University of South Florida
Scholar Commons
#(..+.&%.1#)%-#%1!#3+27!-$2!9
3"+)#!2).-1
#(..+.&%.1#)%-#%1
REVEL: A Model for Recent Plate Velocities from
Space Geodesy
Giovanni F. Sella
Louisiana State University
Timothy H. Dixon
University of Miami2($31&%$3
Ailin Mao
University of Miami
.++.52()1!-$!$$)2).-!+5.0*1!2 (;/11#(.+!0#.,,.-131&%$3'%. &!#/3"
!02.&2(% !02(#)%-#%1.,,.-1
:)102)#+%)1"0.3'(22.7.3&.0&0%%!-$./%-!##%11"72(%#(..+.&%.1#)%-#%1!2#(.+!0.,,.-12(!1"%%-!##%/2%$&.0)-#+31).-)-#(..+
.&%.1#)%-#%1!#3+27!-$2!93"+)#!2).-1"7!-!32(.0)8%$!$,)-)120!2.0.&#(.+!0.,,.-1.0,.0%)-&.0,!2).-/+%!1%#.-2!#2
1#(.+!0#.,,.-131&%$3
#(.+!0.,,.-1)2!2).-
%++!).4!--))6.-),.2(7!-$!.)+)-.$%+&.0%#%-2+!2%%+.#)2)%1&0.,/!#%%.$%17
School of Geosciences Faculty and Sta Publications
(;/11#(.+!0#.,,.-131&%$3'%. &!#/3"
REVEL: A model for Recent plate velocities from space geodesy
Giovanni F. Sella
1
Department of Geology and Geophysics, Louisiana State University, Baton Rouge, Louisiana
Timothy H. Dixon and Ailin Mao
2
Rosenstiel School of Marine and Atmospheric Science, University of Miami, Miami, Florida
Received 30 October 2000; revised 21 January 2002; accepted 31 January 2002; published 27 April 2002.
[1] We present a new global model for Recent plate velocities, REVEL, describing the relative
velocities of 19 plates and continental blocks. The model is derived from publicly available space
geodetic (pri marily GPS) data for the period 1993–2000. We include an independent and rigorous
estimate for GPS velocity uncertainties to assess plate rigidity and propagate these uncertainties
to the velocity estimates. The velocity fields for North America, Eurasia, and Antarctica clearly show
the effects of glacial isostatic adjustment, and Australia appears to depart from rigid plat e behavior
in a manner consistent with the mapped intraplate stress field. Two thirds of tested plate pairs
agree with the NUVEL-1A geologic (3 Myr average) velocities within uncertainties. Three plate
pairs (Caribbean-North America, Caribbean-South America, and North America-Pacific) exhibit
significant differences between the geodetic and geologic model that may reflect systematic errors in
NUVEL-1A due to the use of seafloor magnetic rate data that do not reflect the full plate rate because
of tectonic complexities. Most other differences probably reflect real velocity changes over the last
few million years. Several plate pairs (Arabia-Eurasia, Arabia-Nubia, Eurasia-India) move more
slowly than the 3 Myr NUVEL-1A average, perhaps reflecting long-term deceleration associated
with continental collision. Several other plate pairs, including Nazca-Pacific, Nazca-South America
and Nubia-South America, are experiencing slowing that began 25 Ma, the beginning of the
current phase of Andean crustal shortening. I
NDEX TERMS: 1243 Geodesy and Gravity: Space
geodetic surveys; 8107 Tectonophysics: Continental neotectonics; 8150 Evolution of the Earth:
Plate boundary—general (3040); 8158 Evolution of the Earth: Plate motions—present and recent
(3040); K
EYWORDS: Plate tectonics, geodesy, GPS, global plate model, present-day, REVEL
1. Introduction
[2] The present-day velocities of the Earth’s lithospheric plates
are an important kinematic boundary condition for many geologic
and geophysical studies, including regional neotectonics, seismo-
genic zone processes, and earthquake hazards. Currently, the most
comprehensive picture of geologically young plate motion comes
from the global geologic model NUVEL-1A [DeMets et al.,
1990, 1994], a significant update of earlier global models [Chase,
1972, 1978; Minster et al., 1974; Minster and Jordan, 1978].
NUVEL-1A is based in large part on mid-ocean ridge spreading
rates dated from magnetic anomaly 2A (3 Ma, or mid-Pliocene
time) and thus describes relative plate velocities averaged over
Pliocene to Recent time. NUVEL-1A is a robust model based on
a large data set, but it nevertheless has some deficiencies. First,
some smaller plates are necessarily omitted because of lack of
data. Second, the geologic model may be biased due to poor or
insufficient kinematic data, e.g., North America-Pacific relative
motion [DeMets, 1995; DeMets and Dixon, 1999] and motion of
the Caribbean plate relative to North and South America [Dixon
et al., 1998; DeMets et al., 2000; Weber et al., 2001; Perez et al.,
2001]. Third, the 3-Myr average velocity predicted by NUVEL-
1A or any other geologic model may yield biased estimates of
present-day velocity for some plate pairs because the plates are
speeding up, slowing down, or changing direction, e.g., Nazca-
South America [Norabuena et al., 1998, 1999; Angermann et al.,
1999].
[
3] Space geodesy has the potential to measure relative plate
velocities directly over periods of just a few years, as demon-
strated by satellite laser ranging (SLR) [Smith et al., 1990;
Robbins et al., 1993; Cazenave et al., 1993; Sengoku, 1998],
very long baseline interferometry (VLBI) [Argus and Gordon,
1990; Robaudo and Harrison, 1993; Ryan et al., 1993; Sato,
1993]; Doppler Orbitography and Radiopositioning Integrated by
Satellite (DORIS) [ Cazenave et al., 1992; Soudarin and Caze-
nave, 1993, 1995; Cretaux et al., 1998] and the Global Position-
ing System (GPS) [Dixon et al., 1991a; Dixon, 1993; Argus and
Heflin, 1995; Larson et al., 1997; Dixon and Mao, 1997]. Most
studies to date suggest that the great majority of plate velocities
estimated from space geodesy are consistent with the NUVEL-1A
model within 95% confidence. However, uncertainties in the
geodetic estimates have been large enough that important differ-
ences may have been missed. The true uncertainty of space
geodetic data has also been difficult to quantify. Therefore it
has been difficult to address an important tectonophysical prob-
lem, namely, the extent to which individual plate velocities may
be changing over the last few million years and to what extent
such changes, if they occur, can be understood in terms of simple
plate-driving forces.
[
4] New space geodetic data and new analytical techniques
now permit a significant refinement of our description of present-
day plate motion. In this paper we present a comprehensive
velocity model for most major and several minor plates and
JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 107, NO. B4, 2081, 10.1029/2000JB000033, 2002
1
Now at Department of Geological Sciences, Northwestern University,
Evanston, Illinois.
2
Also at Magellan Systems, San Dimas, California.
Copyright 2002 by the American Geophysical Union.
0148-0227/02/2000JB000033$09.00
ETG 11 - 1
continental blocks, based primarily on GPS. Our study differs
from previous global studies in several respects:
1. A very large geodetic data set is now publicly available
through the efforts of many individuals, institutions, and geodetic
agencies, permitting a more accurate and more comprehensive
geodetic plate motion model. This data set represents both a large
number of sites, giving generally good geographic distribution, and
long, nearly continuous time series at many individual sites,
resulting in precise site velocity estimates. Most major plates now
have at least two GPS sites, the minimum number to determine a
plate’s angular velocity with space geodesy.
2. We include velocity estimates for the Amuria, Anatolia,
Caribbean, Nubia, Okhotsk, Philippine, Sierra Nevada, Somalia,
South China, and Sunda plates or continental blocks. While these
and other plates and continental blocks have been the subject of
earlier local studies, most previous global geologic and geodetic
plate motion models have omitted one or more of these plates
because of sparse data or have approximated Nubia (west Africa)
by combining it with Somalia (east Africa).
3. We incorporate a rigorous, independent estimate for GPS
velocity errors. This permits simple, objective tests of whether the
GPS site velocity for a given location is consistent with rigid plate
behavior and whether plate motions averaged over the last few
years differ significantly from motions averaged over the last few
million years.
[
5] The velocity predictions of geologic plate motion models
are sometimes termed ‘‘present-day’’ or ‘‘current’’ because they
describe geologically young plate motion, derived from the young-
est easily identified magnetic anomaly, typically 2A (3 Ma, or
mid-Pliocene) [Minster and Jordan, 1978; DeMets et al., 1990,
1994] to, in some cases, anomaly 1N (0.8 Ma, or mid-Pleisto-
cene) [DeMets, 1995; Conder and Forsyth, 2000]. Our geodetic
plate motion model is derived from data over a very different time
span, roughly the last decade. It is probably representative of plate
motions over the Holocene or Recent epoch (last 10,000 years)
and possibly the late Pleistocene epoch (last few hundred thousand
years) provided that we account for, or avoid, short-term strain
effects related to the seismic cycle and isostatic effects associated
with the last glacial cycle. The former can impact site velocities in
the vicinity of active plate boundary zone faults, while the latter
may impact site velocities in parts of North America, Eurasia,
Greenland, and Antarctica. To emphasize the time span over which
we believe our model to be valid, we have termed it ‘‘REVEL’’ (for
Recent velocities) with the suffix 2000, to indicate the last year of
data included in the model. We expect that the model can be
improved significantly in subsequent years as additional data are
added and time series lengthen.
2. Data Analysis
[6] Uncertainty in the positions of the GPS satellites is a major
error source for the coordinate time series used in this study. By
1993 the global tracking network for GPS satellites became
sufficiently robust to produce more accurate satellite ephemerides
(satellite positions as a function of time) compared to earlier
periods. Our data span the time period 1 January 1993 to 31
December 2000. To ensure consistency, all data were reanalyzed
specifically for this study, resulting in a uniform set of site velocity
and error estimates. The great majority of sites used in this study
are continuous, in the sense that they record data for 20 – 24 hours
per day, typically for at least 300 days per year (Table 1). For more
than 98% of the continuous sites presented here, all known existing
data acquired after 1 January 1993 were processed (for the
remaining sites, there may be additional data that were not
available to us). In some cases (the Caribbean and Philippine
plates and the Sierra Nevada block), continuous site distribution is
limited, and we have augmented these data with data from
‘‘episodic’’ sites occupied periodically, usually every year or two.
A total of 345,000 station days of data were analyzed for this
study, most of which lie in stable plate interiors and are used to
estimate plate velocities (Table 1). The data were analyzed at the
University of Miami, generally following procedures outlined by
Dixon et al. [1993, 1997], although some of these procedures have
been updated considerably. Salient points are listed here:
1. We use GIPSY/OASIS II, Release 5.0 software developed at
the Jet Propulsion Laboratory (JPL) and nonfiducial satellite orbit
and clock files provided by JPL [Zumberge et al., 1997].
2. We use the ionosphere-free combination of both undiffer-
enced carrier phase and P code pseudorange data (data weights of 1
cm and 1 m, respectively), typically recorded at 30-s intervals, with
a 5-min decimation rate.
3. We created a comprehensive antenna height and type change
file going back to 1993 for all sites, and used the relative dome-free
models of Mader [1999] for first-order antenna phase center
corrections.
4. To improve the definition of horizontal atmospheric
gradients and to reduce their effect on the final position estimates,
we use an elevation angle cutoff of 10 (where data are available)
and the horizontal gradient model of Bar-Sever et al. [1998]. We
use the same random walk model for the zenith delay as Bar-Sever
et al. (3 mm/
p
h), but a looser constraint for the two orthogonal
horizontal gradients (5 mm/
p
h). We use the mapping function of
Niell [1996], which describes how the average atmospheric path
delay varies as a function of elevation angle.
5. All sites are corrected for ocean tidal loading, using the
NLOADF program of Agnew [1997], and the Schwiderski [1980]
ocean tide model.
6. Carrier phase cycle ambiguities are estimated, not fixed.
7. We estimated offset parameters at the date of each change of
antenna height or model to correct for second-order effects such as
inaccuracy of phase center estimates and incorrect records of
antenna height changes; these can exceed 1 cm in the vertical
component, but generally are a few millimeters or less for horizontal
components. Offsets were also calculated for antenna dome changes.
8. Daily position estimates are generated with loose constraints
[Heflin et al., 1992; Blewitt et al., 1992], then transformed to
International Terrestrial Reference Frame (ITRF)-97 [Boucher
et al., 1999] using up to 51 colocated sites whose positions are
defined in ITRF-97. The number of colocated sites depends on
station availability. For 1998 and later, this usually exceeds 45, but
in 1993, for example, it was typically 15 – 20. Transformation
parameters are computed each day; hence each daily position
estimate is essentially independent.
9. Position estimates with formal errors >9.9 m are removed
from the database (typically <2 days per site per year).
[
7] Velocity estimates are based on a weighted least squares
line fit to the daily posit ion est imates, including the offset
parameters described above (Figure 1). Outliers, defined as points
that lie off the best fit line by more than 3 times the formal error,
are flagged, but not removed from the database, and are not used
in the line fit (the fitting and outlier definition are done iter-
atively). Sites listed in upper case letters in Table 1 and in figures
(e.g., STJO) lie on the stable plate and are used to define t he rigid
plate angular velocity. Sites in lowercase (e.g., cic*, algo) are not
used in the rigid plate definition because they may lie in the
deforming boundary zone or may be affected by glacial isostatic
adjustment or other nonrigid process. In some cases a station was
moved or offset during the 1993–2000 period of this study.
Where possible, we used publicly available vector tie information
(‘‘site ties’’) to link the two time series. If vector tie information
was unavailable, we used a procedure similar to that described in
point 7 above to tie the two time series together, in effect
estimating the vector tie along with the slope (velocity) parameter
of interest (Figure 1 shows an example). In such cases we are
estimating three parameters (slope, intercept, and offset) rather
ETG 11 - 2 SELLA ET AL.: REVEL—RECENT PLATE VELOCITIES FROM SPACE GEODESY
than the normal two parameters for a straight line. This allows us
to take advantage of the reduction of velocity uncertainty asso-
ciated with longer time series. Site names for such linked time
series are indicated by asterisks in Table 1 and in the figures; e.g.,
cic* represents the combination of time series from sites Cice and
Cic1 (Figure 1). Velocities for the original sites are also listed in
Table 1 immediately preceding the combined solution with only
the first letter capitalized. In a few cases, tim e series have
obvious offsets, presumably associated with an equipment or
other significant change, even though no change is reported. In
cases where such offsets are large enough to affect the velocity
significantly, we have estimated the offset following the proce-
dure outlined above. These sites are noted in Table 1.
[
8] Our velocity error estimates account for both white (uncor-
related) and colored (time-correlated) noise [Langbein and John-
son, 1997] and are modeled following Mao et al. [1999]. This
approach is based on a numerical analysis of 23 globally
distributed GPS sites with time series spanning the period
1994– 1997. The error model avoids any assumptions concerning
the fit of GPS data to a particular geophysical model (e.g., rigid
plate model) and thus provides independent error estimates. To
account for improvements in analytical techniques since 1997 and
for site-specific effects, we exploit the correlation between
WRMS (the weighted root mean square scatter of the daily
position estimates about a best fit straight line) and white and
flicker noise amplitudes observed in the data of Mao et al.
[1999], as outlined by Dixon et al. [2000a]. Random walk noise
is assumed to be zero. When this error model is applied to our
data set of 64 site velocities from the stable interior of North
America, we obtain a c
2
per degree of freedom of 1.05 for the
rigid plate model, close to the expected value of 1.0. This
suggests that the error model is reasonable and that the region
sampled by these data can be assumed rigid within data uncer-
tainty. This is discussed in more detail in section 4.1.
Figure 1. Example coordinate time series for site cic* (Table 1), a composite time series composed of Cice prior to
1 March 1999 and Cic1 thereafter. Note offset estimate at date of site change (vertical dashed line) after site tie
correction. Site velocity is given by slope of weighted least squares line fit (thin solid line) through the data (solid
circles), excluding outliers (open triangles). Site velocity error is the uncertainty of the slope estimate, accounting for
uncorrelated and time-correlated noise, total time span of observations, and total number of observations.
SELLA ET AL.: REVEL—RECENT PLATE VELOCITIES FROM SPACE GEODESY
ETG 11 - 3
Table 1. Site Positions and Velocities Relative to ITRF-97
Site
a
Position
b
T,
c
years
Total
Data
d
Velocity
e
WRMS
f
Rate Res.
g
N E North East Vertical N E V N E
Amuria (Am)
Taej
h
36.37 127.37 3.32 1185 17.8 ± 0.9 27.9 ± 1.2 3.8 ± 2.7 3.7 5.9 11.3 – –
Daej
h
36.40 127.37 1.79 619 12.4 ± 1.6 25.6 ± 2.1 1.0 ± 4.8 3.5 5.3 10.2 – –
DAE*
i
36.40 127.37 5.12 1804 17.0 ± 0.6 27.6 ± 0.8 3.0 ± 1.8 4.0 5.8 11.1 1.1 0.2
SUWN 37.28 127.05 3.07 920 14.2 ± 0.9 27.8 ± 1.2 1.9 ± 2.7 3.2 5.2 9.2 1.7 0.2
VLAD 43.20 131.93 4.47 1363 15.7 ± 0.8 24.8 ± 0.8 1.5 ± 2.0 4.4 5.4 11.0 0.3 0.1
bjfs 39.61 115.89 1.20 419 13.8 ± 2.4 29.9 ± 2.8 3.9 ± 7.5 3.4 4.9 11.0 1.3 1.1
irkt 52.22 104.32 5.27 1867 9.1 ± 0.7 26.5 ± 0.7 1.8 ± 1.7 4.3 5.4 10.6 4.6 0.3
xian 34.37 109.22 3.48 991 14.7 ± 0.9 34.5 ± 1.1 0.1 ± 2.5 3.8 5.2 10.0 0.4 3.4
Anatolia (At)
Ankr
h
39.89 32.76 5.52 1580 10.3 ± 0.7 0.9 ± 1.0 0.9 ± 1.5 4.9 7.5 9.4 0.0 0.0
ANK*
j
39.89 32.76 4.14 1280 8.1 ± 0.7 4.8 ± 0.7 3.7 ± 1.9 3.4 4.4 8.6 0.2 0.2
7580
k
37.38 33.19 10 – 12.0 ± 4.0 5.2 ± 3.9 5.2 ± 4.0 – – – 3.1 3.7
7585
k
39.80 34.81 10 – 17.7 ± 4.3 3.2 ± 4.3 3.3 ± 4.3 – – – 6.8 8.2
7589
k
39.89 32.76 10 – 8.0 ± 2.0 5.3 ± 1.7 21.1 ± 2.0 – – – 0.3 0.8
Antarctica (An)
CAS1 66.28 110.52 6.49 1839 11.6 ± 0.7 3.7 ± 0.6 6.4 ± 1.5 5.2 5.5 12.4 0.2 0.2
DAV1 68.58 77.97 6.49 1849 6.6 ± 0.7 2.1 ± 0.7 4.0 ± 1.5 5.2 6.3 12.2 0.3 0.4
KERG 49.35 70.26 6.13 1971 5.3 ± 0.7 5.2 ± 0.8 6.5 ± 1.5 5.1 7.1 11.7 0.1 0.3
MAW1 67.60 62.87 7.00 1830 3.6 ± 0.7 3.5 ± 0.7 3.9 ± 1.3 5.8 7.2 10.2 0.2 0.1
MCM4 77.84 166.67 5.93 2060 12.2 ± 0.6 10.6 ± 0.7 4.6 ± 1.9 4.5 5.7 15.5 0.9 0.5
SYOG 69.01 39.58 4.00 1164 0.3 ± 1.0 4.4 ± 1.2 5.8 ± 2.2 5.0 6.8 10.4 1.2 0.1
VESL 71.67 357.16 2.39 648 8.2 ± 1.8 4.7 ± 1.4 1.1 ± 3.6 5.0 4.8 9.8 1.7 3.3
ohig 63.32 302.10 5.70 1330 10.1 ± 0.9 14.7 ± 0.8 8.7 ± 1.8 6.1 6.5 12.9 2.6 2.3
palm 64.78 295.95 2.46 851 14.1 ± 1.9 11.7 ± 1.5 3.3 ± 3.4 5.7 5.3 9.5 1.8 1.5
Arabia (Ar)
BAHR 26.21 50.61 4.52 1592 26.9 ± 0.6 29.5 ± 0.7 0.9 ± 1.8 3.0 4.9 9.1 0.2 0.6
KATZ 33.00 35.69 4.18 590 19.3 ± 0.7 22.7 ± 1.0 3.9 ± 2.4 3.2 5.3 11.4 0.2 1.1
7832
k
24.91 46.40 5 – 20.3 ± 8.2 25.9 ± 8.5 4.3 ± 8.4 – – – 4.6 4.3
Australia (Au)
ALIC 23.67 133.89 6.45 1574 57.0 ± 0.4 33.9 ± 0.7 2.0 ± 1.5 3.1 6.0 11.8 0.1 0.4
CEDU 31.87 133.81 6.63 915 57.4 ± 0.5 31.5 ± 0.6 3.5 ± 1.4 3.4 5.2 10.0 0.5 0.5
DARW 12.84 131.13 6.61 972 57.4 ± 0.5 38.4 ± 1.0 2.2 ± 1.6 3.4 8.2 12.2 0.2 0.2
HOB2 42.80 147.44 6.49 1873 54.2 ± 0.5 15.5 ± 0.6 2.5 ± 1.4 3.7 5.6 10.7 0.2 1.1
JAB1 12.66 132.89 3.40 501 58.2 ± 0.8 35.3 ± 1.4 1.9 ± 2.8 3.1 6.1 10.5 1.2 2.6
KARR 20.98 117.10 6.44 1576 56.7 ± 0.4 43.3 ± 0.7 1.0 ± 1.4 3.2 6.5 10.9 0.6 2.1
PERT 31.80 115.89 7.36 2387 55.5 ± 0.4 42.1 ± 0.6 0.4 ± 1.3 3.3 6.3 11.8 0.4 1.8
TID2 35.40 148.98 5.00 1386 53.4 ± 0.5 19.5 ± 0.7 3.6 ± 1.7 3.2 4.9 10.0 0.1 1.4
TIDB 35.40 148.98 8.00 2468 53.3 ± 0.5 19.4 ± 0.6 5.6 ± 1.2 4.3 6.7 12.0 0.2 1.5
TOW2 19.27 147.06 5.96 1500 53.5 ± 0.5 30.5 ± 0.8 3.5 ± 1.6 3.5 6.8 11.1 0.7 1.1
YAR1 29.05 115.35 8.00 2718 54.9 ± 0.4 41.9 ± 0.6 3.8 ± 1.2 3.5 6.3 11.7 0.9 0.8
auck 36.60 174.83 5.29 1885 37.5 ± 0.5 5.1 ± 0.7 3.9 ± 1.5 3.4 5.7 8.9 1.8 2.3
coco
l
12.19 96.83 4.55 1363 47.7 ± 0.7 43.0 ± 1.3 0.3 ± 2.2 3.9 8.5 12.4 1.1 2.2
dgar 7.27 72.37 4.60 1556 29.8 ± 0.7 46.0 ± 1.4 2.5 ± 2.2 4.1 8.8 13.5 2.3 0.5
hyde 17.42 78.55 5.47 268 32.2 ± 0.8 42.1 ± 1.5 2.1 ± 2.5 4.4 9.5 15.2 4.8 17.0
iisc 13.02 77.57 5.98 1611 31.8 ± 0.6 42.4 ± 1.0 0.1 ± 1.7 3.9 8.2 12.8 4.5 13.5
mald
l
4.19 73.53 1.39 485 30.1 ± 2.8 41.5 ± 4.2 8.6 ± 6.5 4.7 8.2 11.0 3.0 5.1
noum 22.27 166.41 3.00 986 43.1 ± 0.9 21.8 ± 1.6 1.8 ± 2.9 3.1 6.7 10.2 1.9 1.8
Caribbean (Ca)
AVES 15.67 63.62 3.87 29 11.5 ± 0.9 13.2 ± 2.1 5.0 ± 3.4 2.5 7.0 9.4 1.1 0.6
BARB 13.09 300.39 3.09 569 12.6 ± 1.0 13.1 ± 1.8 0.7 ± 3.3 3.3 7.3 11.8 0.2 0.3
CRO1 17.76 295.42 6.96 1828 10.7 ± 0.4 10.5 ± 0.7 2.5 ± 1.5 3.5 6.8 13.1 0.1 0.6
ROJO 17.90 71.67 6.25 21 6.2 ± 0.9 11.5 ± 1.9 5.5 ± 3.7 3.8 9.0 18.4 0.8 1.7
SANA 12.52 81.73 6.25 29 4.8 ± 0.9 14.1 ± 1.2 0.8 ± 3.8 4.1 6.2 20.8 1.1 1.0
TDAD 10.68 61.40 4.05 7 10.7 ± 1.4 15.9 ± 3.2 5.9 ± 4.1 3.2 8.1 9.4 1.5 1.1
isab 18.47 67.05 4.18 26 9.3 ± 1.4 11.1 ± 2.0 0.8 ± 7.8 4.2 7.0 30.8 0.4 0.4
pur3 18.46 292.93 3.56 1096 9.5 ± 0.8 7.5 ± 1.4 0.2 ± 2.9 3.2 7.0 13.7 0.2 3.2
Eurasia (Eu)
ARTU 56.43 58.56 1.40 388 4.8 ± 1.6 24.5 ± 1.6 0.9 ± 5.6 2.6 3.3 8.2 0.2 0.8
BOGO 52.48 21.04 4.00 1293 12.7 ± 0.5 22.2 ± 0.5 0.9 ± 2.0 2.4 3.0 8.8 0.1 0.8
BOR1 52.28 17.07 6.26 2221 12.5 ± 0.3 20.4 ± 0.4 1.9 ± 1.2 2.6 3.8 8.1 0.6 0.4
GLSV 50.36 30.50 2.85 948 11.6 ± 0.9 22.4 ± 0.8 0.0 ± 2.7 2.9 3.5 8.1 0.6 1.0
GOPE 49.91 14.79 5.30 1826 13.0 ± 0.4 21.6 ± 0.5 4.9 ± 1.7 2.6 4.0 10.9 0.4 0.7
JOZE 52.10 21.03 7.36 2577 12.5 ± 0.3 21.3 ± 0.4 5.2 ± 1.2 3.0 4.1 10.8 0.1 0.2
KSTU 55.99 92.79 2.92 713 5.6 ± 1.1 25.3 ± 1.3 0.4 ± 2.9 3.6 5.2 9.5 1.7 0.4
LAMA 53.89 20.67 6.03 1844 12.8 ± 0.4 20.9 ± 0.5 2.6 ± 1.3 3.1 4.4 8.2 0.2 0.1
ETG 11 - 4 SELLA ET AL.: REVEL—RECENT PLATE VELOCITIES FROM SPACE GEODESY
![Figure 19. Nubia-South America spreading rates (full rate) versus age for last 25 Myr, from Cande and Kent [1992] (shaded area shows 1 standard error), compared to NUVEL-1A and REVEL-2000 predictions (arrows) calculated at 28 S, 348 E. REVEL-2000 differs significantly from NUVEL-1A but agrees closely with an extrapolation of interval spreading rates to Recent time.](/figures/figure-19-nubia-south-america-spreading-rates-full-rate-2vn4w1ot.webp)