SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. ICESat Data and Processing
  5. 3. Results
  6. 4. Discussion
  7. Acknowledgments
  8. References
  9. Supporting Information

[1] We present a technique for investigating the grounding zone (GZ) of Antarctic ice shelves using laser altimetry from the Ice, Cloud and land Elevation Satellite (ICESat). Most surface height variability in the GZ is easily resolved by the ICESat laser's ∼65 m footprint and ∼172 m along-track spacing. Comparisons of repeated tracks sampled at different phases of the ocean tide identify the landward and seaward limits of tide-forced ice flexure, providing GZ location and width information for each track. Using ICESat data in the Institute Ice Stream region of southern Ronne Ice Shelf, we demonstrate that the location of the GZ based on feature identification in satellite imagery or digital elevation models may be in error by several km. Our results show that ICESat will contribute significantly to improving knowledge of GZ structure and to studies requiring accurate GZ locations, e.g., ice mass balance calculations and ice-sheet/ocean modeling.

1. Introduction

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. ICESat Data and Processing
  5. 3. Results
  6. 4. Discussion
  7. Acknowledgments
  8. References
  9. Supporting Information

[2] The grounding zone (GZ) is the transition region between the fully grounded ice sheet and the free-floating ice shelf, the latter being in constant vertical motion with the ocean tides (see Figure 1). The width of the GZ across which the ice flexes (roughly between points F and H in Figure 1) is typically a few km. The GZ therefore occupies only a small fraction of the total area of Antarctic ice shelves; however, it is a critical gateway for grounded ice flowing off the continent into the ocean. The balance of forces controlling ice flow changes rapidly as the ice first begins to float, and ice dynamics may be further affected by rapid basal melt in the GZ [Rignot and Jacobs, 2002]. Recent surveys of Antarctica's ice mass budget suggest that the ice sheet is losing mass [Zwally et al., 2005] and it is well established that the location of the GZ can vary rapidly as ice thickness changes [Rignot, 1998a, 1998b]. Monitoring the GZ is therefore an important part of ice sheet change detection, the primary objective of the ICESat mission.

image

Figure 1. Schematic of an ice shelf GZ, adapted from Vaughan [1994]. F is the limit of ice flexure from tidal movement; G is the limit of ice flotation, I is the inflexion point where ice is depressed below the hydrostatic level due to longitudinal stresses associated with the tide-induced bending; and H is the inshore limit of the hydrostatic zone of free-floating ice shelf, or the seaward limit of ice flexure. We define the region between F and H as the GZ, which is typically several km wide. The exact distance between F, G, H and I will change depending on local ice thickness and properties, and local bedrock topography and properties. F and H are the points that are detected by InSAR. F is the landward limit of the dense fringe band which has also been referred to as the “hinge-line” [Rignot, 1998a, 1998b], and H is the seaward limit of the fringe band. F can also be detected through tiltmeter observations [e.g., Stephenson et al., 1979; Stephenson, 1984; Riedel et al., 1999], static [Reidel et al., 1999] and kinematic [Vaughan, 1994, 1995] GPS, and H can also estimated through buoyancy calculations [e.g., Fricker et al., 2002].

Download figure to PowerPoint

[3] Previous remote-sensing studies of the Antarctic GZ have used ERS and RADARSAT interferometric Synthetic Aperture Radar (InSAR) [e.g., Goldstein et al., 1993; Rignot, 1998a, 1998b; Gray et al., 2002] and ERS radar altimetry combined with radio-echo sounding data in buoyancy calculations [e.g., Fricker et al., 2002]. Although these studies have dramatically improved our understanding of the present location, structure and time variability of ice shelf GZs, each technique is limited in resolution, spatial coverage, or temporal sampling. In this paper we consider the contribution to GZ studies by the Geoscience Laser Altimeter System (GLAS) on ICESat. In particular, we present a technique for extracting information about the location and structure of the GZ from analyses of repeated ICESat tracks sampled at different phases of the ocean tide.

2. ICESat Data and Processing

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. ICESat Data and Processing
  5. 3. Results
  6. 4. Discussion
  7. Acknowledgments
  8. References
  9. Supporting Information

[4] Since October 04, 2003, ICESat has operated in a 91-day exact repeat orbit with a 33-day sub-cycle. For all 91-day orbit operations periods since the first (Laser 2a), the last 33 days of Laser 2a have been repeated. As of May 2006, ICESat has completed eight repeats of the same 33-day sub-cycle (Table 1). The GLAS altimetry channel (1064 nm) samples at 40 Hz or every ∼172 m along-track, with a footprint of ∼65 m. We obtained geo-located GLAS footprint locations, and ocean and load tide corrections, from the GLA12 product. We then merged these data with the return GLAS energy and receiver gain from the GLA01 product. These parameters are used to filter data that are affected by clouds.

Table 1. Dates and Latest Data Release Number, for the Eight 91-Day ICESat Operations Periods Acquired to May 2006a
OPS PERIODDATERELEASE
  • a

    Release number is associated with improvements in the data processing but does not necessarily correlate to absolute refinement.

Laser 2a10/04/03–11/19/0326
Laser 2b02/17/04–03/21/0426
Laser 2c05/18/04–06/21/0417
Laser 3a10/03/04–11/08/0423
Laser 3b02/17/05–03/24/0519
Laser 3c05/20/05–06/23/0522
Laser 3d10/21/05–11/24/0526
Laser 3e02/22/06–03/28/0626

[5] At the time of writing, GLA12 elevations are routinely corrected for ocean and load tide using the GOT99.2 global ocean model [Ray, 1999]. Since our analysis depends on measuring vertical displacements due to the tide, we do not want a tide correction applied, so we “retided” the data, i.e., added these corrections back. (We note here that GOT99.2 does not perform well in Antarctica relative to more recent global tide models [King and Padman, 2005]. Users who wish to work with detided data are, therefore, urged to retide the ICESat data and then apply a better model correction.) To account for saturation of the GLAS detector, we applied a saturation correction to the elevations for all points where receiver gain was equal to 13 and the return energy was greater than 13.1 fJ [Sun et al., 2005; X. Sun, personal communication, 2005]. This correction can be tens of centimeters [Fricker et al., 2005]. This correction is available in the GLA12 product for data Releases higher than 24.

[6] For each repeated track, we interpolated each transect of elevation hi onto a common set of evenly-spaced latitude (ϕ) values, then calculated the mean elevation profile for all repeats. We then defined the “elevation anomaly” hi(ϕ) for each repeat as the difference of each elevation profile from the mean profile. We also plotted gain and energy against ϕ for each repeat to assist with data filtering for clouds.

[7] Data from the ICESat operations periods currently exist at different levels of processing sophistication (hence the different data Release numbers in Table 1). The main difference between Releases is the successive refinement of instrument attitude biases; see Luthcke et al. [2005]. Remaining errors manifest themselves as elevation biases and currently limit the potential of using ICESat data for change detection [Fricker et al., 2005]. The ICESat Science Team aims to bring all data to the target pointing accuracy of 2 arcsec (1 arcsec of pointing error combined with a 1° slope leads to 5 cm error in footprint elevation) [Schutz et al., 2005]; however, this is not required for the current application. Here, we are concerned with studying the GZ, which has a much shorter length-scale than that of the elevation biases. Our technique uses the hi(ϕ) profiles, the shapes of which are relatively insensitive to biases.

[8] As a test of our technique, we analyzed all ICESat tracks from near the Institute Ice Stream as it enters the southern Ronne Ice Shelf (Figure 2). This region experiences a large tidal range [Padman et al., 2002], which facilitates demonstrating the technique. The availability of many repeats sampled at different stages in the tidal cycle also enables us to show that the technique can also be used in regions of much lower tidal amplitude (as low as ∼0.4 m range). Detailed analyses for two tracks, Track 218 and Track 1304, are presented below. Track 218 passes approximately along the center line of Institute Ice Stream, and Track 1304, ∼50 km to the east, crosses the grounding line near Bungenstockrücken.

image

Figure 2. Locations of the ICESat tracks (identified by white numbering) crossing the region near the entrance of Institute Ice Stream into the southern Ronne Ice Shelf. Tracks are color-coded by elevation relative to the WGS-84 ellipsoid. The background image is from MOA, and the black line is the break-in-slope from MOA (Scambos et al., manuscript in preparation, 2006). Asterisks indicate GZ features identified from our analyses of ICESat repeated tracks. Red cross on the inset map shows location of this region in Antarctica (RIS: Ronne Ice Shelf).

Download figure to PowerPoint

3. Results

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. ICESat Data and Processing
  5. 3. Results
  6. 4. Discussion
  7. Acknowledgments
  8. References
  9. Supporting Information

3.1. Track 218

[9] In the plot of hi(ϕ) for all repeats of Track 218 (Figure 3a), elevations differ between repeats at the northern end but are similar further south. We interpret this as evidence that the ice at the northern end is moving with the ocean tide, while the ice further south is grounded. Between these regions there is an inflexion point (cf. point I on Figure 1). Profiles of hi(ϕ) (Figure 3b) clearly identify floating and grounded ice, and the inshore and offshore limits of flexure. The profiles for Laser 2c and 3d are noisy, however, which we suspect is due to the presence of clouds: this is confirmed by plots of gain and return energy (Figures 3c and 3d). The preset lowest gain value for GLAS is 13, and where gain is at this value we are confident that there are no significant clouds. Laser 2c and Laser 3d have high gain values – for Laser 2c it is the maximum value of 250. The Laser 2b data are also noisy, for reasons that are less obvious, although there are some places where gain is >13. Since with this particular case we have many repeats acquired under clear conditions, we discard all repeats where gain >13. The maximum relative displacement of ∼4.5 m offshore for the remaining four repeats (Figure 3e) is a consequence of ICESat's temporal sampling: the modeled maximum tidal range at this location is ∼7.6 m, and the typical spring range is 6-7 m. We identify the inshore limit of flexure from Figure 3e as the point at which tidal displacement of the ice shelf in each repeat first becomes significant: this is point ‘F’ in Figure 1. On each repeat, F is in a slightly different location: we postulate that this is because it migrates with the tide. The gray box around F in Figure 3e encompasses all locations for F based on all repeats. We interpret the point on the track at which the tidal amplitude has reached its maximum as the seaward limit of the grounding zone, i.e., the hydrostatic point ‘H’ in Figure 1.

image

Figure 3. (a) ICESat “retided” elevation profiles vs latitude for the seven repeats of Track 218 (there are no Laser 3e data in this range); (b) elevation anomaly, (c) gain and (d) return energy values for each repeat; (e) elevation anomaly profiles calculated for the four repeats remaining after filtering based on Figures 3c and 3d. Horizontal lines at right on Figures 3b and 3e are the equivalent anomalies based on tide predictions from CATS02.01. The inflexion point (I) is marked on Figures 3a and 3e. Flexure limits (F and H) are shown on Figure 3e, where gray boxes indicate the range of values for F and H based on individual profiles.

Download figure to PowerPoint

3.2. Track 1304

[10] For Track 1304 (Figure 4), the most obvious feature in hi(ϕ) is the break-in-slope (marked as point I) near 81.0°S, several km south of point F. Downstream of the break-in-slope is a flat section extending about 10 km along-track, which we interpret as an “ice plain” similar to that found on Pine Island Glacier by Corr et al. [2001]. The flexure limits (F and H) are clearly identified, but there is no obvious inflexion point between F and H. Instead, hi increases offshore from F.

image

Figure 4. (a) ICESat “retided” elevations and (b) elevation anomalies for five repeats along Track 1304. Flexure limits (F and H) and the break-in-slope (I) are indicated, and gray boxes indicate the range of possible values for F and H based on individual profiles. (c) Elevation anomaly plots calculated from Laser 3a and 3b only, both acquired at low tide with a tidal difference of ∼0.37 m.

Download figure to PowerPoint

[11] The break-in-slope (point I) does not represent an inflexion point of the type shown in Figure 1. Instead, it corresponds to the “coupling line” described by Corr et al. [2001], and their Figure 3 is a more appropriate model of this region of the GZ than our Figure 1, which was drawn originally for Rutford Ice Stream [Vaughan, 1994].

4. Discussion

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. ICESat Data and Processing
  5. 3. Results
  6. 4. Discussion
  7. Acknowledgments
  8. References
  9. Supporting Information

[12] Full interpretation of the ICESat along-track data requires knowledge of the orientation of tracks to the ice topography across the GZ, therefore plotting the tracks on satellite imagery or a DEM from a similar epoch is essential. For this purpose we use the recent Mosaic of Antarctica (MOA) [Haran et al., 2005; T. A. Scambos et al., A MODIS-based Mosaic of Antarctica: MOA, manuscript in preparation, 2006], a compilation of multiple MODIS images with a consistent illumination. The MOA, shown for our study region as the background in Figure 2, allows us to convert the along-track distance of GZ features to distance normal to the GZ. For Tracks 1304 and 218, the width of the GZ is ∼4.2 and ∼6.4 km, respectively. The inflexion point for Track 218 and the break-in-slope for Track 1304 are coincident with topographic features visible in MOA. For these tracks, there are no visible expressions in MOA of the limits of ice flexure. The locus of the points F coarsely defines the grounding line (Figure 2). Since track spacing decreases from ∼15 km at 81°S for the present study toward ICESat's southern latitude of 86°S, the combination of ICESat with MOA clearly provides a mechanism for improved mapping of the GZ for the large southern Ross and Filchner-Ronne ice shelves, which will provide a baseline for future change measurements. This mapping task is currently underway. Improved knowledge of the present-day GZ combined with ICESat-derived surface elevations will also contribute to improved calculations of the ice sheet's mass balance.

[13] Beyond mapping applications, repeat-track analyses of ICESat data also provide information on physical structure around the GZ. The spread of locations for the grounding line (represented by the gray boxes around point F in Figures 3e and 4b) is greater for Track 1304 than for Track 218, even though the tidal range captured by ICESat is lower (∼2.1 m for Track 1304 vs. ∼4.5 m for Track 218). This result suggests that the bedrock slope is lower at the edge of Bungenstockrücken than along the main channel of Institute Ice Stream. ICESat also easily identifies ice plains which cannot easily be detected in satellite imagery. The diamond-shaped ice plain on the eastern side of the ice stream, previously discussed by Scambos et al. [2004] and visible in MOA (Figure 2), is clearly seen in the ICESat tracks which cross the plain. However, ICESat also reveals ice plains that are not easily discernible in MOA; e.g., the ice plain revealed by Track 1304 (Figure 4).

[14] Several factors need to be considered when assessing the value of this technique for a specific study. There are few repeats of each track: at the time of writing there is a maximum of eight. Clouds can cause reduced accuracy or total data loss for specific operations periods (see Figure 3, Laser 2a and 3d): some regions we have investigated have persistent cloud and few, if any, good data. The technique is sensitive to sampling of the tidal phase by the time separation of repeat tracks. It is possible to estimate the limits of flexure when the tidal range is as low as ∼40 cm (Figure 4c), although location errors (especially for point H) increase as the signal-to-noise ratio decreases. Finally, ICESat tracks do not repeat exactly, but have across-track separations of up to ±∼100 m. In regions of large across-track slope, significant differences in elevations between repeats can be caused by this track offset.

[15] In spite of these caveats, we have shown that ICESat provides valuable information about the Antarctic GZ which is not available from satellite imagery or DEMs. The GZ products from ICESat most closely compare with those from InSAR. Both sensors can clearly identify the flexure limits F and H given suitable tide height differences between the times of satellite passes. InSAR provides better spatial coverage, where image pairs exist, because it can locate the boundaries of the GZ throughout the area covered by the SAR swath instead of just along-track. ICESat coverage, however, extends to 86°S compared with ∼81.5°S for ERS and Envisat. Furthermore, ICESat's direct measurement of high-resolution along-track ice surface topography provides improved topographic context for interpreting the flexural signals. Comparisons of ICESat-based GZ locations with InSAR data acquired in the previous decade may identify regions north of 81.5°S that have undergone change that is sufficiently large to be detectable above the various measurement uncertainties.

Acknowledgments

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. ICESat Data and Processing
  5. 3. Results
  6. 4. Discussion
  7. Acknowledgments
  8. References
  9. Supporting Information

[16] This work was supported by NASA contract NAS5-99006 to ICESat Science Team member Bernard Minster, and NASA grant NNG05GR58G to ESR. We thank NASA's ICESat Science Project and the NSIDC for distribution of the ICESat data (see http://icesat.gsfc.nasa.gov and http://nsidc.org/data/icesat) and NSIDC for MOA. Ted Scambos provided the MOA coastline. We thank David Vaughan and Jonathan Bamber for constructive criticism of the original manuscript, and Jeremy Bassis, Richard Coleman, Eric Rignot, Bob Schutz and Christopher Shuman for helpful comments. This is ESR contribution number 75.

References

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. ICESat Data and Processing
  5. 3. Results
  6. 4. Discussion
  7. Acknowledgments
  8. References
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. ICESat Data and Processing
  5. 3. Results
  6. 4. Discussion
  7. Acknowledgments
  8. References
  9. Supporting Information
FilenameFormatSizeDescription
grl21865-sup-0001-t01.txtplain text document1KTab-delimited Table 1.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.