A moving Earth, measured to the millimetre
Plate boundaries by kind (teeth on the overriding plate along subduction zones), present plate motion measured by GNSS, and earthquakes of magnitude 6 and above since 1976.
- Subduction
- Ridges and rifts
- Transform faults
- Continental collisions
- Plate motion
- Earthquakes Mw ≥ 6
Particles: present plate motion (Euler poles fitted on GNSS stations, PB2002 for plates without stations), much exaggerated: one cycle of the animation stands for about 10 million years at the present rate. Dots: Global CMT earthquakes of Mw ≥ 6, orange shallow (< 70 km), pink intermediate, violet deep (> 300 km).
Three independent data sets are brought together. The PB2002 plate model (Bird, 2003) divides the surface into 52 plates and classes its 5,819 boundary steps: subduction, ridge, transform fault, rift, collision. GNSS velocities from the Nevada Geodetic Laboratory give the present motion of 19,993 stations: 1,946 of them, far from boundaries, are used to fit the rotation of 14 plates (Euler poles), 10 of which are rigid to better than 1.5 mm/yr. The Global CMT catalogue provides 71,453 focal mechanisms (1976 → 2026-10-28), which tell how the rock broke: in extension, in strike-slip or in compression.
Main result: plates move today as they did on average over the last three million years. For the rigid plates, the rotation axis measured by GNSS differs by only 1.3° (median) from that of the geological model, and the rotation rate by 2 % (median).
Three ways for two plates to meet
- 94,805 km of boundaries
- 3.26 km²/yr of sea floor created
- mean relative speed 46 mm/yr (up to 202)
- 60 % of earthquakes on normal faults (extension)
- 73,918 km of boundaries
- mean relative speed 40 mm/yr (up to 263)
- 67 % of earthquakes in strike-slip
- San Andreas: 48 mm/yr of right-lateral slip (GNSS)
- 51,309 km of subduction zones, 40,450 km of other convergent boundaries
- 3.29 km²/yr consumed
- 68 % of earthquakes on thrust faults (compression)
- 68 % of the shallow seismic moment; 96 % of deep earthquakes
Sketches not to scale; numbers computed from the data of this page. Regimes are those of shallow earthquakes (≤ 70 km) within 100 km of a step of the kind shown (oceanic ridges, oceanic transform faults, subduction zones).
Plate rotation measured by GNSS
GNSS station velocities, MIDAS, IGS20 frame (Nevada Geodetic Laboratory) · Free use with citation of Blewitt et al. (2016, 2018) · 10/28/26Each rigid plate rotates about an axis through the centre of the Earth: its Euler pole. It is fitted on the velocities of stations in the interior of the plate (1,946 stations, IGS20 frame, file dated 2,026.75). The last column compares the rotation obtained, relative to the Pacific plate, with that of the geological model PB2002 (sea-floor magnetic anomalies, transform faults and focal mechanisms, averaged over about 3 million years): angle between the two axes and difference in rate.
| Plate | Area (10⁶ km²) | Mean speed (mm/yr) | GNSS pole (lat., lon.) | Rotation (°/Myr) | Stations | Residual (mm/yr) | Difference from geology | Fit |
|---|---|---|---|---|---|---|---|---|
| Pacific PA | 104.5 | 68 | -62.6° ± 0.1, 109.9° ± 0.3 | 0.6815 ± 0.0008 | 26/30 | 0.38 | — | rigid |
| Africa AF | 58.6 | 27 | 49.7° ± 0.1, -80.8° ± 0.3 | 0.2660 ± 0.0004 | 85/88 | 0.37 | 0.4°, +2 % | rigid |
| Antarctica AN | 57.9 | 13 | 60.0° ± 0.4, -127.4° ± 0.4 | 0.2179 ± 0.0023 | 52/56 | 0.65 | 0.1°, +1 % | rigid |
| North America NA | 55.4 | 18 | -6.6° ± 0.2, -87.6° ± 0.1 | 0.1878 ± 0.0006 | 725/789 | 0.61 | 1.4°, +2 % | rigid |
| Eurasia EU | 48.6 | 25 | 54.2° ± 0.1, -100.6° ± 0.2 | 0.2552 ± 0.0004 | 406/416 | 0.39 | 2.9°, +8 % | rigid |
| Australia AU | 46.1 | 62 | 32.4° ± 0.0, 38.2° ± 0.1 | 0.6289 ± 0.0004 | 263/273 | 0.29 | 1.2°, +0 % | rigid |
| South America SA | 41.9 | 12 | -19.6° ± 0.2, -134.5° ± 0.6 | 0.1157 ± 0.0005 | 219/236 | 0.53 | 1.6°, +7 % | rigid |
| Somalia SO | 19.2 | 25 | 49.5° ± 0.3, -94.0° ± 0.8 | 0.3056 ± 0.0027 | 38/39 | 0.67 | 1.2°, −0 % | rigid |
| Nazca NZ | 16.1 | 72 | not enough stations: PB2002 rotation, attached to the Pacific plate measured by GNSS | PB2002 | ||||
| India IN | 12.4 | 54 | 51.5° ± 0.1, 6.3° ± 1.2 | 0.5267 ± 0.0033 | 20/21 | 0.47 | 3.2°, +3 % | rigid |
| Sunda SU | 8.8 | 37 | 72.8° ± 2.4, -145.6° ± 20.0 | 0.2444 ± 0.0050 | 25/26 | 2.27 | 13.2°, −18 % | deforming |
| Philippine Sea PS | 5.6 | 44 | not enough stations: PB2002 rotation, attached to the Pacific plate measured by GNSS | PB2002 | ||||
| Amur AM | 5.3 | 30 | 55.6° ± 11.4, -89.8° ± 15.4 | 0.2452 ± 0.0160 | 34/34 | 4.41 | 5.2°, −1 % | deforming |
| Arabia AR | 4.9 | 43 | 51.2° ± 0.3, -8.4° ± 2.3 | 0.5045 ± 0.0110 | 10/10 | 0.50 | 3.5°, −2 % | rigid |
| Okhotsk OK | 3.0 | 23 | -42.1° ± 3.6, -47.9° ± 6.0 | 0.5259 ± 0.1595 | 18/19 | 5.95 | 40.0°, −13 % | deforming |
| Caribbean CA | 3.0 | 15 | 33.2° ± 0.7, -95.3° ± 1.4 | 0.2887 ± 0.0112 | 25/25 | 0.94 | 1.0°, +3 % | rigid |
“Deforming”: mean residual above 1.5 mm/yr, a sign that the plate is not rigid or that stations move for other reasons (deformation after the great Japanese earthquakes for Amur and Okhotsk, Sumatra for Sunda). The mean speed is that of the plate surface in the GNSS frame. Uncertainties at 1 σ, scaled by the reduced χ².
Some famous boundaries
| Place | Kind | Plates | Relative speed (mm/yr) | Opening (+) / closing (−) | Right-lateral slip | PB2002 |
|---|---|---|---|---|---|---|
| Iceland (Mid-Atlantic Ridge) | Oceanic ridge | Eurasia / North America | 15.8 | +14.1 | -7.2 | 18.6 |
| East Pacific Rise (20° S) | Oceanic ridge | Nazca / Pacific | 147.8* | +147.6 | -6.9 | 147.8 |
| San Andreas Fault | Continental transform fault | North America / Pacific | 48.0 | -3.1 | 47.9 | 45.7 |
| Himalaya (Nepal) | Continental collision | Eurasia / India | 38.2 | -38.1 | 2.7 | 48.3 |
| Japan Trench (Tohoku) | Subduction | Pacific / Okhotsk | 92.0* | -88.8 | -23.8 | 92.0 |
| Tonga Trench | Subduction | Tonga / Pacific | 175.3* | -174.7 | -13.8 | 175.3 |
| Chile Trench (31° S) | Subduction | Nazca / South America | 75.9* | -72.4 | 22.9 | 80.6 |
| Ethiopian Rift | Continental rift | Somalia / Africa | 4.1 | +3.5 | -2.1 | 5.8 |
| Red Sea | Continental rift | Arabia / Africa | 13.8 | +13.7 | -1.4 | 16.7 |
| Hellenic Arc (Crete) | Subduction | Aegean Sea / Africa | 38.8* | -37.0 | 11.6 | 37.4 |
Velocity of the right-hand plate relative to the left-hand one, at the boundary step nearest to the place. Without asterisk: both plates are measured by GNSS; *: at least one plate comes from the PB2002 model. Negative slip: left-lateral. The last column gives the speed of the geological model at the same place.
How rock breaks: stress regimes of earthquakes
Global CMT earthquake focal mechanisms since 1976 (centroid moment tensors) · Free use with citation of Dziewonski et al. (1981) and Ekström et al. (2012) · 10/28/26The focal mechanism of an earthquake gives the axes of shortening (P) and lengthening (T) of the rock. Their plunges class the earthquake (Zoback, 1992, World Stress Map rules): normal fault (extension, P vertical), strike-slip (horizontal sliding), thrust fault (compression, T vertical), and the intermediate cases. Table: 59,047 shallow earthquakes (≤ 70 km) out of 71,453, placed by the nearest boundary step within 100 km, else “orogen” (zone of diffuse deformation in PB2002) or “intraplate”. Last column: share of the seismic moment (the energy released) of shallow earthquakes.
| Where | Earthquakes | Regimes | Seismic moment |
|---|---|---|---|
| Subduction | 20,885 | 67.9 % | |
| Oceanic convergence | 1,405 | 2.1 % | |
| Continental collision | 2,276 | 2.3 % | |
| Oceanic ridge | 7,375 | 0.7 % | |
| Continental rift | 2,726 | 4.4 % | |
| Oceanic transform fault | 10,210 | 3.6 % | |
| Continental transform fault | 2,099 | 2.6 % | |
| Orogen (diffuse deformation) | 5,181 | 10.1 % | |
| Intraplate | 6,890 | 6.4 % |
Each kind of boundary has its signature: compression at subduction zones, extension at ridges, sliding on transform faults. The normal faults of subduction zones are mostly in the plate that bends before sinking. At depth, 96 % of the 12,406 earthquakes deeper than 70 km lie within 1,000 km of a subduction zone: they follow the sinking slab.
Subduction sections: earthquakes outline the sinking slab
Depth of Global CMT earthquakes against distance from the trench, in a 200 km wide band perpendicular to the trench, toward the overriding plate. Under each trench the foci line up on a dipping plane, the Wadati–Benioff zone: the cold oceanic plate sinking into the mantle. Its dip is fitted between 70 and 300 km depth (white line). Colour: stress regime; size: magnitude.
Dips vary from one subduction zone to another: under Tonga, where an old and dense Pacific plate sinks fast, earthquakes reach almost 700 km, and the Mariana plane is one of the steepest; under central Chile (near 31° S) the Nazca plate stays almost flat for several hundred kilometres before sinking again, a well-known “flat slab” (Gutscher et al., 2000). Global CMT depths are centroid depths, less precise than relocated catalogues: the reference geometry of slabs is the Slab2 model (Hayes et al., 2018).
Plates and boundaries. PB2002 model (Bird, 2003): outlines of the 52 plates, 229 boundary curves, 5,819 steps of about 45 km classed as subduction zone, oceanic spreading ridge, oceanic transform fault, oceanic convergent boundary, continental rift, continental transform fault and continental convergent boundary, and 13 orogens (zones of diffuse deformation such as the Alps or Tibet).
Plate motion. MIDAS velocities of the Nevada Geodetic Laboratory (Blewitt et al., 2016, 2018), IGS20 frame (ITRF2020). Each station is placed in its plate (winding number on the sphere). Kept for the fit: at least 3 years of data, horizontal uncertainty ≤ 1 mm/yr, more than 200 km from any boundary (a locked fault strains the plate over hundreds of kilometres), outside the orogens, vertical rate ≤ 3 mm/yr (glacial rebound, local subsidence), and the best station of each 1° cell so that dense networks (United States, Japan, Europe) do not weigh like whole plates. The rotation vector ω of each plate minimises the weighted misfit between the measured velocities and v = ω × r (MIDAS uncertainties plus 0.3 mm/yr), stations beyond 3 robust standard deviations (and 1.5 mm/yr) being set aside in turn; its covariance is scaled by the reduced χ². A plate counts as rigid with at least 8 stations and a residual ≤ 1.5 mm/yr.
Comparison with the geological model. PB2002 poles are relative to the Pacific plate; they are compared with the GNSS poles minus the GNSS pole of the Pacific. For plates without enough stations (Nazca, Cocos, Philippine Sea, small plates), the PB2002 rotation is attached to the measured Pacific: a “hybrid” set of 52 plates used for the boundaries, maps and globe.
Boundaries. At the midpoint of each step, velocity of the right-hand plate relative to the left-hand one, split into opening (perpendicular to the step, positive when the plates move apart) and right-lateral slip; conventions checked by recomputing the PB2002 columns from its own poles (correlation 1.000). Area created: sum of opening × length over ridges and rifts; area consumed: sum of closing × length over convergent boundaries.
Stress. Global CMT mechanisms (Dziewonski et al., 1981; Ekström et al., 2012), catalogue 1976–2020, reviewed months since 2021 and quick CMTs for recent months. Regime from the plunges of the P, B and T axes (Zoback, 1992); direction of the maximum horizontal stress SHmax according to the regime; axial mean of directions per 2° cell (at least 3 earthquakes, consistency ≥ 0.5), weighted by the cube root of the moment. Moment magnitude Mw = 2/3 (log₁₀ M₀ − 16.1), M₀ in dyne·cm.
Sections. Earthquakes within 100 km of a great circle perpendicular to the subduction step nearest to the chosen place, from −150 to 900 km from the trench toward the overriding plate (PB2002 polarity); dip fitted by least squares between 70 and 300 km depth. Recomputed every week (Python pipeline, recipe tectonics).
PB2002 dates from 2003 and simplifies: some boundaries are wide zones, not lines, and the kind of a step depends on its direction relative to the relative motion. Absolute velocities depend on the frame (here IGS20): only relative velocities between plates have a physical meaning independent of the frame. Oceanic plates (Nazca, Cocos, Philippine Sea) have too few stations on stable islands: their motion comes from the geological model, marked with an asterisk.
The seismic moment is dominated by a few giant earthquakes (Sumatra 2004, Chile 2010, Japan 2011): its distribution changes a lot from one decade to the next, and the catalogue starts in 1976 (without Chile 1960 or Alaska 1964). A distance of 100 km to the nearest step does not tell on which fault an earthquake occurred. Shallow earthquake depths are often fixed (10 to 15 km) in Global CMT, hence horizontal bands on the sections.
Arrows and the globe animation exaggerate the motion: at 5 cm/yr a plate travels 500 km in 10 million years.
- Bird (2003)
-
Bird, P. (2003). An updated digital model of plate boundaries. Geochemistry, Geophysics, Geosystems, 4(3), 1027.
· doi:10.1029/2001GC000252
Licence : Published literature; cited - Blewitt et al. (2016)
-
Blewitt, G., Kreemer, C., Hammond, W. C., & Gazeaux, J. (2016). MIDAS robust trend estimator for accurate GPS station velocities without step detection. Journal of Geophysical Research: Solid Earth, 121(3), 2054–2068.
· doi:10.1002/2015JB012552
Licence : Published literature; cited - Blewitt et al. (2018)
-
Blewitt, G., Hammond, W. C., & Kreemer, C. (2018). Harnessing the GPS data explosion for interdisciplinary science. Eos, 99.
· doi:10.1029/2018EO104623
Licence : Published literature; cited - Dziewonski et al. (1981)
-
Dziewonski, A. M., Chou, T.-A., & Woodhouse, J. H. (1981). Determination of earthquake source parameters from waveform data for studies of global and regional seismicity. Journal of Geophysical Research, 86(B4), 2825–2852.
· doi:10.1029/JB086iB04p02825
Licence : Published literature; cited - Ekström et al. (2012)
-
Ekström, G., Nettles, M., & Dziewoński, A. M. (2012). The global CMT project 2004–2010: Centroid-moment tensors for 13,017 earthquakes. Physics of the Earth and Planetary Interiors, 200–201, 1–9.
· doi:10.1016/j.pepi.2012.04.002
Licence : Published literature; cited - Zoback (1992)
-
Zoback, M. L. (1992). First- and second-order patterns of stress in the lithosphere: The World Stress Map Project. Journal of Geophysical Research, 97(B8), 11703–11728.
· doi:10.1029/92JB00132
Licence : Published literature; cited - DeMets et al. (2010)
-
DeMets, C., Gordon, R. G., & Argus, D. F. (2010). Geologically current plate motions. Geophysical Journal International, 181(1), 1–80.
· doi:10.1111/j.1365-246X.2009.04491.x
Licence : Published literature; cited - Altamimi et al. (2023)
-
Altamimi, Z., Métivier, L., Rebischung, P., Collilieux, X., Chanard, K., & Barnéoud, J. (2023). ITRF2020 plate motion model. Geophysical Research Letters, 50(24), e2023GL106373.
· doi:10.1029/2023GL106373
Licence : Published literature; cited - Hayes et al. (2018)
-
Hayes, G. P., Moore, G. L., Portner, D. E., Hearne, M., Flamme, H., Furtney, M., & Smoczyk, G. M. (2018). Slab2, a comprehensive subduction zone geometry model. Science, 362(6410), 58–61.
· doi:10.1126/science.aat4723
Licence : Published literature; cited - Gutscher et al. (2000)
-
Gutscher, M.-A., Spakman, W., Bijwaard, H., & Engdahl, E. R. (2000). Geodynamics of flat subduction: Seismicity and tomographic constraints from the Andean margin. Tectonics, 19(5), 814–833.
· doi:10.1029/1999TC001152
Licence : Published literature; cited - Plate boundary model PB2002 (Bird, 2003): 52 plates, boundaries, Euler poles and boundary steps
-
Bird, P. (2003). An updated digital model of plate boundaries. Geochemistry, Geophysics, Geosystems, 4(3), 1027.
· doi:10.1029/2001GC000252 · source page ↗
Licence : Free use with citation (data supplement of the article); HTTPS mirror under ODC-BY 1.0 · retrieved on Oct 28, 2026 · SHA-256 e21f98c341f0… - GNSS station velocities, MIDAS, IGS20 frame (Nevada Geodetic Laboratory)
-
Blewitt, G., Hammond, W. C. and Kreemer, C. (2018). Harnessing the GPS data explosion for interdisciplinary science. Eos, 99. Velocities: Blewitt, G., Kreemer, C., Hammond, W. C. and Gazeaux, J. (2016). MIDAS robust trend estimator for accurate GPS station velocities without step detection. Journal of Geophysical Research: Solid Earth, 121, 2054–2068.
· doi:10.1029/2018EO104623 · source page ↗
Licence : Free use with citation of Blewitt et al. (2016, 2018) · retrieved on Oct 28, 2026 · SHA-256 80329fabd681… - Global CMT earthquake focal mechanisms since 1976 (centroid moment tensors)
-
Dziewonski, A. M., Chou, T.-A. and Woodhouse, J. H. (1981). Determination of earthquake source parameters from waveform data for studies of global and regional seismicity. Journal of Geophysical Research, 86, 2825–2852. Ekström, G., Nettles, M. and Dziewonski, A. M. (2012). The global CMT project 2004–2010: centroid-moment tensors for 13,017 earthquakes. Physics of the Earth and Planetary Interiors, 200–201, 1–9.
· doi:10.1016/j.pepi.2012.04.002 · source page ↗
Licence : Free use with citation of Dziewonski et al. (1981) and Ekström et al. (2012) · retrieved on Oct 28, 2026 · SHA-256 d9f31e121b13…