Dr. Stuart N. Thomson
Research Scientist
Department of Geosciences
University of Arizona

The Making of Antarctica's Fjords

Below is a vertically-exaggerated reconstruction of the Lambert Glacier and Prydz Bay region of East Antarctica both at the present day (a and b) and 34 million years ago (c).

  1. Modern ice cover from the RAMP radarsat dataset. The Lambert Glacier is the world's largest glacier (100 km wide, over 400 km long).
  2. Modern subglacial topography from the Bedmap 2 dataset. If the Lambert Glacier were removed, the Lambert Trough would form the world's largest fjord system, reaching depths of over 2300m (7600ft) below sea level).
  3. A reconstruction of the subdued, slowly eroding Antarctic landscape immediately before the onset of ice sheet expansion around 34 million years ago.
Our study published in the March 2013 edition of Nature Geoscience shows that the subglacial fjords of this sector of East Antarctica - including the 2345m (7690ft) deep Lambert Trough - were created almost entirely since initial expansion of the ice sheet around 34 million years ago (Eocene/Oligocene transition). Before this, Antarctica would have been characterized by subdued topography occupied by large slow-moving rivers.

This work was carried out as part of NSF Antarctic Earth Sciences award (ANT #0838722) conducted in collaboration with Peter Reiners, George Gehrels, and graduate student Clare Tochilin at the University of Arizona and Sidney Hemming at Columbia University.

Collaborative Research: Erosion history and sediment provenance of East Antarctica from multi-method detrital geo- and thermochronology

Publications and Abstracts

Work related to this project now published in Nature Geoscience

and Geochemistry, Geophysics, Geosystems

Thomson, S.N., Reiners, P.W., Hemming, S.R. & Gehrels, G.E. (2013). The contribution of glacial erosion to shaping the hidden East Antarctic landscape. Nature Geoscience, 6, p. 203-207, doi:10.1038/ngeo1722.

In the News:
Featured on NPR's "The Academic Minute", April 22nd, 2013
Nature Geoscience News & Views, February 27th, 2013
University of Arizona UANews, March 5th, 2013
Our Amazing Planet, March 7th, 2013
Scientific American, March 7th, 2013
Planetsave (blog), March 6th, 2013
Daily Mail (UK), March 9th, 2013

Tochilin, C.J, Reiners, P.W., Thomson, S.N., Gehrels, G.E., Hemming, S.R. & Pierce, E.L. (2012). Erosional history of the Prydz Bay sector of East Antarctica from detrital apatite and zircon geo- and thermochronology multidating. Geochemistry, Geophysics, Geosystems, 13, Q11015, doi:10.1029/2012GC004364.

Thomson, S.N., Reiners, P.W., Tochilin, C.J., Hemming, S.R. & Gehrels, G.E. (2011). The Cenozoic history of East Antarctic subglacial erosion and sediment flux from the offshore detrital thermochronometric record. Eos Transactions AGU, Fall Meeting Supplement, Abstract PP33B-1934.

Thomson, S.N., Reiners. P.W., Hemming, S.R., Cox, S.E. & Gehrels, G.E. (2011). An offshore thermochronometric record of post-Eocene East Antarctic subglacial erosion and landscape evolution. 11th International Symposium on Antarctic Earth Sciences, Edinburgh, UK.

Thomson, S.N., Hemming, S.R., Reiners, P.W. & Cox, S.E. (2009). Revealing the subglacial erosion and landscape evolution history below the East Antarctic ice sheet using detrital thermochronology. GSA Abstracts with Programs, Vol. 41, No. 7, p. 52


  • Novel single-grain fission-track, U-Pb, and (U-Th)/He dating approach.
  • From offshore sediments in Prydz Bay (ODP holes 739C; 1166A; and Holocene jumbo piston cores).
  • Samples from Cretaceous, late Eocene, early Oligocene, and from late Miocene through the Pliocene and Holocene.
  • Apatite U-Pb ages constrain sediment provenance and subglacial bedrock metamorphic and magmatic history.
  • FT and He ages record catchment pre-glacial (ca. 300 to 34 Ma) and glacial (post 34 Ma) thermal and erosional history.
  • Total new single grain detrital ages (August 2012):
        1209 Ap U-Pb ages
        2020 AFT ages
        214 AHe ages
        283 Zr U-Pb ages
        668 ZFT ages
        31 ZHe ages
        151 Hornblende Ar-Ar Ages

  • Figure 1a: 3D vertically exaggerated representation of East Antarctic ice surface (from the Bedmap 2 data-set) showing location of Lambert Glacier and offshore cores sampled in this study

    Figure 1b: 3D vertically exaggerated representation of East Antarctic subglacial topography (from the Bedmap 2 data-set, surface rebound adjusted isostatically for removal of ice) showing outline of Figure 2 and location of Prime Meridian and Antimeridian (International date-line)

    Figure 2: Map showing location of core hole samples, as well as newly acquired onshore AFT ages (black text) and apatite U-Pb ages (white text with black background) from moraines of the Pagadroma Group. Published bedrock ages shown as colored circles (Arne et al., 1993; Arne, 1994; Lisker et al., 2003; Lisker et al., 2007a; Lisker 2007b)

    Detrital Apatite Data Summary:

    Below are figures showing a summary of all the apatite U-Pb, fission track, and (U-Th)/He data that we have obtained so far from several different cores drilled in Prydz Bay (locations shown in Figures 1 and 2 above).

    Pre-34 Ma: A very slowly eroding stable craton since the Permian

  • Single-grain AFT AHe double dates fall in two clusters:
         1) Slow constant pre-34 Ma erosion rates of 0.005-0.03 km/Myr for over 200 Myr.
         2) Accelerated cooling between ~300 and 250 Ma at >0.05 km/Myr.

  • Cretaceous sediments (~84-94 Ma) show a ~110-120 Ma AFT age peak in Pan-African (~500 Ma U-Pb age) detrital apatite that supports local resetting by basic magmatism of this age.

  • Some AHe ages older than AFT ages can be explained by radiation damage effects inhibiting He diffusion during long residence at low temperatures.

  • Figure 5: (a) Apatite fission-track vs (U-Th)/He “double-date” plot from ODP hole 1166A hole late Eocene sandstones. Points close to the 1:1 line (green field) represent fast cooling at the time represented by the AFT and AHe ages. Points along the red line represent long-term constant erosion rates. Points below the red line (blue field) represent grains that have undergone accelerated cooling since the time represented by the AHe age. (b) AFT versus AHe “triple-date” plot with color-coded U-Pb ages from the same grains from a slightly younger early Oligocene sandstone.

    Figure 6: (a) Simple thermal model of rocks in the Lambert catchment that were at different temperatures (or depths) at 35 Ma. The paths are based on thermochronologic constraints (fast cooling at end Permian followed by very slow cooling until onset of widespread glaciation at ~34 Ma). (b) Predicted AFT and AHe age-depth profiles at 35 Ma using published annealing and diffusion models incorporated into HeFTy thermal modeling software (Ketcham, 2005). The long period of slow cooling produces a diagnostic rapidly decreasing age versus depth profile. (c) Predicted AFT and AHe age-depth profiles at 35 Ma with an additional 50°C mid-Creaceous heating episode.

    Post-34 Ma: Up to 3km localized glacial incision in Lambert Glacier catchment

  • Much younger AFT and AHe ages in late Miocene and younger glacial sediments require an increase in source region erosion rates to more than 0.05 km/Myr since the time represented by the AHe age (since 80 Ma for several grains).

  • Matching AFT-AHe age pairs to the thermal model predicted age-depth profile in Figure 6 shows grains were at temperatures of up to 60°C at 35 Ma, requiring about 2 to 3 km of incision since the onset of glaciation.

  • The spread in detrital and bedrock AFT ages is best explained by post-34 Ma glacial incision. We find NO evidence to indicate enhanced Cretaceous cooling or erosion linked to rifting.

  • AFT Data from early Oligocene in hole 739C imply most incision occurred during the early stages of Antarctic glaciation, within a maximum time frame of about 5 Myr.

  • Figure 7: AFT-AHe age pairs fitted to predicted 34 Ma age-depth profiles (from Figure 6 above) compared to Bedmap2 bedrock topography and single-dated grain age data.
    Location of section shown in (b).
    (b) Bedrock topography and regional maximum, mean, and minimum elevations calculated from Bedmap2 topography using a 100 km radius moving circle. The maximum elevation surface matches well with the projected remnants of a prominent near-planar pre-glacial land surface seen (where sporadically exposed) throughout much of East Antarctica. The current Bedmap2 bedrock surface represents an estimate of incision into the pre-glacial land surface by the Lambert Glacier and its tributaries following initial expansion of the East Antarctic ice sheet at ~34 Ma.
    (c) AFT-AHe double-dated grain age pairs fit to predicted age-depth curves from Figure 6 (above), providing estimate of temperature and depth of grains at ~34 Ma before onset of glacial erosion. These are overlain on the published subglacial bedrock topography shown in (b). AFT and AHe ages with open circle represent youngest Pliocene and younger grain ages not double-dated. The youngest AFT and AHe detrital grains ages signify those grains that were at greatest depth at ~34 Ma - and hence record the greatest amount of post-34 Ma erosion. When fit to the predicted age-depth curves (see dotted blue and green lines and black arrows), the youngest ages match well with the deepest part of the Lambert trough (~2 km below sea level, and ~ 3.5 km below the projected pre-glacial land surface), further supporting that excavation of this overdeepened trough was accomplished almost entirely since initial expansion of the EAIS at ~34 Ma.
    (d) Same as (c) but with AFT-AHe age pairs fit to predicted 34 Ma age-depth profiles with additional 50°C mid-Cretaceous reheating.

    Figure 8: Cartoon illustrating the concept of thermochronometric “stratigraphy”. If sudden incision occurs into an old stable landsurface where a steep age versus depth profile has developed, then ongoing incision will quickly yield apatite with younger AHe and AFT ages that previously resided at depth. The more incision that occurs, the younger the ages that will be found in detrital apatite deposited in Prydz Bay (or from the bedrock at the bottom of deeply incised troughs in the hinterland).

    Implications for Cenozoic Sediment Flux Estimates

  • If most incision occurred during early Oligocene (ca. 5 Myr), this would imply incision (erosion) rates of greater than 0.4 mm/yr at this time.

  • Minimum rock volume removed from incised troughs in area of Figure 10 = ~1.1 x 106 km3 (assumes little erosion of pre-glacial East Antarctic erosion surface).

  • Agrees well with post-34 Ma sediment volume estimated eroded source rock volume for Lambert catchment of 0.75 to 1.17 x 106 km3 (Wilson et al., 2012 - Palaeo-Cubed, v. 335-336, p. 24-34).

  • Minimum sediment flux rate estimates:
         Since 34 Ma = ~32,165 km3/Myr (= ca. 90 x 106 t/yr)
         If most during early Oligocene = ~218,000 km3/Myr (= ca. 625 x 106 t/yr)

    For comparison Yukon River has a sediment load of ~60 x 106 t/yr; Mississippi River has a load of ~400 x 106 t/yr; and the Amazon River has a load of ~1200 x 106 t/yr

  • Figure 9: Plot of geophysical (local) relief (elevation difference between a smooth surface connecting the highest points in the current landscape calculated using a moving 50km radius circle, and the current topography using Bedmap2. The values agree well with the depths of incision implyed by detrital apatite thermochronometry. The total volume represented by the geophysical relief in this figure is 1,100,000 km3.

    Last Modified: April 23rd, 2013