Research ArticleGEOLOGY

Kinematics and dynamics of the East Pacific Rise linked to a stable, deep-mantle upwelling

See allHide authors and affiliations

Science Advances  23 Dec 2016:
Vol. 2, no. 12, e1601107
DOI: 10.1126/sciadv.1601107
  • Fig. 1 RRTs and ridge axis migration relative to the fixed Indo-Atlantic reference frame with detailed inset (A) of the EPR.

    RRTs and ridge axis migration relative to the Indo-Atlantic hot spot frame of O’Neill et al. (21). Unshaded regions are for RRTs <1 My. Gray dotted lines are 33.5 Ma (C13no), and solid gray lines are 83 Ma (C34ny) isochrons from the age grid (19). The asymmetric distribution of the 33.5-Ma isochrons relative to the EPR on the Nazca and Pacific plates is illustrated. Red lines are the modern-day plate boundaries (75). Contour encloses regions with upwelling velocities of >2 cm/year at 650 km. (A) Top left inset map shows details of the RRTs along the EPR, with modeled radial flow velocity contour of 2 cm/year at 650-km depth. The predicted positions of the EPR were spreading to have been persistently symmetric over 50 Ma and 80 Ma, and are shown by the two isochrons. (B) South polar view of the same data. Red dots show the reconstructed positions of the present-day south pole (−90°) in the Indo-Atlantic frame of reference in 10-My intervals back to 80 Ma, demonstrating that it has drifted less than 5° latitudinally and hence has been essentially fixed in this frame of reference for the past 80 My.

  • Fig. 2 RRTs and ridge axis migration relative to the fixed Pacific hot spot reference frame (22).

    (Inset) South polar view (same as in Fig. 1B).

  • Fig. 3 Reconstruction of Pacific isochrons and associated uncertainties using various plate circuits and in the Pacific hot spot frame of reference.

    (A) Reconstructions using the more traditional Pacific–West Antarctica–East Antarctica plate circuit. (B) Our preferred reconstruction of EPR migration since 83 Ma along a plate circuit using constraints from the Emerald Basin (30) and assuming no earlier motion between the Campbell Plateau and Lord Howe Rise. Contours of modeled radial flow velocity in the underlying mantle are overlaid to show the spatial relationship between paleogeographic stability of the EPR axis and locus of mantle upwelling.

  • Fig. 4 Cumulative ridge-perpendicular migration relative to cumulative spreading at 10-My intervals from 10 Ma to 80 Ma.

    The thicker black dashed line (slope = 2) represents expected (half-spreading rate) ridge migration relative to a stationary plate. Horizontal error bars represent projections of the 95% confidence ellipses in the ridge-perpendicular direction. Ridges are MAR, Southwest Indian Ridge (SWIR), Central Indian Ridge (CIR), Carlsberg Ridge (CR), Southwest Pacific Ridge (SWPR), EPR, and Gorda Rise (GR). Plates are Eurasia (Eu), North America (NA), Greenland (Gr), Nubia (Nu), South America (SA), Somalia (So), India (In), Capricorn (Ca), East Antarctica (EA), Australia (Au), West Antarctica (WA), Pacific (Pa), Nazca/Farallon (Na), Cocos/Farallon (Co), and Juan de Fuca/Vancouver/Farallon (JdF).

  • Fig. 5 Subduction-buoyancy flux in the Pacific basin.

    (A) Comparison of the subduction-related buoyancy flux proxy for the Pacific plate relative to the Farallon and Farallon-derived plates over the past 78 My. Each symbol represents the per My relationship. Black dots are not scaled by area, and blue diamonds are scaled by plate area as a function of age. Gray line is the 1:1 relationship. (B) Relationship of length-weighted mean spreading rate of the EPR as a function of age derived from plate boundary polygons (72) and associated rotations (73) and the sum of Pacific and Farallon plate subduction-related buoyancy flux as a function of age. Each dot represents the per My relationship in the interval from 0 to 78 Ma. (C) Pacific relative to Farallon plate subduction-related buoyancy flux ratio as a function of age versus Pacific Plate spreading rate fraction along the EPR as a function of age from 3 Ma to 49 Ma based on Rowan and Rowley (24) (black dots). Pacific relative to Farallon plate subduction-related buoyancy flux ratio normalized by plate area as a function of age (blue diamonds). Each symbol represents the per My relationship. Note that Pacific fraction is ≤0.5, so Farallon is accreting asymmetrically faster.

  • Fig. 6 Geodynamic inferences of mantle viscosity and implications for shallow mantle flow beneath the EPR.

    The solid blue and green curves, labeled V1 and V2, respectively, are the depth-dependent effective viscosities derived in Occam-style inversions of combined glacial isostatic adjustment and convection-related data sets (41). The dashed gray lines illustrate the 2-σ uncertainties in the viscosity inference determined by varying the smoothing weights in the Occam inversions. The solid black curve, labeled L4, is a four-layer model (45) from fitting mantle flow to seismically inferred anisotropy below the African plate.

  • Fig. 7 Predicted present-day convective flow at three different depths in the mantle.

    (A) Asthenosphere. (B) Base of the transition zone. (C) Top of the seismic D″ layer. The mantle buoyancy distribution is given by model TX2008, obtained from joint seismic-geodynamic inversions by Simmons et al. (40). The viscous response of the mantle is calculated on the basis of the “V2” viscosity profile (39), derived from the joint glacial isostatic adjustment convection inversions (41) shown in Fig. 6. The flow calculations are described in detail by Forte et al. (39).

  • Fig. 8 Cross sections across the EPR showing mantle density anomalies as a function of depth together with computed flow velocities.

    (A) Mantle flow at −55 Ma [before the present (B.P.)] predicted on the basis of a tomography-based, time-reversed mantle convection simulation (51) that is initiated with the same present-day flow shown in Fig. 7, based on the joint tomography model TX2008 (40) and geodynamically inferred V2 viscosity profile (Fig. 6). (B) Present-day mantle flow with surface plates whose motions are predicted on the basis of viscous coupling to the underlying buoyancy-driven flow in the mantle. This is the same flow calculation used in Fig. 7 that also serves as the starting point for the time-reversed simulation in (A). (C) Present-day mantle flow calculated with a global, rigid-surface boundary condition. All other inputs (for example, mantle buoyancy and viscosity structure) are identical to those used in (B). (D) Evolution of the flow pattern shown in (C) after 100 My, calculated by forward integration in the tomography-based mantle convection simulations with a rigid-surface boundary condition by Glišović et al. (44) and subsequently filtered to maximum harmonic degree 32 to simulate the effective filtering arising from the tomographic inverse procedure (63). The amplitude spectrum of this filtered prediction agrees closely with the present-day structure in (B).

  • Fig. 9 Cross sections across the EPR showing the vertical and horizontal components of buoyancy-driven asthenospheric flow at 250-km depth.

    The blue curves in (A) to (D) show the component of the horizontal flow in the plane of the cross sections presented in Fig. 8. The black curves show the vertical component of the flow.

  • Fig. 10 Horizontal components of lithosphere and asthenosphere flow.

    (A) Horizontal component of mantle flow under the EPR region in the mid-lithosphere (50-km depth) and asthenosphere (250-km depth) calculated using the three viscosity profiles in Fig. 6. In each case, the mantle density anomalies are from the joint seismic-geodynamic tomography model TX2008 (40). The flow is represented in the same mantle cross section as in Figs. 8 and 9. Positive and negative values indicate eastward- and westward-directed flow, respectively. (B) Observed and predicted surface lithospheric flow velocity, projected onto the same cross section, as in (A). The blue curve represents the NUVEL-1A plate velocity in the NNR reference frame. The black and red curves show the prediction obtained using the V1 and V2 viscosity profiles, respectively (Fig. 6). The dashed curves show the predicted lithospheric velocity when we remove the positive buoyancy below the EPR at all depths in the lower mantle (curves labeled δρ > 670 km) and at all depths below the asthenosphere (curve labeled δρ > 250 km). In these flow tests, we searched and removed all positive buoyancy in a geographic region centered on the EPR axis, extending from 20°N to 50°S and from 90°W to 120°W. In all cases, the lithospheric flow is represented up to spherical harmonic degree 32 (hence, the slight oscillations in the flow profiles).

  • Fig. 11 3D representation of the buoyancy distribution (shaded volumes) and corresponding flow field (cones) centered on the EPR.

    Lower boundary is the CMB, and upper boundary is the surface of Earth. Black lines are plate boundaries from Bird (75). Coastline of South America is shown as the speckled line. Blue shaded volumes are characterized by δρ/ρ ≥ 0.1, and red shaded volumes are δρ/ρ ≤ −0.1. Axes of the cones point in flow direction, and the size is proportional to the velocity. Note the clear asymmetry of the flow velocities on either side of the EPR.

  • Fig. 12 Dependence of mantle flow under EPR on distribution of deep-mantle buoyancy.

    (A) Mantle flow predicted on the basis of the S20RTS tomography model converted to density using the constant density-velocity scaling used by Behn et al. (45) and Conrad and Behn (46). The L4 viscosity profile (see Fig. 6) is used in this calculation. (B) Mantle flow predicted using the density anomalies from the TX2008 tomography model (40) and the V2 viscosity profile (see Fig. 7). (C) Mantle flow predicted using an optimized, Occam-inverted, depth-dependent density-velocity scaling [see Forte (42) for details] of the S20RTS tomography model. The V2 viscosity profile is used in this calculation. (D) Mantle flow predicted using the TX2008 density anomalies and the L4 viscosity profile. The mantle cross section used here is identical to that in Fig. 8.

  • Fig. 13 Free-air gravity anomalies in the central and eastern Pacific.

    (A) Observed free-air gravity anomalies from the GRACE geopotential model (65). (B) Free-air gravity anomalies predicted by the TX2008-based mantle flow model shown in Fig. 12B. (C) Free-air gravity anomalies predicted by the S20RTS-based mantle flow model shown in Fig. 12A. (D) Free-air gravity anomalies predicted when the mantle flow calculated in (C) is modeled with a free-slip surface boundary rather than viscously coupled tectonic plates. All fields are truncated at spherical harmonic degree = 16.

  • Table 1 Global fits to convection-related observables, using TX2008 mantle density anomalies.

    All fits, with the exception of the CMB ellipticity, are expressed as percent variance reduction.

    Viscosity
    model
    Plate
    velocities*
    Free-air
    gravity
    GeoidDynamic
    topography
    CMB
    ellipticity§
    V190%68%99%66%430 m
    V282%61%90%65%540 m
    L455%40%41%61%690 m

    *Fits between the predicted and NUVEL-1A vector field of plate velocities (48), in the NNR frame, are calculated on a global 5° × 5° grid.

    †Fits are relative to the global free-air gravity and nonhydrostatic geoid anomalies derived from the GRACE geopotential solution (65), truncated at spherical harmonic degree 32.

    ‡Fits to the CRUST2.0-corrected dynamic topography (42) are for predicted and observed fields truncated at spherical harmonic degree 32.

    §The space-geodetic inference of excess CMB ellipticity is 400 m (76).

    • Table 2 Global fits to long-wavelength ( = 20) convection-related observables.

      All fits, with the exception of the CMB ellipticity, are expressed as percent variance reduction, where the predicted and observed fields are truncated at spherical harmonic degree = 20. N/A, not applicable.

      Flow
      model
      Plate
      velocities*
      Free-air
      gravity
      GeoidDynamic
      topography
      CMB
      ellipticity**
      S20RTS + L444%−7%−60%−50%1440 m
      S20RTSfs + L4N/A12%64%−17%1430 m
      S20RTSinv+ V2§48%28%57%40%450 m
      TX2008 + V280%74%90%73%540 m
      TX2008 + L4||60%50%41%69%690 m

      *Fits between the predicted and NUVEL-1A vector field of plate velocities (48) in the NNR frame are calculated on a global 5° × 5° grid.

      †Fits are calculated for the flow model using a posteriori density conversion of tomography model S20RTS used by Behn et al. (45) and L4 viscosity model (Fig. 6). Plate motions are viscously coupled to underlying mantle flow.

      ‡Fits are calculated for a theoretical free-slip surface boundary condition (no surface plates) using identical density and viscosity inputs as in the previous footnote (see previous model above).

      §Fits are calculated for the flow model using optimized density-velocity scaling of model S20RTS from Occam inversion of surface geodynamic data (42) and using V2 viscosity model (Fig. 5). Plate motions are viscously coupled to underlying mantle flow.

      ¶Fits are calculated for the flow model using the TX2008 density heterogeneity model and the V2 viscosity model. Plate motions are viscously coupled to underlying mantle flow.

      ||Fits are calculated for the flow model using the TX2008 density heterogeneity model and the L4 viscosity model. Plate motions are viscously coupled to underlying mantle flow.

      **The space-geodetic inference of excess CMB ellipticity is 400 m (76).

      Supplementary Materials

      • Supplementary material for this article is available at http://advances.sciencemag.org/cgi/content/full/2/12/e1601107/DC1

        fig. S1. RRT in the NNR frame of reference from present to 50 Ma.

        fig. S2. Data sources for ridge drift calculations.

        fig. S3. Subduction-related buoyancy flux components of the Pacific and Farallon-related plates.

        table S1. Rotation poles and associated covariance parameters used in our reconstructions.

        table S2. Rotation poles for the major mid-ocean ridges relative to the Indo-Atlantic hot spot reference frame, listed in 5-My increments.

        References (7791)

      • Supplementary Materials

        This PDF file includes:

        • fig. S1. RRT in the NNR frame of reference from present to 50 Ma.
        • fig. S2. Data sources for ridge drift calculations.
        • fig. S3. Subduction-related buoyancy flux components of the Pacific and Farallon-related plates.
        • table S1. Rotation poles and associated covariance parameters used in our reconstructions.
        • table S2. Rotation poles for the major mid-ocean ridges relative to the Indo-Atlantic hot spot reference frame, listed in 5-My increments.
        • References (77–91)

        Download PDF

        Files in this Data Supplement:

      Stay Connected to Science Advances

      Navigate This Article