Geophys. J. Int. (2007) 170, 1373–1380 doi: 10.1111/j.1365-246X.2007.03494.

New constraints on relative motion between the Pacific Plate and

Baja California microplate (Mexico) from GPS measurements

C. Plattner,1 R. Malservisi,1 T. H. Dixon,2 P. LaFemina,2, ∗ G. F. Sella,3 J. Fletcher4

and F. Suarez-Vidal4
1 SektionGeophysik der Ludwig-Maximilians Universität, Theresienstr. 41, 80333 München, Germany. E-mail: plattner@geophysik.uni-muenchen.de
2 Rosenstiel School of Marine and Atmospheric Sciences, University of Miami, 4600 Rickenbacker Causeway, Miami, FL 33149, USA
3 National Geodetic Survey, NOAA, 1315 East-West Highway, Silver Spring, MD 20910, USA
4 Centro de Investigacion Cientifica y Educacion Superior de Ensenada, Km. 107 Carretera Tijuana Ensenada, Ensenada BC 2732, Mexico

Accepted 2007 May 14. Received 2007 May 14; in original form 2007 March 13

SUMMARY
We present a new surface velocity field for Baja California using GPS data to test the rigidity of
this microplate, calculate its motion in a global reference frame, determine its relative motion
with respect to the North American and the Pacific plates, and compare those results to our
estimate for Pacific–North America motion. Determination of Pacific Plate motion is improved
by the inclusion of four sites from the South Pacific Sea Level and Climate Monitoring Project.
These analyses reveal that Baja California moves as a quasi-rigid block but at a slower rate
in the same direction, as the Pacific Plate relative to North America. This is consistent with
seismic activity along the western edge of Baja California (the Baja California shear zone),

GJI Tectonics and geodynamics

and may reflect resistance to motion of the eastern edge of the Pacific Plate caused by the ‘big
bend’ of the San Andreas fault and the Transverse Ranges in southern California.
Key words: geodesy, global positioning system (GPS), plate tectonics, tectonics.

Data analysis
I N T RO D U C T I O N
For the Pacific Plate we use only continuous GPS (CGPS) with
Plate rigidity is a key assumption in plate tectonics. While this as-
time-series longer than 3 yr, giving time-series durations from 3
sumption works well for plate interiors, plate boundaries can include
to 10 yr (Table 1). Compared to previous studies of Pacific Plate
a broad region of deformation and the development of multiple
motion we add four new sites from the South Pacific Sea Level
blocks or ‘microplates’. This is particularly true for the Pacific–
and Climate Monitoring Project (SPSLCMP) improving kinematic
North American Plate boundary (e.g. Atwater & Stock 1988). Iden-
constraints in the central and western Pacific. To reduce the uncer-
tifying these rigid blocks provides important kinematic boundary
tainty in the time-series of GUAX on Guadalupe Island, we added
conditions for tectonic studies of western North America.
episodic GPS (EGPS) data from GAIR through a vector tie (Sella et
Constraints on North America–Pacific Plate motion are also im-
al. 2002), which extends the time-series back to 1993. The uncer-
portant for kinematic tectonic studies of western North America
tainty of this new time-series, here called GUAZ, is reduced by 40
and parts of the circum-Pacific region. Models of this motion on
per cent, while the velocity changes only by 0.1 mm yr−1 in all three
geological timescales (e.g. DeMets et al. 1990, 1994) may use
components.
magnetic anomalies from the spreading centre in the southern
Thirty-two of 33 GPS stations in Baja California are episodic
Gulf of California. However, evidence is accumulating that Baja
GPS (EGPS) sites. We use EGPS data from sites with at least three
California’s motion is distinct from that of the Pacific Plate (Fig. 1)
occupation episodes of two and more 24 hr days and a minimum
and thus behaves as a separate block or microplate (Dixon
total time span of 6 yr since 1993. We did not use data from before
et al. 2000b; Fletcher & Munguia 2000; Gonzalez-Garcia et al.
1993 due to incomplete orbit information.
2003; Michaud et al. 2004). Here we use new GPS data to
All data were processed using GIPSY/OASIS II, Release 5.0
quantify current North America–Pacific Plate motion and inves-
software and non-fiducial satellite orbit and clock files provided by
tigate coupling and rigidity of Baja California and the Pacific
the Jet Propulsion Laboratory (Zumberge et al. 1997). The data
Plate.
analysis follows Sella et al. (2002), but the daily solutions are
aligned to IGb00 (Ray et al. 2004). The velocity and its uncertainty
for each site are then calculated by linear regression. Outliers with an
∗ offset of more than three times the formal error are not considered in
Now at: Pennslyvania State University, University Park, PA 16802, USA.


C 2007 The Authors 1373
Journal compilation 
C 2007 RAS
1374 C. Plattner et al.

the regression. Velocity uncertainties are calculated following Mao

et al. (1999) and Dixon et al. (2000a).
We calculate the stable plate reference frames for Pacific and Baja
32˚ North American plate California by the best fitting Euler vector (Minster et al. 1974; Ward
1990), testing for plate rigidity by comparing velocity residuals to
uncertainties. Using the stable North America reference frame of
Sella et al. (2006) we calculate its relative plate motion with the
Pacific Plate and Baja California microplate.
30˚
ABf
SPMf Stable Pacific Plate reference frame
Using standard geological criteria for the definition of a stable plate
(e.g. Sella et al. 2002) we initially identified 21 CGPS sites on the
28˚
Pacific Plate interior. We excluded FALE due to its location in a
Subduction-Transform Edge Propagator (STEP) region (Govers &
Wortel 2005). FARB, KOKB, NAUR, PAMA, TAHI and UP01 are
Pacific plate known to have technical problems or large uncertainties and are,
26˚ TAf therefore, excluded. We use only one station from the Hawaiian
Islands to avoid an overconstraint; the station importance (Minster
et al. 1974) would sum up to 48 per cent. We choose KOK1 as it
ABf = Agua Blanca fault Mf is located furthest from volcanic activity. However, we include all
Bf = Bonfil fault Bf Hawaiian Island stations to compare residuals with respect to the
24˚ CAf = Carrizal fault stable Pacific Plate.
Mf = Magdalena fault CAf As the stable Pacific Plate reference frame is based upon a limited
SJf = San Jose Cabo fault number of GPS stations we test its sensitivity to each GPS station.
SPMf = San Pedro Martir fault SJf Using a jackknife method we compute 11 rotation poles from the
TAf = Tosco Abreojos fault
data set of 11 GPS stations (Table 1), leaving out one station at a
22˚ time. We compare the rotation pole locations and the corresponding
average residual motion within the Pacific Plate. For every model
Figure 1. Baja California, Mexico: Faults above latitude 28 from Instituto we apply the F-test (Stein & Gordon 1984) to test whether we obtain
Nacional Estadistica Geografia e Informatica de Mexico and Dixon et al. significant improvement. We recognize the limitations of these tests,
(2002), faults south of latitude 28 by Paul Umhoefer (personal communica- in the sense that our sample size is small, and the tests assume
tion, 2006) and Michaud et al. (2004). Epicentres from National Earthquake normal distribution. While the 2-D-error ellipses for all rotation
Center Information (1973 to present). The geologically defined rigid block poles overlap at 95 per cent confidence, however, at the level of one
is outlined by Agua Blanca fault (ABf), San Pedro Martir fault (SPMf), San standard error, the solutions are sensitive to exclusion of stations
Jose del Cabo fault (SJf), Bonfil fault (Bf) and Carrizal fault (CAf), Tosco GUAZ and CHAT (Fig. 2, Table 2). The F-test implies a significant
Abreojos fault (TAf) and Magdalena fault (Maf).
improvement in the definition of the Pacific Euler vector for the

Table 1. GPS data used for the computation of Pacific Euler vector and their residual rate.
IGb00 Residual to PA1c PA1
Site ida Lon. Lat. T Ve σ Veb Vn σ Vnb Rate σ b rate Azi. Importance
(◦ E) (◦ N) (yr) (mm yr−1 ) (mm yr−1 ) (mm yr−1 ) (mm yr−1 ) (mm yr−1 ) (mm yr−1 ) (◦ )d (per cent)
chat1 −176.57 −43.96 10 −41.2 0.3 32.3 0.3 1.0 0.5 78 –
CKIS2 −159.80 −21.20 4 −63.5 0.8 33.8 0.5 0.4 0.7 −46 13
guaz4 −118.29 28.88 12 −47.7 0.3 23.4 0.3 2.0 0.6 149 –
hnlc1 −157.86 21.30 5 −63.6 0.8 33.8 0.6 0.9 0.9 −71 –
KIRI2 172.92 1.35 3 −67.10 1.1 30.7 0.6 1.4 1.0 50 8
KOK11 −159.76 21.98 7 −63.0 0.6 33.8 0.4 0.3 0.5 −22 21
KWJ1 167.73 8.72 6 −69.7 0.7 27.7 0.4 0.6 0.5 176 10
MARC3 153.98 24.29 5 −73.3 1.4 21.5 1.2 2.3 1.4 −138 5
maui1 −156.26 20.71 7 −63.3 0.4 32.8 0.3 0.9 0.4 −142 –
mkea1 −155.46 19.80 9 −63.7 0.4 33.4 0.3 0.7 0.5 −96 –
POHN2 158.21 6.96 4 −70.2 1.3 26.5 0.8 1.6 0.9 −7 7
THTI1 −149.61 −17.58 7 −66.6 0.8 32.6 0.4 0.5 0.6 −159 20
TRUK3 151.89 7.45 4 −72.0 1.6 22.5 0.9 1.5 1.6 −83 6
TUVA2 179.20 −8.53 4 −64.2 0.9 31.2 0.5 1.1 1.0 97 10
a Only upper case sites are used to compute the Pacific Euler vector. 1 IGS, 2 SPSLCMP, 3 WING, 4 SCEC time-series has been tied using guax and gair.
b Uncertainties are 1σ .
c Velocity after removing rigid motion of the Pacific Plate (PA1) from the IGb00 velocities at each site. See Table 2 for angular velocity.
d Azimuth is the angle of the rate residual in degrees clockwise from North.


C 2007 The Authors, GJI, 170, 1373–1380
Journal compilation 
C 2007 RAS
 Pacific Plate and Baja California microplate kinematics 1375

11 models. In the model for which GUAZ was excluded we also

obtain the largest residual for GUAZ (2.6 ± 0.6 mm yr−1 ), while
CHAT shows a near perfect fit (0.4 ± 0.4 mm yr−1 ). Stations on the
Hawaiian Islands, TUVA, MARC, and TRUK show low residuals.
In the model for which CHAT is excluded, the residual of CHAT
increases to 1.9 ± 0.5 mm yr−1 , while GUAZ is fit better (0.6 ±
0.4 mm yr−1 ), together with good fits at THTI, CKIS, KIRI, and
POHN. In general, these results are consistent with a rigid Pacific
Plate, within limits defined by our data uncertainty, ∼2 mm yr−1 .
Our best-fitting Pacific Plate Euler vector PA1 is based on nine
stations, excludes GUAZ and CHAT, and has a reduced χ 2 of 1.00
(Tables 1 and 2). The station importance for this solution varies
between 5 and 21 per cent (Table 1). The average residual ve-
Figure 2. Stability of location of Pacific-IGb00 pole of rotation. Using 10
out of 11 stations we compute 11 rotation poles, leaving out one station at
locity of the 14 stations on the interior of the Pacific Plate is
a time. The rotation pole shows increased sensitivity to GUAZ and CHAT. 1.1 mm yr−1 . GUAZ shows significant residual motion, while CHAT
We exclude GUAZ and CHAT and use nine stations (Table 1) to obtain and MAUI have residual velocities close to the error limit (Fig. 3,
our best-fitting pole of rotation. All rotation pole error ellipses are colour Table 1).
coded by the average residual motion, calculated from of the 11 stations plus
HNLC, MKEA and MAUI. For visibility only the 1σ error ellipses showing
the 2-D-error are shown.
Rigidity of the Baja California microplate
Geological observations suggest rigid block behaviour for Baja
exclusion of each and both stations. This can be explained by the California (e.g. Gastil et al. 1975; Suarez-Vidal et al. 1991;
geographic location of GUAZ and CHAT and their resulting relative Umhoefer & Dorsey 1997; Umhoefer 2000). The northern and
importance that varies between 26–33 per cent for CHAT and 32– southern ends of the peninsula are cut by several faults, but there
38 per cent for GUAZ. The residual velocities of GUAZ and CHAT is no apparent deformation along the main body of the peninsu-
with respect to the Pacific Plate Euler vectors are close to the limit lar batholith. We calculate a Baja California Euler vector using
of the 95 per cent confidence interval error ellipse for 9 of the GPS data from stations located within the geologically rigid block

Table 2. Pacific, North American and Baja California Euler vectors.

Rotation polea Lon. (◦ E) Lat. (◦ N) Omega (deg Myr−1 ) σ max σ mina Azi. (◦ )b χ / d.o.f.
Pacific Plate–IGb00
PA1 (This study nine stationsc ) – IGb00 109.81 −63.67 0.681 ± 0.003 0.6 0.3 79 1.00
PA1 + GUAZ (10 stations)–IGb00 107.50 −63.75 0.677 ± 0.003 0.4 0.3 86 1.62
PA1 + CHAT (10 stations)–IGb00 111.31 −63.43 0.679 ± 0.003 0.5 0.2 90 1.47
North America–IGb00
Sella et al. 2006 −83.82 −5.66 0.195 ± 0.001 0.4 0.1 −1
North America–Pacific (geodetic)
Sella et al. 2006 – PA1 −75.89 50.16 0.769 ± 0.004 0.5 0.3 −85
Sella et al. 2006 – PA1 + GUAZ −77.32 50.11 0.766 ± 0.004 0.4 0.2 −87
Sella et al. 2006 – PA1 + CHAT −74.95 49.98 0.7680 ± 0.004 0.5 0.2 77
Beavan et al. 2002 −75.04 50.26 0.773 ± 0.005 0.4 0.2 94
DeMets & Dixon 1999 −73.70 51.50 0.765 ± 0.016 2.0 1.0 −85
Gonzalez-Garcia et al. 2003 −77.01 49.89 0.766 ± 0.007 0.3 0.2 70
Sella et al. 2002 −72.11 50.38 0.755 ± 0.004 0.6 0.4 −79
North America–Pacific (geological)
NUVEL-1A (DeMets 1994) −78.2 48.7 0.749 ± 0.012 1.3 1.2 −61
Baja California–IGb00
BAC1d –IGb00 106.63 −64.73 0.637 ± 0.034 4.4 0.4 −53 3.50
North America–Baja California
Sella et al. 2006 – BAC1 −78.11 50.16 0.725 ± 0.039 3.14 0.4 62
Note: The first plate rotates counter-clockwise relative to the second plate around the stated rotation pole.
a Lengths in degrees of the semi-major axes sig maj and semi-minor axes sig min of the 1σ pole error ellipse. Both axes are derived from a 2-D error
distribution.
b Azimuth of the semi-major ellipse axis in degrees clockwise from north.
c List of nine sites used see Table 1.
d List of 10 sites used see Table 3.

Angular velocity of PA1 relative to IGb00 in cartesian coordinates with covariance matrix. The X , Y , Z axes are parallel to (0◦ N,0◦ E), (0◦ N, 90◦ E), and
(90◦ N), respectively.
Omega (10−3 rads Myr−1 ): omegaX = −1.8186589 omegaY = 4.9377246 omegaZ = −10.6312041.
Covariance matrix (10−6 rads Myr– ): xx = 0.0064931 xy = 0.0006229 xz = −0.0002533 yy = 0.0010012 yz = −0.0000000 zz = 0.0026199.
Angular velocity of BAC1 relative to IGb00 in cartesian coordinates with covariance matrix:
Omega (10−3 rads Myr−1 ): omegaX = −1.3575097 omegaY = 4.5444985 omegaZ = −10.0461015.
Covariance matrix (10−6 rads Myr– ): xx = 0.0865653 xy = 0.1960590 xz = −0.1088797 yy = 0.4638906 yz = −0.2558797 zz = 0.1455125.


C 2007 The Authors, GJI, 170, 1373–1380
Journal compilation 
C 2007 RAS
1376 C. Plattner et al.

30˚ MARC
GUAZ

KWJ1
TRUKPOHN
0˚
KIRI KOK1
TUVA 22˚ HNLC
MAUI
THTI 20˚ MKEA
CKIS
30˚

200˚ 204˚
2mm/yr CHAT

120˚ 150˚ 180˚ 210˚ 240˚

Figure 3. Residual velocity with respect to best fitting Pacific Plate Euler vector (Table 1). Stations used for the computation of the Euler vector are shown
in red. Error ellipses indicate 95 per cent confidence interval, to distinguish significant residual motion. Fault and Plate boundaries from the same sources of
Fig. 1.

(Table 3). The Agua Blanca fault along with the San Pedro Martir other stations within the geologically rigid block are unaffected by
fault marks the northern boundary of the block (Fig. 1). The south- strain accumulation. Therefore, we believe our Baja California Eu-
ern block boundary includes the Bonfil fault, the Carrizal fault and ler vector adequately represents the rigid microplate motion within
the San Jose Cabo fault (Fig. 1). We exclude GPS stations LOSA the defined uncertainty limits.
and SPMX, located close to the Agua Blanca fault and San Pedro We find that sites AGUA, CARD, TOSA and CABO in the south-
Martir fault, respectively, and possibly influenced by strain accu- ern network move with very similar rate and direction with respect
mulation (Dixon et al. 2002). The shape of the Baja California to rigid Baja California. Therefore, our geodetic measurements can-
peninsula poses problems to an Euler vector calculation due to its not resolve motion across the Carrizal fault, which has sometimes
limited east–west extent, reflected in the orientation of the ellipsoid been described to cut through the peninsula (e.g. Hausback 1984).
describing its uncertainty (Table 2). The best fitting Euler vector This suggests that the southwestern tip of the peninsula belongs to
has a reduced χ 2 misfit of 3.5. All residual rates within the geolog- the rigid microplate (Fig. 4). Other stations in the southern tip of
ically rigid block are within uncertainties at 95 per cent confidence Baja California may represent the motion of smaller blocks that are
(Fig. 4; Table 3). However, the azimuths of the residuals do not ap- bounded by active normal faults. This is compatible with the recent
pear to be randomly oriented, as the northern network has its residual geological observations of Busch et al. (2006).
motion directed towards the south and vice versa (Fig. 4). For the
northern network the mean residual rate is 1.7 ± 0.8; for the south
Pacific – Baja California motion
it is 1.3 ± 0.8 mm yr−1 . This apparent convergence may reflect data
uncertainty, or perhaps internal deformation of the block. In the lat- We tested the significance of a separate Baja California microplate
ter case the average shortening strain rate between the two networks compared to a larger Pacific Plate including Baja California by ap-
is ∼1 × 10−16 s−1 . Additional EGPS data from Baja California will plying the F-test (Stein & Gordon 1984). Baja California acts as a
be required to distinguish between these hypotheses. separate microplate with 99 per cent confidence. Since Guadalupe
We tested for the effect of elastic strain accumulation at the Island shows significant residual motion with respect to our Pacific
edges of the microplate using the block model code DEFNODE Euler vector PA1 we also tested the possibility that it may be part of
(McCaffrey 2002). We found that stations LOSA and SPMX are the Baja California microplate. The F-test indicates that this is not
affected by strain accumulation when assuming a standard lock- the case at 99 per cent confidence, implying that the western border
ing depth of the block bounding faults between 10 and 20 km. All of the Baja California microplate lies east of Guadalupe Island.


C 2007 The Authors, GJI, 170, 1373–1380
Journal compilation 
C 2007 RAS
 Pacific Plate and Baja California microplate kinematics 1377

Table 3. Velocity of GPS stations in Baja California and residual motion with respect to the stable Pacific Plate PA1.a
IGb00 Residual to PA 1 Residual to BAC1

Site id Lon Lat. Ve σ Ve Vn σ Vn Rate σ rate Azi. Rate σ rate Azi.

(◦ E) (◦ N) (mm yr−1 ) (mm yr−1 ) (mm yr−1 ) (mm yr−1 ) (mm yr−1 ) (mm yr−1 ) (◦ ) (mm yr−1 ) (mm yr−1 ) (◦ )
AGUA −111.30 25.59 −48.4 0.3 19.6 0.6 3.2 0.7 141 1.3 0.6 −39
ancn −110.03 23.74 −51.0 1.7 18.6 0.7 3.1 1.0 163 2.1 1.7 −75
blnd −110.31 24.33 −48.7 0.7 20.2 0.7 3.1 0.8 118 2.0 0.8 −5
burr −110.07 23.52 −48.7 1.3 19.8 0.8 3.8 1.3 117 1.8 0.1 15
CABO −109.86 22.92 −50.4 0.7 18.6 0.5 3.6 0.7 143 1.0 0.7 −51
CADG −116.32 31.36 −41.6 0.8 19.5 0.7 6.4 0.9 137 2.0 0.9 123
CARD −110.78 24.15 −49.3 1.3 19.7 0.5 3.2 1.1 133 1.5 0.8 −23
cice −116.67 31.87 −40.2 0.9 17.4 0.8 8.8 1.0 142 4.3 0.9 141
COLO −116.21 31.10 −42.3 0.8 19.8 0.7 5.9 0.9 138 1.4 0.9 122
CONC −111.81 26.62 −45.7 1.1 18.7 0.8 5.3 1.1 133 1.1 1.2 95
ecer −109.81 24.18 −48.3 0.9 20.4 1.0 3.3 1.0 108 2.5 1.1 6
elal −116.21 31.85 −39.8 0.7 20.4 1.0 6.8 0.9 123 3.1 0.9 93
elch −115.05 31.49 −37.0 0.6 16.2 0.5 11.4 0.7 131 7.1 0.7 123
elco −116.17 32.47 −37.5 0.7 17.1 0.9 10.2 0.9 134 5.9 0.9 126
elja −115.76 31.49 −39.9 0.8 18.4 0.6 8.1 0.8 134 3.8 0.9 122
elmo −116.99 32.27 −39.8 1.1 18.1 1.0 8.4 1.1 139 3.9 1.1 135
emir −109.74 23.37 −51.3 1.0 17.7 0.6 3.8 0.8 167 2.1 1.1 −89
filo −116.44 31.72 −40.9 1.5 21.3 1.2 5.6 1.5 122 2.2 1.6 73
inde −115.94 31.55 −39.4 1.0 18.4 0.6 8.5 1.0 132 4.2 1.0 119
lagh −115.96 31.97 −37.8 0.9 18.1 0.7 9.6 0.9 128 5.4 1.0 116
losa −116.3 31.46 −41.6 0.9 19.6 0.6 6.3 0.9 137 1.9 0.9 122
mayo −115.24 31.99 −35.7 1.7 15.1 0.9 12.8 1.5 133 8.5 1.5 127
MELRa −115.74 30.98 −42.5 0.6 19.7 0.8 5.7 0.9 139 1.3 0.8 123
rive −109.53 23.55 −49.5 1.0 20.0 0.5 2.8 1.0 118 2.2 0.7 −11
rlov −116.63 32.12 −39.2 0.9 18.4 0.8 8.5 1.0 134 4.2 1.0 124
SAIS −116.22 31.19 −41.7 0.9 19.5 0.6 6.4 0.9 137 2.0 0.9 121
sald −115.39 31.77 −36.6 1.0 16.5 0.9 11.4 1.1 130 7.2 1.1 121
sfai −114.81 30.93 −42.8 1.0 16.1 0.7 8.2 0.9 156 4.0 0.8 171
SLRE −116.16 31.26 −42.4 0.7 19.0 0.5 6.4 0.7 141 1.9 0.7 134
sm01 −115.83 31.62 −41.0 0.9 17.9 1.3 7.7 1.3 143 3.2 1.2 142
spmx −115.47 31.05 −43.5 0.6 18.3 0.5 6.2 0.7 155 2.0 0.7 −179
TOSA −110.13 23.54 −49.6 0.7 19.5 0.6 3.2 0.8 130 1.5 0.7 −17
wmar −111.98 24.51 −50.0 1.4 21.1 1.0 2.0 1.3 131 2.6 1.2 −32
a Column headings are analogous to Table 1.

The magnitude of relative motion of Baja California with respect cated on Isla Margarita, shows a velocity with respect to the Pacific
to the Pacific Plate depends on the chosen stable Pacific Plate refer- Plate that is zero within uncertainty (Fig. 5b).
ence frame. The relative motion increases when using PA1 + CHAT,
while it decreases when using PA1 + GUAZ, with a range of dif-
North America–Pacific motion
ference of 1.8 mm yr−1 for the mean relative motion. However, all
models lead to significant relative motion of the Baja California mi- The precision of the geodetic estimates of NA-Pacific motion has
croplate with respect to the Pacific Plate at 95 per cent confidence improved with time, as more stations become available and GPS
interval. In the following we use model PA1. time-series become longer (Argus & Heflin 1995; Larson et al. 1997;
Except for WMAR, all the EGPS velocities in Baja California DeMets & Dixon 1999; Freymueller et al. 1999; Beavan et al. 2002;
relative to the stable Pacific Plate are significant at 95 per cent con- Sella et al. 2002; Gonzales-Garcia et al. 2003). Our result for NA-
fidence interval (Table 3; Figs 5a and b). On the rigid microplate the PA1 together with previous results is listed in Table 2, and illustrated
average velocity with respect to the Pacific Plate is 4.9 mm yr−1 , in Fig. 6.
ranging from 3.2 ± 0.7 (AGUA) to 6.4 ± 0.9 mm yr−1 (CADG). We obtain a shift in the location of the North America–Pacific
The average rate of the northern network (CADG, COLO, LOSA; pole of rotation (Table 2) when we use PA1+CHAT or PA1+GUAZ
MELR; SAIS, SLRE) is 6.2 mm yr−1 . In the southern network instead of PA1 for the Pacific Plate. The results are comparable to the
(AGUA, CABO, CARD and TOSA) the average rate is 3.7 mm difference in location of the North America–Pacific pole of rotation
yr−1 . Outside the rigid microplate, deformation of parts of the plate from Beavan et al. (2002) compared to the one of Gonzalez-Garcia
boundary zone can be observed. The northern part of the penin- et al. (2003). This may be because Beavan et al. (2002) use CHAT
sula shows an increase of velocity from west to east, approach- and EGPS data from the Campbell Plateau (analogous to PA1 +
ing the North America–Baja California boundary, that is, the San CHAT), while Gonzalez-Garcia et al. (2003) use one CGPS and
Andreas/Gulf of California system. The velocity reaches 12.8 ± three EGPS stations from Guadalupe Island (analogous to PA1 +
1.5 mm at MAYO, indicating significant strain accumulation along GUAZ).
the northern faults (Fig. 5a). In the southern part of the peninsula no Comparing our Euler vectors we see that the North America–
such pattern is observed. West of the peninsula, station WMAR, lo- Pacific rotation rate is significantly faster than North America–Baja


C 2007 The Authors, GJI, 170, 1373–1380
Journal compilation 
C 2007 RAS
1378 C. Plattner et al.

32˚ his velocity (51.6 ± 1.9 mm yr−1 ) agrees with our results (Fig. 7).
This implies that the average spreading rate in the Gulf of California
during the past 3 Myr is comparable to the geodetic rate over the last
decade and the same is true for the rate of North America–Pacific
Plate motion. We can exclude thermal contraction of the seafloor
to be responsible for the velocity difference between Baja, and also
30˚ Guadalupe, with respect to the stable Pacific Plate as the contraction
can explain only ∼1.35 per cent of the spreading rate (Kumar et al.
2006), not the observed ∼10 per cent of difference.

28˚
33˚

ELCO
26˚ ELMO
RLOV
LAGH MAYO
32˚ ELAL
CICE SALD
FILO SM01
24˚
LOSA INDE ELJA ELCH
CADG
2mm/yr SLRE
SAIS
COLO MELR
31˚ SPMX
SFAI
Figure 4. Stations on the Baja California microplate: residual motion with
respect to Baja California Euler vector (Table 1). Error ellipses indicate 95
per cent confidence interval.
6mm/yr
California. At the location 23.5◦ N, 108.5◦ E, at the spreading cen-
tre in the Gulf of California (DeMets 1995), we calculate North 27˚
America–Pacific motion to be 51.1 ± 0.4 mm yr−1 at an azimuth
of 125 ± 1◦ clockwise from north. Relative motion between North CONC
America and Baja California at the same location is only 46.8 ±
0.4 mm yr−1 , 124 ± 1◦ clockwise from north. This leaves a residual
motion of 4.3 ± 0.8 mm yr−1 between Baja California and the Pacific 26˚
Plate. The negligible difference in azimuth, which is also indicated
by the proximity of the two Euler poles, shows that Baja California AGUA
is moving in approximately the same direction as the Pacific Plate
with respect to North America.
Along the Gulf of California, the geodetic rate for North 25˚
America–Baja California derived from our Euler vector decreases
from south to north, 46.8 ± 0.4 mm yr−1 (23.5◦ N, 108◦ W) to WMAR
43.1 ± 0.4 mm yr−1 at the Colorado River delta (31.8◦ N, BLND ECER
−114.5◦ W).
A key constraint for the North America–Pacific Euler vector mag- CARD
24˚
nitude in NUVEL-1A (DeMets et al. 1994) is magnetic anomaly data ANCN
from the Gulf of California. Due to the rigid block motion of Baja RIVE
TOSA
California and incomplete coupling with the Pacific Plate, this rate BURR
must in reality represent North America–Baja California relative EMIR
motion. When we compare our geodetic rate for North America–
23˚
Baja California (46.8 ± 0.4 mm yr−1 ) with the NUVEL-1A rate CABO
(47.4 ± 1.2 mm yr−1 ) at the latitude of the spreading centre (23.5◦ N, 3mm/yr
108.5◦ E) we see that these rates agree within uncertainties. On the
other hand, when we compare at the same location our geodetic
North America–Pacific rate (51.1 ± 0.4 mm yr−1 ) with an estimate Figure 5. Northern (a) and southern (b) GPS network on Baja California:
of DeMets (1995) that excludes the magnetic anomaly data from the velocity with respect to stable Pacific Plate Euler vector 1 (Table 1). Error
Gulf of California and other problematic data sets from NUVEL-1A, ellipses indicate 95 per cent confidence interval.


C 2007 The Authors, GJI, 170, 1373–1380
Journal compilation 
C 2007 RAS
 Pacific Plate and Baja California microplate kinematics 1379

Omega & Normark 1979; Legg et al. 1991; Fletcher & Munguia 2000;
[°/Myr] Michaud et al. 2004).
A possible explanation for why the Baja California microplate is
NA-BAC1 B02 0.78
50˚ only partially coupled to the Pacific Plate and for activity along the
REVEL western Baja California shear zone (Dixon et al. 2000b) is the colli-
GG03 NA-PA
0.76 sion of the northwestward moving microplate with North America
along the Transverse Ranges and the big bend of the San Andreas
48˚ fault. This impact may cause the microplate to shear off the Pa-
0.74
cific Plate, along an inherited weak zone, the former Farallon–North
NUVEL-1A
American Plate boundary along the western coast of Baja California.
280˚ 288˚ 0.72 In this case the shear zone may have formed from north to south. This
284˚
may explain why the northern part of the Baja California microplate
Figure 6. Location and magnitude of North America–Pacific (NA-PA) Eu- shows larger relative motion with respect to Pacific Plate than south-
ler vectors and North America–Baja California (NA-BAC1) Euler vector. If ern part. An alternative explanation for this observed pattern, also
not otherwise indicated, results are from this study. NUVEL-1A (DeMets consistent with the idea of collision along the northern boundary, is
et al. 1994), Revel (Sella et al. 2002), B02 (Beavan et al. 2002) and GG03 internal deformation within the Baja California microplate.
(Gonzalez-Garcia et al. 2003). Ellipses show the 1σ error from a 2-D distri-
bution.
AC K N OW L E D G M E N T S

Geologic Geodetic Geodetic This research is partially supported by DFG. CP is partially

supported by the international graduate school thesis within the
54 Elitenetzwerk Bayern. RM received support from the BMBF, Ger-
many. THD is partially supported by NSF grant OCE0505075. We
53 are very grateful to the operators of the CGPS sites used in this study,
Velocity [mm/yr]

52 including the South Pacific Sea Level and Climate Monitoring Pro-
gram funded by the Australian Government Aid project operated by
51 Geoscience Australia, the member institutions of IGS, the Western
50 Pacific Integrated Network of GPS (WING) operated by the ERI,
and SOPAC. Some of the EGPS data was collected by SCEC, in par-
49 ticular Javier Gonzalez Garcia and Duncan Agnew. We express our
48 gratitude to all people that helped in the collection of the GPS data,
in particular the technicians from CICESE. We are very thankful to
47 Paul Umhoefer for providing fault map data, and to Kevin Furlong,
46 NA-Baja Jim Ray, and the reviewers Chuck DeMets and Joann Stock, whose
NA-PA NA-PA comments helped to improve this paper.
NUVEL-1A DeMets(1995) This study REVEL B02 GG03

REFERENCES
Figure 7. Magnitude of NA-PA and NA-Baja relative motion computed for a
location at the spreading centre in the Gulf of California (23.5◦ N, 108.5◦ W). Argus, D.F. & Heflin, M.B., 1995. Plate motion and crustal deformation es-
The geodetic rates of NA-PA from different studies (for abbreviations see timated with geodetic data from the Global Positioning System, Geophys.
Fig. 6) agree with each other and with the rate from geological model of Res. Lett., 22, 1973–1976.
DeMets (1995). Nuvel1A (DeMets et al. 1994) agrees with NA-Baja. Atwater, T. & Stock, J., 1988. Pacific–North America Plate tectonics of the
neogene Southwestern United States—an update, in Integrated Earth and
Environmental Evolution of the Southwestern United States: The Clarence
DISCUSSION A. Hall, Jr. Volume, pp. 393–420, eds Ernst, W.G. & Nelson, C.A.,
Baja California and its rigid block motion is an analogue for other Bellwether Publishing, Columbia.
Beavan, J., Tregoning, P., Bevis, M., Kato, T. & Meertens, C., 2002. Mo-
terranes that are transported northwest with the Pacific Plate while
tion and rigidity of the Pacific Plate and implications for plate boundary
interacting with North America (Atwater & Stock 1988; McQuarrie deformations, J. Geophys. Res., 107, 2261, doi:10.1029/2001JB000282.
& Wernicke 2005). The essentially rigid behaviour of such a mi- Busch, M. et al. 2006. Late Quaternary faulting in the Cabo San Lucas-La
croplate on geodetic timescales preserves the coherence of a terrane Paz Region, Baja California, Eos Trans. AGU, 87(52), Fall Meet. Supol.,
during translation over geological timescales. The correspondence Abstract T41D-1612.
of terrane transport direction with the Pacific Plate motion is con- DeMets, C., 1995. A reappraisal of seafloor spreading lineations in the Gulf
sistent with the idea that the Pacific Plate is the driving force for of California: implications for the transfer of Baja California to the Pacific
terrane transport in this region. The subducted extent of the oceanic plate and estimates of Pacific-North America motion, Geophys. Res. Lett.,
microplates that were captured by the Pacific Plate (Nicholson et al. 22, 3545–3548.
1994; Stock & Lee 1994) may also be an important influence. DeMets, C. & Dixon, T.H., 1999. New kinematic models for Pacific-North
America motion from 3 Ma to present, I: evidence for steady motion and
The direction of Pacific–North America and Baja California–
biases in the NUVEL-1A model, Geophys. Res. Lett., 26, 1921–1924.
North America motion is similar, but Baja California moves sig- DeMets, C., Gordon, R.G., Argus, D.F. & Stein, S., 1990. Current plate
nificantly slower. This result supports the suggestion of a western motions, Geophys. J. Int., 101, 425–478.
Baja California shear zone (Dixon et al. 2000b), and is consistent DeMets, C., Gordon, R.G., Argus, D.F. & Stein, S., 1994. Effect of recent
with observations of right-lateral offset on Quaternary faults, and revisions to the geomagnetic reversal time-scale on estimates of current
seismicity along the southwestern coast of Baja California (Spencer plate motions, Geophys. Res. Lett., 21, 2191.


C 2007 The Authors, GJI, 170, 1373–1380
Journal compilation 
C 2007 RAS
1380 C. Plattner et al.

Dixon, T.H., Miller, M., Farina, F., Wang, H. & Johnson, D., 2000a. Present Boundary Zones, Geodyn. Ser., Vol. 30, pp. 101–122, eds S. Stein & J.
day motion of the Sierra Nevada block and some implications for the Basin Freymueller.
and Range province, North America Cordillera, Tectonics, 19, 1–24. McQuarrie, N. & Wernicke, B.P., 2005. An animated tectonic reconstruc-
Dixon, T.H. et al. 2000b. New kinematics for Pacific-North America motion tion of southwestern North America since 36 Ma. Geosphere, 1(3), doi:
from 3 Ma to present, II: evidence for a “Baja California shear zone”, 10.1130/GES00016.1, 147–172.
Geophys. Res. Lett., 27, 3961–3964. Michaud F. et al. 2004. Motion partitioning between the Pacific plate, Baja
Dixon, T.H. et al. 2002. Seismic cycle and rheological effects on esti- California and the North America plate: the Tosco-Abreojos fault revis-
mation of present-day slip rates for the Agua Blanca and San Miguel- ited, Geophys. Res. Lett., 31, L08604, doi:10.1029/2004GL019665.
Vallecitos faults, northern Baja California, Mexico, J. Geophys. Res., 107, Minster, J.B., Jordon, T.H., Molnar, P. & Haines, E., 1974. Numerical mod-
doi:10.1029/2000JB000099. elling of instantaneous plate tectonics, Geophys. J. R. astr. Soc., 36, 541–
Fletcher, J.M. & Munguia, L., 2000. Active continental rifting in southern 576.
Baja California, Mexico: implications for plate motion partitioning and Nicholson, C., Sorlien, C.C., Atwater, T., Crowell, J.C. & Luyendyk, B.P.,
the transition to seafloor spreading in the Gulf of California, Tectonics, 1994. Microplate capture, rotation of the western Transverse Ranges, and
19, 1107–1123. initiation of the San Andreas transform as a low-angle fault system, Ge-
Freymueller, J.T., Murray, M.H., Segall, P. & Castillo, D., 1999. Kinematics ology, 22, 491–195.
of the Pacific–North America plate boundary zone, northern California, Ray, J., Dong, D. & Altamimi, Z., 2004. IGS reference frames: status and
J. Geophys. Res., 104, 7419–7441. future improvements, GPS Solutions, 8, 251–266.
Gastil, R.G., Phillips, P.R. & Allison, C.E., 1975. Reconnaissance Geology Sella, G.F., Stein, S., Dixon, T.H., Craymer, M., James, T.S., Mazzotti,
of the State of Baja California, Geol. Soc. Am. Mem., 140, 170 p. S. & Dokka, R.K., 2006. Observation of glacial isostatic adjustment
Gonzales-Garcia, J.J., Prawirodirdjo, L., Bock, Y. & Agnew, D., 2003. in “stable” North America with GPS, Geophy. Res. Lett., 34, L02306,
Guadalupe Island, Mexico as a new constraint for Pacific plate motion, doi: 10.1029/2006GL027081.
Geophys. Res. Lett., 30(16), 1872, doi:10.1029/2003GL017732. Sella, G.F., Dixon, T.H. & Mao, A.L., 2002. REVEL: a model for
Govers, R. & Wortel, M.-J.-R., 2005. Lithosphere tearing at STEP faults: Recent plate velocities from space geodesy, J. Geophys. Res., 107,
response to edges of subduction zones, Earth Plant. Sc. Lett., 236(1–2), doi:10.1029/2000JB000033.
505–523. Spencer, J.-E. & Normark, W., 1979. Tosco-Abreojos fault zone: a neogene
Hausback, B.P., 1984. Cenozoic volcanic and tectonic evolution of Baja transform plate boundary within the Pacific margin of southern Baja Cal-
California Sur, Mexico, in Geology of the Baja Peninsula, Frizell, V.A. ifornia, Mexico, Geology, 7, 554–557.
(ed.), Pac. Sect. Soc. of Econ. Paleontol. and Mineral., Vol. 39. pp. 219– Stein, S. & Gordon, R.G., 1984. Statistical tests of additional plate boundaries
236, Los Angeles, CA. from plate motion inversions, Earth Plan. Sci. Lett., 69(2), 401–412.
Kumar, R.R., Gordon, R.G. & Landis, C.M., 2006. Horizontal thermal con- Stock, J. M. & J. Lee, 1994. Do microplates in subduction zones leave a ge-
tractional strain of oceanic lithosphere: the ulimate limit to the rigid ological record? Tectonics, 13(6), 1472–1487, doi: 10.1029/94TC01808.
plate hypothesis, Eos Trans. AGU, 87(52), Fall Meet. Suppl., Abstract Suarez-Vidal, F., Armijo, R., Morgan, G., Bodin, P. & Gastil, R.G., 1991.
G41A-04. Framework of recent and active faulting in Northern Baja California, in
Lachenbruch, A.H., Sass, J.H., & Galanis, S.P., Jr., 1985. Heat flow in south- Gulf and Peninsular Province of the Californias, Vol.47, pp. 285–300, eds
ernmost California and the origin of the Salton trough, J. Geophys. Res., J. Dauphin & B. Simoneit, Am. Assoc. Pet. Geol. Mem.
90, 6709–6736. Umhoefer, P.J., 2000. Where are the missing faults in translated terranes?,
Larson, K.M., Freymueller, J.T. & Philipsen, S., 1997. Global plate velocities Tectonophysics, 326, 23–35.
from the Global Positioning System, J. Geophys. Res., 102, 9962–9982. Umhoefer, P.J. & Dorsey, R.J., 1997. Translation of terranes; lessons from
Legg, M.R., Wong, V.O. & Suarez-Vidal, F., 1991. Geologic structure and central Baja California, Mexico, Geology, 25, 1007–1010.
tectonics of the inner continental borderland of northern Baja California, Ward, S., 1990. Pacific-North America Plate motions: new results from
in Gulf and Peninsular Province of the Californias, Vol.47, pp. 145–177, very long baseline interferometry, J. Geophys. Res., 95, 21 965–
eds J. Dauphin & B. Simoneit, Am. Assoc. Pet. Geol. Mem.. 21 981.
Mao, A., Harrison, G.A. & Dixon, T.H., 1999. Noise in GPS coordinate time Zumberge, J.F., Heflin, M.B., Jefferson, D.J., Watkins, M.M. & Webb, F.H.,
series, J. Geophys. Res., 104, 2797–2816. 1997. Precise point positioning for the efficient and robust analysis of
McCaffrey, R., 2002. Crustal block rotations and plate coupling, in Plate GPS data from large networks, J. Geophys. Res., 102, 5005–5017.


C 2007 The Authors, GJI, 170, 1373–1380
Journal compilation 
C 2007 RAS