← Back to Part 3

3.2 Plate Velocity Fields

NUVEL-1A: The First Modern Plate Motion Model

The NUVEL-1A model (DeMets et al., 1990; revised 1994) was the first comprehensive global plate motion model that combined three independent data types to constrain angular velocities for 16 major plates. The three data types are:

277

Spreading Rates

From magnetic anomaly profiles at mid-ocean ridges

121

Transform Azimuths

Strike directions of oceanic transform faults

724

Slip Vectors

Earthquake first-motion fault-plane solutions

Spreading rates provide the magnitude of relative velocity perpendicular to ridge crests, averaged over the last ~3 Ma (anomaly 2A). Transform azimuths constrain the direction of relative motion. Earthquake slip vectors give the direction of plate motion at subduction zones and transform boundaries. NUVEL-1A uses the revised geomagnetic polarity timescale (GPTS-94) which corrected a systematic ~4% overestimate in NUVEL-1 spreading rates.

MORVEL: The Current Standard

The MORVEL model (DeMets, Gordon & Argus, 2010) updated NUVEL-1A with significantly more data and expanded the plate set from 16 to 25 plates. Key improvements include:

  • Over 28,000 km of new magnetic anomaly profiles from previously unsurveyed ridges
  • Multibeam bathymetric mapping of transform faults for precise azimuth determination
  • Separate treatment of the Capricorn, Lwandle, and Somali plates (previously lumped into Indo-Australian and African plates)
  • GPS-derived constraints to validate geological rates and detect non-closure
  • Improved uncertainty estimation using bootstrap methods

MORVEL angular velocities represent average plate motions over the last ~0.78 Ma (since the Brunhes–Matuyama reversal), making them geological-timescale estimates. The model achieves typical misfits of 1–2 mm/yr when compared with GPS velocities, confirming that most major plates have moved at nearly constant rates over the Quaternary.

GPS Geodesy: Direct Velocity Measurement

Global Navigation Satellite Systems (GNSS/GPS) have revolutionized plate kinematics since the 1990s by providing direct, instantaneous measurements of station velocities to sub-millimeter per year precision. Continuous GPS networks (e.g., IGS, PBO/NOTA, GEONET) now include over 10,000 stations worldwide.

GPS velocity fields reveal that the rigid-plate model holds remarkably well for plate interiors: residual velocities after removing the best-fit Euler rotation are typically <1 mm/yr. However, plate boundary zones can be hundreds of kilometers wide, with distributed deformation that violates the rigid-plate assumption. The Basin and Range Province, the Aegean, and the Tibetan Plateau are prominent examples of diffuse deformation zones.

Geological vs. Geodetic Rate Comparison

For most plate pairs, geological rates (MORVEL, ~0.78 Ma average) agree with geodetic rates (GPS, ~20 yr average) to within 2–3 mm/yr. Significant discrepancies occur at:

  • India–Eurasia: GPS gives ~36 mm/yr convergence; MORVEL gives ~40 mm/yr — possible deceleration
  • Nazca–South America: GPS shows ~58 mm/yr vs. ~61 mm/yr geological — marginal difference
  • Pacific–North America: near-perfect agreement at ~50 mm/yr across the San Andreas system

Plate Velocity Closure Condition

For a set of three plates A, B, and C, the angular velocity vectors must satisfy a closure condition. This arises because if we know how A moves relative to B, and B relative to C, then A's motion relative to C is fully determined.

Plate Circuit Closure

\[ \boldsymbol{\omega}_{AB} + \boldsymbol{\omega}_{BC} + \boldsymbol{\omega}_{CA} = \mathbf{0} \]

This vector equation must hold exactly for perfectly rigid plates. Any residual non-closure in a plate circuit indicates either data errors, plate non-rigidity, or an unrecognized plate boundary. MORVEL uses circuit closure as a quality control criterion.

Velocity Gradient Tensor

\[ L_{ij} = \frac{\partial v_i}{\partial x_j} = \dot{\varepsilon}_{ij} + \dot{\omega}_{ij} \]

The velocity gradient tensor L decomposes into a symmetric strain rate tensor (ε̇) and an antisymmetric rotation rate tensor (ω̇). Within rigid plates, L = 0 everywhere; at boundaries, strain rates reach 10−15 to 10−14 s−1.

Selected Absolute Plate Velocities

Absolute plate velocities depend on the chosen reference frame. Below are representative velocities in the NNR (no-net-rotation) frame from NNR-MORVEL56, given at a characteristic point on each plate.

PlateSpeed (cm/yr)Azimuth (°)Reference Point
Pacific6.72950°N, 180°E
Australian6.302025°S, 135°E
Nazca5.507815°S, 90°W
India5.103520°N, 78°E
North America2.124940°N, 100°W
Eurasia2.106050°N, 10°E
Africa2.20320°N, 25°E
South America1.228615°S, 55°W
Antarctic1.118075°S, 0°E

NNR-MORVEL56 velocities. Fastest plates are those attached to large slabs (Pacific, Nazca, Australian); slowest plates lack significant slab pull (Antarctic, South American).

What Controls Plate Speed?

Fast Plates (~5–10 cm/yr)

The Pacific (~10 cm/yr in hotspot frame), Nazca, Australian, and Indian plates are the fastest movers. All have substantial attached subducting slabs, confirming that slab pull is the dominant driving force. The Pacific plate's large area and continuous subduction zones along its western and northern margins produce the highest velocity.

Slow Plates (~1–2 cm/yr)

The Antarctic, South American, and Eurasian plates are the slowest. These plates have minimal or no attached subducting slabs. Their motion is driven primarily by ridge push and basal drag from mantle convection. Continental plates with thick cratonic roots may also experience increased basal resistance.

Second Invariant of Strain Rate

\[ \dot{\varepsilon}_{II} = \sqrt{\frac{1}{2} \dot{\varepsilon}_{ij} \dot{\varepsilon}_{ij}} \]

The second invariant of the strain rate tensor is a scalar measure of deformation intensity. Global Strain Rate Map (GSRM v2.1) values range from ~10−16 s−1 in stable plate interiors to ~10−14 s−1 in active orogens like the Himalayas and the Andes.