Research ArticleCLIMATOLOGY

Sea level fingerprinting of the Bering Strait flooding history detects the source of the Younger Dryas climate event

See allHide authors and affiliations

Science Advances  26 Feb 2020:
Vol. 6, no. 9, eaay2935
DOI: 10.1126/sciadv.aay2935
  • Fig. 1 Bathymetry in the Bering Shelf region from ARDEM (11).

    Locations of observations constraining the timing of the Bering Strait resubmergence are shown by circles, triangles, and squares, which correspond to markers in Figs. 2 and 5. Pacific species found on uplifted beach terraces were collected on Victoria and Banks Island, Canada (79). Dotted gray line and shaded region show schematic of Last Glacial Maximum ice margin [Dyke (1)].

  • Fig. 2 Relative sea level predictions for sites in the Bering Strait region compared with observations on the timing of Bering Strait resubmergence.

    Predictions at the Bering Strait sill are shown by the black line. Present-day sill depth is shown by horizontal dotted red line at −53 m. Upward-pointing triangles denote dates on marine deposits, implying relative sea level higher than the plotted elevation. Downward-pointing triangles denote terrestrial deposits, implying relative sea level lower than the plotted elevation. Open circles are observations signaling an open connection at the Bering Strait (e.g., Pacific species found in the Arctic), which means that Bering Strait had flooded sufficiently for tidal flows to allow species dispersal in a planktonic larval (veliger) stage, whereas squares are observations signaling significant flows through the Bering Strait. Error bars represent 1σ uncertainties (see the Supplementary Materials). Overlapping filled and open squares or triangles represent geochemical evidence of a transition from a closed to a significantly open connection at the Bering Strait between the Arctic and Pacific Oceans with net northward water flows. The gray rectangle highlights the interval from 13 to 11.5 ka ago. The blue, red, and orange lines show relative sea level histories at the corresponding site of observation (Fig. 1) using ice history GI-31.

  • Fig. 3 Eustatic contributions of each ice sheet.

    Cordilleran and western Laurentide Ice Sheet (CIS) in black, Laurentide Ice Sheet (east of 110°W; LIS) in gray, Fennoscandian Ice Sheet (FIS) in blue, and Antarctic Ice Sheet (AIS) in red, for both ICE-6G (solid) and an alternate ice model, GI-31 (dashed). The shaded gray rectangle highlights the interval of 13 to 11.5 ka. The Younger Dryas (YD; 12.9 to 11.7 ka) and Meltwater Pulse-1a (MWP-1a; 14.5 to 14 ka ago) are labeled. The inset compares total global eustatic histories for ICE-6G (gray) and GI-31 (pink), which differ by less than 2 m from 13 to 11.5 ka ago.

  • Fig. 4 Ice melting scenario from 13 to 11.5 ka ago.

    Contours at 500-m intervals represent ice thickness at 13 ka ago in the GI-31 ice history (right-hand colorbar). Squares represent calibrated radiocarbon ages on organic material postdating ice loss, circles represent luminescence ages on aeolian dune deposits, and triangles represent cosmogenic ages on moraines, including cirque and valley moraines (see fig. S8). The GI-31 ice distribution from 15 to 13 ka ago is consistent with the median ages reported in this dataset (n = 818), constraining the deglaciation chronology of this region. For any dates that round to 13 ka ago, we ensure that these sites become ice free in the time step from 13 to 12.75 ka ago. Interior colors represent ages rounded to nearest integer (top colorbar). Dashed gray lines enclosing the areas filled with gray shading represent the limits of ice extent at 11.5 ka ago. Inset shows the map area in black rectangle. The magenta line shows the zero thickness contour at 13 ka ago.

  • Fig. 5 Map of relative sea level predictions at 13 and 11 ka ago.

    Sites with observations on resubmergence of the Bering Strait are shown by upward- and downward-pointing triangles, which denote marine or terrestrial deposits, respectively. Adjacent white boxes show dates associated with each observable. At 13 ka ago, the relative sea level at the Bering Strait sill is higher than the −53-m threshold; however, the relative sea level is lower at other sites on the Bering Shelf.

  • Table 1 Observational data and calibrated radiocarbon ages.

    ReferenceSite
    name
    LatitudeLongitudeMaterialMarker
    depth (m)
    Raw 14C
    age
    (years)
    Raw
    age
    error
    ΔRΔR
    error
    Calendar
    year
    (before
    present)
    1σ−
    years
    1σ+
    years
    Calibration
    curve
    Jakobsson
    et al. (4)
    4PC-172.8175.7Mollusc124.0710,2003025020010,900230510Marine13
    Keigwin
    et al. (5)
    JPC0267.4165.6Elphidium
    excavatum
    53.45510,90014025020011,950350490Marine13
    Hill and
    Driscoll (6)
    VBC03*70.7165.4Marine
    bivalve
    (Portlandia)
    60.4211,50076525020012,5901120910Marine13
    JPC1070.8165.5Marine
    bivalve
    59.210,2005525020010,910240250Marine13
    Elias et al. (12)85-6970165.7Screened
    peat
    44.9511,000600012,8709080Intcal13
    Dyke and Savelle (8)Bowhead whale bone10,210703205010,850140130Marine13
    Dyke et al. (9)Macoma
    balthica
    11,4001003205012,60090110Marine13
    England and Furze (7)Cyrtodaria
    kurriana
    12,3801103205013,520140130Marine13
    England and Furze (7)C. kurriana12,170253205013,3206060Marine13
    England and Furze (7)C. kurriana11,800703205012,960110121Marine13

    *Portlandia arctica raw 14C age reflects correction (800 ± 700 years subtracted from published raw 14C age).

    Supplementary Materials

    • Supplementary material for this article is available at http://advances.sciencemag.org/cgi/content/full/6/9/eaay2935/DC1

      Section S1. Uncertainty on elevation of Bering Strait sill at 13 to 11.5 ka ago

      Section S2. Local observations of flooding as sea level markers

      Section S3. Contributions to relative sea level: Gravitational versus deformational effects

      Section S4. Sensitivity to ice model

      Section S5. Meltwater flux volumes to Arctic Ocean

      Section S6. Meltwater pulses recorded in Arctic

      Section S7. Terrestrial geologic data constraining CIS and LIS retreat

      Section S8. Radiocarbon reservoir age corrections

      Section S9. Fitting relative sea level constraints in far field

      Section S10. Fitting glacial lake shoreline tilts and local relative sea level histories

      Fig. S1. Relative sea level predictions at each site of observation for ice model GI-31 and ICE-6G adopting the Earth model described in the main text.

      Fig. S2. Ice thickness at 13 ka ago for various ice histories.

      Fig. S3. Snapshots of ice thickness from 13 to 11.5 ka ago for ice history GI-31 (ice history adopted in the main text).

      Fig. S4. Decomposition of total relative sea level at the Bering Strait sill into components associated with the direct gravitational effect of the surface load and crustal deformation, including the local gravitational effect of this deformation.

      Fig. S5. Map of the difference in relative sea level predictions at 13 ka BP predicted using the GI-31 and ICE-6G ice histories and the Earth model described in the main text.

      Fig. S6. Ice melting scenario from 13-11.5 ka ago for GI-34 and GI-30.

      Fig. S7. Relative sea level predictions based on a suite of ice models, testing the sensitivity of the predictions to changes in the regional distribution and duration of ice melt.

      Fig. S8. Earth model sensitivity.

      Fig. S9. Oxygen isotope record from planktonic foraminifera from Mackenzie delta.

      Fig. S10. The location of cirque and valley glacier moraines in the Menounos et al. (18) study is shown on Fig. 4.

      Fig. S11. Relative sea level predictions for sites in the Bering Strait region compared with observations using radiocarbon dates calibrated with additional uncertainty.

      Fig. S12. Relative sea level predictions using ice history GI-31, GI-30, and GI-34.

      Fig. S13. Comparison of measured and predicted tilt using ice history NAICE-D and NAICE.

      Fig. S14. Paleotopography compared to observed shoreline elevations.

      Fig. S15. Misfit between the observed and predicted paleotopography for each glacial Lake Agassiz shoreline.

      Fig. S16. Relative sea-level predictions compared with relative sea-level markers in the Arctic older than 11 ka ago.

      Fig. S17. Possible marine retreat of ice sheet.

      Table S1. Compilation of ice models used in this study.

      Table S2. Calibrated radiocarbon ages using a larger uncertainty on reservoir ages than in main text (ΔR = 300 ± 200 years).

      Data file S1. Compilation of ages constraining timing of ice retreat in CIS/Western LIS.

    • Supplementary Materials

      The PDF file includes:

      • Section S1. Uncertainty on elevation of Bering Strait sill at 13 to 11.5 ka ago
      • Section S2. Local observations of flooding as sea level markers
      • Section S3. Contributions to relative sea level: Gravitational versus deformational effects
      • Section S4. Sensitivity to ice model
      • Section S5. Meltwater flux volumes to Arctic Ocean
      • Section S6. Meltwater pulses recorded in Arctic
      • Section S7. Terrestrial geologic data constraining CIS and LIS retreat
      • Section S8. Radiocarbon reservoir age corrections
      • Section S9. Fitting relative sea level constraints in far field
      • Section S10. Fitting glacial lake shoreline tilts and local relative sea level histories
      • Fig. S1. Relative sea level predictions at each site of observation for ice model GI-31 and ICE-6G adopting the Earth model described in the main text.
      • Fig. S2. Ice thickness at 13 ka ago for various ice histories.
      • Fig. S3. Snapshots of ice thickness from 13 to 11.5 ka ago for ice history GI-31 (ice history adopted in the main text).
      • Fig. S4. Decomposition of total relative sea level at the Bering Strait sill into components associated with the direct gravitational effect of the surface load and crustal deformation, including the local gravitational effect of this deformation.
      • Fig. S5. Map of the difference in relative sea level predictions at 13 ka BP predicted using the GI-31 and ICE-6G ice histories and the Earth model described in the main text.
      • Fig. S6. Ice melting scenario from 13-11.5 ka ago for GI-34 and GI-30.
      • Fig. S7. Relative sea level predictions based on a suite of ice models, testing the sensitivity of the predictions to changes in the regional distribution and duration of ice melt.
      • Fig. S8. Earth model sensitivity.
      • Fig. S9. Oxygen isotope record from planktonic foraminifera from Mackenzie delta.
      • Fig. S10. The location of cirque and valley glacier moraines in the Menounos et al. (18) study is shown on Fig. 4.
      • Fig. S11. Relative sea level predictions for sites in the Bering Strait region compared with observations using radiocarbon dates calibrated with additional uncertainty.
      • Fig. S12. Relative sea level predictions using ice history GI-31, GI-30, and GI-34.
      • Fig. S13. Comparison of measured and predicted tilt using ice history NAICE-D and NAICE.
      • Fig. S14. Paleotopography compared to observed shoreline elevations.
      • Fig. S15. Misfit between the observed and predicted paleotopography for each glacial Lake Agassiz shoreline.
      • Fig. S16. Relative sea-level predictions compared with relative sea-level markers in the Arctic older than 11 ka ago.
      • Fig. S17. Possible marine retreat of ice sheet.
      • Table S1. Compilation of ice models used in this study.
      • Table S2. Calibrated radiocarbon ages using a larger uncertainty on reservoir ages than in main text (ΔR = 300 ± 200 years).
      • Legend for data file S1

      Download PDF

      Other Supplementary Material for this manuscript includes the following:

      • Data file S1 (Microsoft Excel format). Compilation of ages constraining timing of ice retreat in CIS/Western LIS.

      Files in this Data Supplement:

    Stay Connected to Science Advances

    Navigate This Article