchment of the glacier, and sampled at approximately 15 meters along track. Most of the radar data used for this dataset has been processed using a 1-D focusing algorithm, to reduce the along track resolution to tens of meters, to improve boundary conditions for ice sheet models.


Data are provided in ASCII format.

Background color on
File and Directory Structure

The directory contains two files;

  • ase1_foc1_thk.tllzzhr.txt is the ASCII text file that contains the data set.
  • crossovers.txt contains crossover information between orthogonal lines.
Background color on
File Naming Convention

File name example: ase1_foc1_thk.tllzzhr.txt



Table 1. File Naming Convention
Variable Description


Field season name


Dominant processing method 
(one-dimensional focused SAR)


Data product (ice thickness)


t = time 
l = longitude 
l = latitude 
z = elevation (bed) 
z = elevation (surface) 
h = ice thickness 
r = range to surface

Background color on
File Size

Two data files are 80 KB and 234 MB.

Background color on

The entire data set is approximately 234 MB.

Background color on
Spatial Coverage

Amundsen Sea Embayment, Antarctica, including Thwaites Glacier, Smith Glacier, and Haynes Glacier
Northernmost Latitude: 73°S
Southernmost Latitude: 80°S
Westernmost Longitude: 125° W
Easternmost Longitude: 90° W

Spatial Resolution

5 mm depth

Background color on
Temporal Coverage

10 December 2004 to 29 January 2005

Background color on
Parameter or Variable

Ice thickness

Ice sheet bed elevation

Ice sheet surface elevation

Parameter Description

Table 2 provides descriptions of each of the columns in the ase1_foc1_thk.tllzzhr.txt file.

Table 2. Description of columns in the ase1_foc1_thk.tllzzhr file
1 Year
2 Day of Year
3 Seconds of day
4 Longitude (decimal degrees)
5 Latitude (decimal degrees)
6 Northing (polar stereographic -71 meters)
7 Easting (polar stereographic -71 meters)
8 Bed elevation (meters; WGS-84)
9 Surface elevation (radar-derived; meters, WGS-84)
10 Ice thickness (meters; 169 m/usec; no firn correction)
11 Range to surface (meters)
12 Name of line (text)
13 Processing (text)

The cross_overs.txt file contains cross over information between orthogonal lines: The structure of the cross over file is seven space-delimited ASCII columns, as described in Table 3.

Table 3. Description of columns in the ase1_foc1_thk.tllzzhr file
1 Longitude
2 Latitude
3 Easting (polar stereographic -71 meters)
4 Northing (polar stereographic -71 meters)
5 Absolute cross over difference between lines (meters)
6 Name of first line in cross over
7 Name of second line in cross over
Background color on

Software and Tools

Data are readable using standard spreadsheet software.

Background color on

Data Acquisition and Processing

The AGASEA survey was conducted using a Twin Otter C-FSJB aircraft equipped with radar, laser, magnetometers, and a gravimeter. AGASEA is described in detail in Holt, et. al., 2006. The aircraft operated from two camps, one to the south of Thwaites Glacier, and one to the East, near Pine Island Glacier, with a range of 1000 km. Seventy-five flights were conducted over 50 days.

Data Acquisition Methods

Data were acquired using the High Capacity Radar Sounder (HiCARS) coherent ice penetrating radar delveloped by the University of Texas Institute of Geophysics. The radar antennas were mounted as dipoles under each wing. The radar had operated with a 60 MHz center frequency and 15 MHz of bandwidth, and used a 8 kW transmitter with a 1 μsec FM chirp and a pulse repetition frequency of 6400 Hz.

The across track beam width is controlled by the antenna beam pattern, and has a main cross-track half-power beam width of 12 degrees, and side lobes at ±22 degrees. The data were down converted to a 10 MHz center frequency, and recorded on two gain offset channels sampled at 50 MHz. Total dynamic range between the 2 channels is 90 dB. These data were coherently stacked 32 times to provide 16-bit 3200 sample coherent records at 200 Hz. In post processing, the data were range compressed with a synthetic chirp, and a 10 MHz coherent system noise was removed

Background color on
Derivation Techniques and Algorithms

Processing Steps

In this data set, two further forms of post processing were applied. In primary processing, (1-D focused SAR or "foc1"), in the frequency domain, matched correlation filters derived from GPS trajectories and surface range measurements were used to focus the data in the along track direction (Peters et al, 2007). These data were resampled to 4 Hz along track, converted to amplitude from complex data, and logarithmically scaled for interpretation. The effective beam width was approximately 10 meters on the bed.

The second form of processing ("pik1") was used where good GPS solutions were missing, or issues with aircraft recording prevented focusing. The data was coherently stacked, converted to amplitudes and incoherently stacked 5 times to yield a 4 Hz sampling of the along track data. Given an average aircraft velocity of 70 meters per second this yields a effective coherent aperture of 14 meters and an along-track half-power beam width of 10 degrees. These beam widths illuminate a patch on the bed (at 800 meters above the surface and 2000 meters ice thickness) of 414 meters along track and 920 meters across track (in the main and secondary lobe). For similar conditions the pulse limited footprint is 390 meters.

Once processed, the data were interpreted by a team of undergraduate research assistants, who defined bounds about the bed and surface returns in the high and low gain channels respectively. The closest return to the aircraft was selected for the bed. Between the bounds, the brightest echo with a parabolic intensity profile was automatically detected. No adjustment at cross overs was applied, to preserve the integrity of the error statistics along the full length of the line.

Background color on
Error Sources

Strictly speaking, what was obtained for each point along the transect was the range between the closest part of the bed over a ~2 km swath and the aircraft. Cross over differences between orthogonal lines are high over mountainous regions to the west and south (~150 meters) and low over the central catchment (~33 meters). The primary cause of these differences is cross track topography. The cross over differences for the entire foc1 data set are 102 meters, implying an uncertainty of 72 meters.

Background color on

References and Related Publications

Contacts and Acknowledgments

Investigators Name and Title

Dr. Donald D. Blankenship
University of Texas at Austin
Institue for Geophysics
Mail Code R2200
10100 Burnet Road
Austin, TX 78758 

Dr. Duncan Young
University of Texas at Austin
Institute for Geophysics
Mail Code R2200
10100 Burnet Road
Austin, TX 78758 


This research was supported by the National Science Foundation (NSF) Office of Polar Programs (OPP) grant number 0636724.

Document Information

Document Creation Date

April 12, 2012