Crustal structure of central Lake Baikal:
Insights into intracontinental rifting
Uri S. ten Brink and Michael H. Taylor1
U.S. Geological Survey, Woods Hole, Massachusetts, USA
Received 26 January 2001; revised 15 October 2001; accepted 25 October 2001; published 16 July 2002.
[1] The Cenozoic rift system of Baikal, located in the interior of the largest continental
mass on Earth, is thought to represent a potential analog of the early stage of breakup of
supercontinents. We present a detailed P wave velocity structure of the crust and
sediments beneath the Central Basin, the deepest basin in the Baikal rift system. The
structure is characterized by a Moho depth of 39–42.5 km; an 8-km-thick, laterally
continuous high-velocity (7.05–7.4 km/s) lower crust, normal upper mantle velocity
(8 km/s), a sedimentary section reaching maximum depths of 9 km, and a gradual increase
of sediment velocity with depth. We interpret the high-velocity lower crust to be part of the
Siberian Platform that was not thinned or altered significantly during rifting. In
comparison to published results from the Siberian Platform, Moho under the basin is
elevated by <3 km. On the basis of these results we propose that the basin was formed by
upper crustal extension, possibly reactivating structures in an ancient fold-and-thrust
belt. The extent and location of upper mantle extension are not revealed by our data, and it
may be offset from the rift. We believe that the Baikal rift structure is similar in many
respects to the Mesozoic Atlantic rift system, the precursor to the formation of the North
Atlantic Ocean. We also propose that the Central Baikal rift evolved by episodic fault
propagation and basin enlargement, rather than by two-stage rift evolution as is commonly
assumed. INDEX TERMS: 8109 Tectonophysics: Continental tectonics—extensional (0905); 8107
Tectonophysics: Continental neotectonics; 8120 Tectonophysics: Dynamics of lithosphere and mantle—
general; 9320 Information Related to Geographic Region: Asia; KEYWORDS: Lake Baikal, continental rifts,
Newark basin, crustal extension, plate driving forces, seismic velocity structure
1. Introduction
[2] Most aspects of the deformation of intracontinental
rifts can be explained by a combination of several physical
parameters [e.g., Ruppel, 1995]: (1) the rheology of the
continental lithosphere, in particular, the crust, (2) preexist-
ing heterogeneities within the lithosphere, (3) the thermal
structure of the lithosphere, and (4) the rate of extension.
Together these parameters determine rift geometry (local-
ized versus distributed crustal extension, reversal in rift
polarity, etc.), subsidence pattern, rift-shoulder uplift, the
amount and composition of rift-related magmatism,
the magnitude of crustal thinning, and rift evolution. The
interaction among these parameters is responsible for a
variety of rifting mechanisms such as pure shear (litho-
spheric necking), simple shear, lower crustal flow, and
combinations of the above [e.g., Ruppel, 1995].
[3] The Baikal rift system is the largest active Eurasian
rift system (Figure 1). It is therefore thought to represent a
potential analog for the breakup of supercontinents and to
provide a natural laboratory in which the early stages of
continental rifting can be observed. The Baikal rift system
has, for many years, been the focus of a debate about the
driving forces of continental rifting. Driving forces for
intracontinental rifting fall into three categories: (1) far-field
stresses within the continental lithosphere due to either plate
boundary forces or sublithospheric drag; (2) active buoy-
ancy forces due to dynamic mantle upwelling; and (3) other
gravitational effects, such as gravitational collapse and
lithospheric delamination [e.g., Ruppel, 1995, and referen-
ces within]. Far-field stresses that could drive the Baikal rift
system are believed to be generated by the Indo-Asian
collision [Tapponnier and Molnar, 1979] and possibly by
subduction of the Pacific plate along its western boundary
[e.g., Calais et al., 1998] and by the configuration of the
Siberian Platform as a rigid promontory (Figure 2) [Petit
et al., 1997]. An alternative driving force for this rift system
is an active asthenospheric upwelling in the vicinity of the
rift system [e.g., Zorin and Lepina, 1985], farther to the SE
[Windley and Allen, 1993], or on a larger scale under Tibet,
Mongolia, and parts of China [Yin, 2000]. A variety of
evidence about volcanism, deformation geometry, and ther-
mal structure as well as comparisons with other rift systems,
such as the East African rift system and the Rio Grande Rift
[Lipman et al., 1989; Lysak, 1987], have been used to
support the different points of view. However, these lines
JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 107, NO. B7, 10.1029/2001JB000300, 2002
1Now at the Department of Earth and Space Sciences, University of
California, Los Angeles, California, USA.
This paper is not subject to U.S. Copyright.
Published in 2002 by the American Geophysical Union.
ETG 2 - 1
of evidence are generally indicative of the particular combi-
nation of the above physical parameters, not the driving
forces. For example, lithospheric extension due to far-field
stresses can lead to passive asthenospheric upwelling, which
produces many of the same features as those generated by
active asthenospheric upwelling [Holbrook and Kelemen,
1993; Mutter et al., 1988]. We therefore discuss the impli-
cations of our results for the rifting mechanisms, not for the
driving forces.
[4] We present a detailed P wave velocity structure of the
sediments and crust beneath the Central Basin of Lake
Baikal (Figures 2 and 3) derived from a joint wide-angle
and multichannel seismic reflection (MCS) experiment,
conducted in 1992 by the U.S. Geological Survey and the
Russian Academy of Sciences. The data offer insight to the
rift geometry at depth and the possible mode of extension.
The velocity model helps to differentiate between some
proposed mechanisms for crustal extension of Lake Baikal
[Burov et al., 1994; Lesne et al., 2000; Petit et al., 1997;
Poort et al., 1998; van der Beek, 1997]. We use the results
to draw an analogy between the geometry of the Baikal rift
system and the early Mesozoic Atlantic rift system, the
precursor to the formation of the North Atlantic Ocean,
where the role of an active asthenopheric upwelling is still
debated [e.g., Holbrook and Kelemen, 1993]. We also
discuss evidence against the commonly cited two-stage
rifting history of the Baikal rift.
2. Tectonic Setting
[5] The 
1800-km rift system is composed of 15 indi-
vidual basins within the Sayan-Baikal fold-and-thrust belt
along its boundary with the Siberian Platform (Figures 1
and 2) [Keller et al., 1995, and references therein]. The rift
system is dominated by extensional tectonics with a strike-
slip component, potentially a result of the southeastward
movement of the Amur plate relative to the stable Siberian
Platform [e.g., Lesne et al., 1998]. GPS measurements show
crustal extension across Lake Baikal at a rate of 4.5 ±
1.2 mm/yr in a WNW-ESE direction [Calais et al., 1998]
and an extension at a rate of 6.3 ± 1.8 mm/yr in a 125
 ± 15

direction between Irkutsk and Ulan Bataar in Mongolia
[Calais and Amarjargal, 2000]. Active extension is also
evidenced by frequent magnitude 5 and larger earthquakes
[Zonenshain and Savostin, 1981; Doser, 1991; Deverchere
et al., 1993]. The Baikal rift system is located within a
mountainous region (Figure 2), which is sometimes referred
to as a domal uplift [Kiselev, 1987; Logatchev and Zorin,
Figure 1. Locations of the 15 basins, which comprise the Baikal rift system and the location of
Cenozoic volcanic fields [Kiselev, 1987]. Inset is a generalized tectonic map of East Asia.
ETG 2 - 2 TEN BRINK AND TAYLOR: CRUSTAL STRUCTURE OF BAIKAL RIFT
1987]. The prerift structural framework of the Baikal rift
zone is characterized by northeast-southwest trending Pre-
cambrian to lower Paleozoic fold-and-thrust belts, widely
believed to be a controlling factor in rift evolution [Keller et
al., 1995; Mats, 1993; Zamarayev and Ruzhich, 1978].
[6] The rift basins of Lake Baikal occupy the central 650
km of the rift system and include three major sedimentary
basins, the Southern Basin, the Central Basin and the
Northern Basin and a deep smaller depocenter, the Selenga
Basin (inset in Figure 3) [Hutchinson et al., 1992]. The floor
of the Central Basin has the lowest elevation of the entire
rift system, 
1650 m below lake level and 
1190 m below
mean sea level. The deformation in this section of the rift is
predominantly confined to the width of the lake, 
60 km.
The Morskiy Fault along the northwest edge of the basin
appears to be the major boundary of the basin (Figure 3).
Bathymetry and overall sediment thickness increase toward
this fault [Hutchinson et al., 1992; Moore et al., 1997;
Scholz et al., 1993], giving the basin an appearance of a half
graben dominated by dip-slip displacements. However, the
detailed basin evolution appears to be more complex and
may involve periods of transtensional motion, as indicated
by periods of relatively symmetric sediment accumulation
[Hutchinson et al., 1992; Moore et al., 1997]. In addition, a
new boundary fault, the Primorskiy Fault (Figure 3) appears
to be progressively replacing the Morskiy Fault as the main
boundary fault of the rift [Agar and Klitgord, 1995].
3. Seismic Data and Model
[7] A 247-km wide-angle seismic refraction line was
recorded by five ocean bottom seismometers (OBS) along
the axis of the central basin, with two 60-L air guns as the
source (Figure 3). The northern 140 km of the line were
repeated with six OBS and a tuned 10-gun array with a total
volume of 27.3 L. Multichannel seismic reflection (MCS)
data were recorded simultaneously by a 96-channel, 2400-
m-long streamer along this portion of the line [Scholz et al.,
1993]. For the southern 100 km, MCS line 11 was collected
at distances between 4 and 13 km southeast of the wide-
angle transect (Figure 3). Continuous high-amplitude wide-
angle reflections and refractions from the sedimentary
Figure 2. Topography of the Baikal rift and the surrounding areas. Basins form depressions and Lake
Baikal. Lake level is at 455 m above sea level (asl). White dashed line separates the Siberian Platform
from the Proterozoic-Paleozoic fold and thrust belt of Trans-Baikal Mongolia. White line shows location
of our seismic refraction profile. The high elevation around and southeast of the Baikal rift zone is
interpreted by some to represent uplift due to mantle upwelling [e.g., Windley and Allen, 1993].
TEN BRINK AND TAYLOR: CRUSTAL STRUCTURE OF BAIKAL RIFT ETG 2 - 3
section and the top of basement were observed in each OBS
record to shot-receiver offsets of 40 km. Discontinuous
arrivals are observed to maximum offsets of 170 km and
represent reflections and refractions from the crystalline
basement and, in four OBS, from the Moho (Figure 4).
[8] Processing of the wide-angle data included a linear
move out correction, spiking deconvolution, low-pass filter
of 14 Hz, and automatic gain control (AGC). Processing of
the MCS data included prestack spiking deconvolution,
iterative velocity analysis, and AGC. Poststack processing
included predictive deconvolution, AGC, vertical sum,
time-variant filtering, and a cascaded migration scheme
[Agena et al., 1994]. Forward travel time modeling of the
wide-angle data was carried out utilizing the interactive
RayGUI code [Loss et al., 1998], based on Zelt and Smith’s
[1992] ray-tracing package. Further improvement to the
solution was achieved by linearized inversion with the same
package (Figure 5). We constructed the starting model by
digitizing the seismic structure of basin fill deposits and
sparse basement reflections derived from coincident vertical
incidence data acquired during the 1992 cruise. Because the
southern 100-km portion of the refraction transect lacks a
coincident MCS profile, we used information from MCS
cross lines there (Figure 3). Tomographic inversion was not
attempted because of the small number of data and the poor
signal-to-noise ratio. Instead, we obtained an estimate of
model resolution by perturbing various parts of the model
and observing the decay in model fit (Figure 6) [e.g.,
Holbrook et al., 1996]. The velocity model consists of nine
layers (excluding the water layer), with five layers of
sedimentary section and four layers composing the crust
and upper mantle (Figure 7a). Ray coverage for the model is
generally very good in the sedimentary section and upper-
most crust, with resolution decreasing with depth to Moho.
An average root-mean-square (RMS) misfit for the entire
model is 0.127 s, and the RMS misfit for the sedimentary
section and basement is 0.1 s.
4. Results
4.1. Sedimentary Section
[9] Our velocity model indicates a maximum thickness of

9 km for the sedimentary section, northeast of the Selenga
Delta (km 75, Figure 7b), where reflection data did not
image basement beneath the deep part of the Central Basin
(Figure 8). Depth to basement reaches 8–9 km under other
parts of the Central Basin (120–170 along the model).
Depth to basement under the Selenga Basin at the south-
ernmost part of the profile reaches 7 km, a value similar to
the depth of 6–7.5 km estimated from MCS data [Hutch-
inson et al., 1992]. The transition to basement is taken as
the 5.2 km/s isovelocity contour, which corresponds to a
clear basement reflection on the MCS data between km 170
and 230 (Figure 8). Velocity of basin fill deposits within the
Central Basin increases gradually with increasing depth
from the lake floor without abrupt changes in velocity
gradient (Figure 7), probably due to sediment compaction.
A few wide-angle reflections can be identified in the
refraction data in the upper part of the sedimentary section
(Figures 4 and 5), but they do not form a coherent reflecting
horizon across the model. The gradual increase in velocity
and the lack of a reflecting horizon do not support a division
of the sedimentary section in the Central Basin into two
units of different physical properties, as previously sug-
gested based on their reflectivity in the MCS data [Hutch-
inson et al., 1992; Scholz et al., 1993]. Amplitude analysis
of these data has indeed shown that the loss of reflectivity at
4–4.5 s two-way travel time cuts across seismic horizons
Figure 3. Location of seismic refraction line (gray line) and multichannel seismic reflection lines shown
(heavy black lines) or used for interpretation (thin black lines). Ocean bottom seismometer (OBS) names
are denoted by letters. Inset shows basins and boundary faults in the vicinity of the Central Basin
(modified from Agar and Klitgord [1995]).
ETG 2 - 4 TEN BRINK AND TAYLOR: CRUSTAL STRUCTURE OF BAIKAL RIFT
and is perhaps caused by silica diagenesis [Lee et al., 1996].
Furthermore, the loss of reflectivity was shown to be
gradual and to start above the suggested upper/lower unit
boundary [Lee et al., 1996].
[10] The basement under the Selenga Delta is shallower
than under the adjacent Central and Selenga Basins (Figure
8b). An MCS grid of profiles in the Selenga Delta indicates
that the basement high comprises a series of fault blocks
with large (3–4 s) vertical throws between them [Scholz
et al., 1993; Scholz and Hutchinson, 2000]. Since the
Selenga River is the largest sediment source of Lake Baikal,
draining much of Mongolia, sedimentation has nearly kept
pace with the tectonic movement, resulting in a shallower
bathymetry than elsewhere in the lake. Excess sediments are
funneled northward to the southern part of the Central Basin
and southward to the Selenga Basin [Scholz et al., 1993;
S. M. Colman et al., Quaternary depositional patterns and
environments in Lake Baikal from high-resolution seismic
stratigraphy and coring, unpublished manuscript, 2001].
[11] Near the north end of the Central Basin an abrupt
decrease in basement depth by 2–3 km can be seen in the
velocity model northeast of 
170–180 km (Figure 7). The
overlying sedimentary section on the MCS profile in this
region thins to the northeast and is cut by faults (Figure 8).
However, modeling the exact location and shape of the
basement rise was difficult because of a lack of travel time
Figure 4. Examples of wide-angle refraction records. Note that discontinuous events can be detected to
a maximum offset of 170 km. Ps, P1, P2, P3 are turning waves from the sedimentary section, upper crust,
middle crust, and the basal high-velocity layer, respectively. PsP, P2P, P3P, and PmP are wide-angle
reflections from basement, the middle crust, the top of the basal high-velocity layer, and Moho,
respectively. Pn is upper mantle head wave.
TEN BRINK AND TAYLOR: CRUSTAL STRUCTURE OF BAIKAL RIFT ETG 2 - 5
reciprocity between adjacent OBSs at this location. We
explain the lack of reciprocity by out-of-plane effects,
perhaps because the profile there lies close to the Morskiy
Fault (Figure 3). The Morskiy Fault is a normal fault
dipping to the southeast, and at depth it likely underlies
our refraction profile. Our velocity model and the grid of
MCS profiles in the area [Moore et al., 1997] show that
basement continues to shallow over a short distance at 190–
200 km and is 
1.5 km below lake bottom in the north-
ernmost 40 km of the line (Figures 7b and 8).
[12] The distribution of high-velocity (3.0–5.2 km/s)
sediments is highly variable (Figure 7b). High-velocity
sediments are 
4 km thick between 80 and 180 km. They
are thinner below the Selenga Delta and are missing along
the northern part of the Central Basin and Barguzin Bay.
This distribution suggests either original deposition in
basins of limited size or exhumation and subsequent erosion
of uplifted blocks. Seismic stratigraphic analysis suggests
that the Central Basin was initially smaller and lengthened
with time [Moore et al., 1997]. For example, the sedimen-
tary thickness below Horizon B4 in the Northern Basin, the
Academic Ridge, and the northern part of the Central Basin
is small (<500 m) compared with the deep part of the
Central Basin (compare Figures 8 and 9). Sediments on the
Academic Ridge appear to have filled local depressions and
are not heavily eroded [Moore et al., 1997]. In contrast to
the high-velocity sediments, lower-velocity sediments (1.5–
3 km/s) cover the entire Central Basin. Their thickness is

3 km at the center of the basin, 2–3 km at the Selenga
Delta, 1–1.5 km at the Barguzin Delta, and 
2 km at the
southern part of the Northern Basin.
[13] Severe attenuation of seismic energy and anoma-
lously low apparent velocities within the shallow sedimen-
tary section are apparent in two OBS records (wide-angle
data from OBS AA3 and B1, not shown) south of Ol’khon
Island that were recorded with two different active source
configurations (Figure 3). The coincident MCS profile at km
100–125 along the model profile shows chaotic or disrupted
reflections capped by a reflector with reverse polarity
(Figures 8 and 3b, shots 4100–4450, of Hutchinson et al.
[1992]). The above seismic characteristics are interpreted to
represent free gas, possibly resulting from discharge of
methane gas in the central part of the Central Basin due to
rapid sediment influx from the Selenga Delta. A bottom-
simulating reflector (BSR) extends from the gas pocket
northward along the Central Basin (Figure 8). A BSR with
reverse polarity indicating the presence of free gas was also
observed south of the Selenga Delta [Vanneste et al., 2001].
4.2. Crust
[14] Crustal and Moho reflections and crustal refractions
are observable at offsets ranging from 25 to 170 km
Figure 5. (a) Observed (dots) and calculated (lines) travel time for all OBS. (b) Ray paths for calculated
travel times. Travel time modeling was carried out by two-point ray tracing using Zelt and Smith’s [1992]
linearized inversion program. RMS misfit between the calculated and observed travel times was 0.127 s.
A much better fit was obtained for the sedimentary section, 0.1 s, because of better ray coverage.
ETG 2 - 6 TEN BRINK AND TAYLOR: CRUSTAL STRUCTURE OF BAIKAL RIFT
(Figure 4). The subsedimentary continental crust in our
model consists of three layers, with velocities ranging from
5.6 to 7.2 km/s (Figure 7a). Variations in upper crustal
velocity gradient along the profile are reflected in the
thickness of the top layer and show the highest gradient
to be located under the Selenga Delta. Upper crust is
defined by velocity values ranging from 5.6 to 6.3 km/s.
P wave velocities for the midcrustal layer range from 6.4 to
6.9 km/s. The lower crustal layer has a uniform thickness of

8 km, beginning at a depth of 
33 km with an average P
wave velocity of 7.2 km/s (Figure 7a). The apparent high-
velocity layer occurs throughout the entire model and is
associated with wide-angle reflections (P3P) from its top
surface (Figure 4). Average depth to Moho is 
40 km, with
small wavelength perturbations incorporated into the model
to fit the observed travel times. PmP (wide-angle reflection
from the Moho) and Pn (refraction from the uppermost
mantle) arrivals are only constrained in the central portion
of the model at 
75–155 km. Both arrivals display reci-
procity in OBS records from the southern and northern
extents of the survey (Figure 5). Pn velocity is 8.0 km/s.
[15] To test the accuracy of our model, we perturbed the
best fitting model by changing Moho depth by 1-km incre-
ments and by changing the basal layer velocity by 1 km/s
increments and keeping a constant velocity gradient within
the basal layer. The variances of these perturbations were
computed (Figure 6). Next, we estimated the ranges of
velocities and depths that satisfy the seismic observations.
Under the assumption of Gaussian distribution of the
residual travel time we can use the F test. If the difference
in RMS value between the best fitting model and the
perturbed model is 0.01 s, then the perturbed model is
different at the 99% confidence level. For the subset of rays
that penetrate the lower crust a difference in RMS value
0.015 s (i.e., an RMS value of 0.142 s), implies that the
perturbed model is different at a 99% confidence level from
the best fitting model. Hence we estimate the range of lower
crustal velocities at 7.05–7.4 km/s, and the range of Moho
depths is estimated at 39–42.5 km. We cannot confidently
assign error bounds to upper mantle velocity due to the
small number of Pn arrivals.
[16] Our lower crustal velocity and Moho depth are
significantly different from previously published wide-
angle seismic reflection and refraction results [Krylov,
1981; M. M. Mandelbaum et al., Deep structure of the
region, unpublished manuscript, 2001], which were based
on data from sparse shots and receivers, none of which
were located within the lake. Their models probably did
not incorporate near-surface and shallow structural con-
straints because of lack of coincident MCS data. Accord-
ing to Krylov [1981] and M. M. Mandelbaum et al. (Deep
structure of the region, unpublished manuscript, 2001)
Moho is 34–38 km deep beneath the Central Basin,
increasing abruptly to 42–44 km in the Northern and
Barguzin Basins. It is 36–37 km deep beneath the Selenga
Basin. They report relatively low basal velocity of the
crust (6.4 km/s) under all the basins and a relatively low-
velocity upper mantle layer (7.6–7.8 km/s) that extends
under the entire length of Lake Baikal and for another
100–150 km southeastward underneath Transbaikal. An
additional interface was placed at depths of 55–60 km
where velocity increased to 8.2–8.3 km/s. Error bounds of
previous measurements in the Baikal rift zone were
estimated as ±2 km for Moho depth and ±0.2–0.3 km/s
for upper mantle velocity (M. M. Mandelbaum et al., Deep
structure of the region, unpublished manuscript, 2001).
Krylov’s [1981] results were used to argue for an active
upwelling driving force for Baikal rifting [e.g., Zorin et al.,
1989]. However, considering the error bars on the above
velocities and using our own results for Pn velocities, it is
unclear whether an upper mantle velocity anomaly exists
under Lake Baikal.
Figure 6. Perturbations to the best fitting model shown in
Figure 7. Moho depth was varied in increments of 1 km,
and the average velocity of the basal layer was varied in
increments of 0.1 km/s, with a constant velocity difference
of 1 km/s between the top and bottom of the layer. The
purpose of these perturbations was to explore the range of
permissible solutions. Using the F test as a guide, we
determine a velocity range (shown by arrows) of 7.05–
7.4 km/s and a depth range to Moho of 39–42.5 km.
TEN BRINK AND TAYLOR: CRUSTAL STRUCTURE OF BAIKAL RIFT ETG 2 - 7
[17] Our Moho depth and Pn velocity are also different
from the Gao et al. [1994] interpretation of teleseismic P
wave travel time delays observed in a profile across the
Central Basin. Assuming that the delay arises from structure
on the lithosphere-asthenosphere boundary, they concluded
that the asthenosphere reaches a level as shallow as 34 km
under the Central Basin and has a velocity contrast of 5%
with the lithosphere (i.e., a velocity of 7.7 km/s). The Gao et
al. [1994] interpretation was challenged by Petit et al.
[1998], who proposed a different distribution of travel time
delays within the upper mantle.
5. Discussion
5.1. Rifting Mechanisms
[18] Our seismic velocity structure for the Central Basin
(Figure 7) is characterized by (1) a Moho depth of 39–
42.5 km, (2) an 8-km-thick, laterally continuous high-
Figure 7. (a) P wave velocity model along the Central Baikal Basin (gray line in Figure 3). (b) An
enlargement of the upper part of the model. Red circles indicate locations of ocean bottom seismometers.
ETG 2 - 8 TEN BRINK AND TAYLOR: CRUSTAL STRUCTURE OF BAIKAL RIFT
velocity (7.05–7.4 km/s) lower crust, (3) a sedimentary
section reaching maximum depths of 9 km, and (4) sedi-
ment velocity increasing gradually with depth, probably
with a normal Pn velocity of 
8 km/s.
[19] The high-velocity basal crustal layer can be inter-
preted in two ways, either as rift-related magmatic addition
to the base of the crust or as relic prerift continental crust.
Travel time delays of teleseismic data indicate that the
asthenosphere under the rift is not much deeper than the
base of the crust [Gao et al., 1994; Rogozhina and Koz-
hevnikov, 1979], although later work [Petit et al., 1998] and
our Pn velocity questioned this result. A shallow astheno-
sphere could be a potential source for magmatic addition to
the base of the crust in the form of dikes [Zorin et al., 1989]
or sills. A basal high-velocity layer was found under other
rifts, for example, under the Oslo Graben, which is a narrow
Middle to Late Paleozoic rift within Precambrian crust
[Newmann et al., 1995] similar to Lake Baikal. Unlike
Lake Baikal, however, the basal layer is associated with
extensive basaltic to rhyolitic volcanoes and shallow intru-
sions in the form of dikes and sills, which were emplaced
throughout the rifting history [Newmann et al., 1995]. In
addition, the crust under the Oslo Basin was thinned
considerably with a Moho depth at 32 km and the top of
the basal layer at 20 km depth [Newmann et al., 1995].
There is no evidence for Cenozoic volcanism in and around
the Central Basin or in other parts of Lake Baikal [Kiselev,
1987].
[20] Volcanism in the Baikal rift system is, in fact, rarely
associated with the basins themselves. Volcanic rocks at the
Sayan-Khamar Daban area west of Lake Baikal are found
mostly outside the basins and extend hundreds of kilometers
south into Mongolia (Figure 1) [Kiselev and Popov, 1992;
Windley and Allen, 1993]. Volcanism east of Lake Baikal is
confined to the Vitim Plateau and Udokan Range, elevated
areas south of the rift system. The largest volume of
volcanism erupted during the Miocene, although volcanism
continued into the Pleistocene [Kiselev and Popov, 1992].
Mantle xenoliths in young (1–4 Ma) volcanic rocks SW of
Lake Baikal have chemical composition and texture similar
Figure 8. Seismic reflection lines (a) 92-13 and (b) 92-11 along the Central Basin. Line 92-13 is
coincident with part of the refraction line, but line 92-11 is not. See Figure 3 for location. The deeper
sedimentary section on line 92-13 is nonreflective for reasons that are not entirely clear, but not because
of the water multiple [Lee et al., 1996]. Basement is deeper than 8 s under the deepest part of the basin
and rises toward Barguzin Bay and under the Selenga Delta, as indicated by the velocity model. The
hummocky character of the reflectors at km 105–132 is likely due to the presence of free gas in the
sediments. Reversed polarity of the top reflector and severe attenuation of the refraction records of OBS
B1 and AA3 at this location (Figure 5) support this interpretation. A bottom-simulating reflector extends
northeastward along the line and is associated with normal polarity. B4 is stratigraphic horizon identified
by Moore et al. [1997].
TEN BRINK AND TAYLOR: CRUSTAL STRUCTURE OF BAIKAL RIFT ETG 2 - 9
to xenoliths in older (5–7 Ma) eruptions. Their equilibrium
temperatures do not indicate progressive shallowing of the
level of magma generation [Ionov et al., 1995; Kiselev and
Popov, 1992]. These temperatures are not particularly high,
arguing against anomalous mantle temperatures or partial
melt beneath the rift zone [Ionov et al., 1995]. We therefore
do not interpret the basal high-velocity layer as representing
magmatic addition to the lower crust during rifting.
[21] The alternative interpretation for the high-velocity
basal layer is that it is a relic of the prerifted crust. The
velocity structure of the Siberian Platform derived from
Russian transcontinental refraction lines is characterized by
a P wave velocity of 7.2–7.4 km/s in the lower 10 km of the
crust and a Moho depth of 40–42 km [Pavlenkova, 1996].
Logatchev and Zorin [1992] hypothesized the existence of
mafic lower crust in the Baikal region, which consists of the
differentiation products of large Precambrian and lower
Paleozoic granitic bodies. World-wide compilation of crus-
tal velocity structure shows that a basal layer velocity of 7–
7.3 km/s is typical of Proterozoic shields and platforms and
of continental arcs [Christensen and Mooney, 1995]. In
particular, the average layer thickness of Proterozoic shields
and platforms is 7 km, and the average Moho depth is
41.5 km [Christensen and Mooney, 1995], although varia-
tions from the average can be large [Durrheim and Mooney,
1991]. We therefore propose that the lower part of the crust
under Lake Baikal is a remnant of the original crust of the
Siberian Platform or of continental arcs within the Sayan-
Baikal belt.
[22] If we accept this interpretation, then rifting of
Central Baikal Basin must be confined to the upper and
middle crust, whereas the lower crust and Moho are not
modified to a degree that can be detected seismically
(Figure 10). One way to achieve this style of rifting is
by extension along a normal fault, which flattens into a
midcrustal detachment. Krylov et al. [1993] present a line
drawing from a deep seismic reflection profile along the
Upper Angara Basin (north of Lake Baikal) that shows a
band of reflectivity extending from the upper crust to a
depth of 18–22 km where it flattens. The interpretation of
Baikal rift basins as narrow rifts rooted into midcrustal
detachments (Figure 10) is also attractive given that the
basins occupy the Sayan-Baikal fold-and-thrust belt at its
boundary with the Archean-Proterozoic crust of the Sibe-
rian Platform (Figure 2) [e.g., Zamarayev and Ruzhich,
1978].
Figure 9. A portion of seismic reflection line 92-15 along
the Academic Ridge in the vicinity of boreholes BDP-98.
See Figure 3 for location. B4 is the stratigraphic horizon
identified by Moore et al. [1997], which is close to the
bottom of the hole. Age of sediments at the bottom of the
hole is identified at 11 Ma (J. King, personal communica-
tion, 2000).
Figure 10. Conceptual model for the extension of the Central Basin. Rifting of the Central Basin is
confined to the upper and middle crust and probably reactivates one of the faults within the fold and
thrust belt, which was accreted during the Late Proterozoic and the Paleozoic onto the Siberian Platform.
The lower crust and Moho are not modified to an extent that can be detected seismically. This style of
rifting can be achieved by extension on faults soling into a midcrustal detachment. Krylov et al. [1993]
presents a line drawing from a deep seismic reflection line in the upper Angara region showing a band of
reflectivity extending from the upper crust and flattening at depths of 18–22 km. The locus of upper
mantle extension is possibly shifted to the southeast where high topography, volcanism, and, possibly, a
thin astenosphere are observed.
ETG 2 - 10 TEN BRINK AND TAYLOR: CRUSTAL STRUCTURE OF BAIKAL RIFT
[23] Poort et al. [1998] and van der Beek [1997] pro-
posed a detachment model to explain the topography and
gravity fields across the Central Basin. They used time-
dependent thermomechanical models in which extension
takes place along a listric normal fault that flattens at a
depth of 20 km and by pure shear in the lower crust and
upper mantle. Their models predicted the Moho under the
rift to shallow by 3 km, which is within our error estimates.
The required thinning factor in their model is small (b =
1.3), and therefore the expected thermal and mechanical
perturbations are small. They and others [e.g., Ruppel et al.,
1993] estimated the elastic thickness of the lithosphere to be
large (30–50 km) implying a rigid cold lithosphere. The
large depth of some earthquakes, 25–30 km, also implies a
cold crust [Deverche`re et al., 1993]. Our interpretation of a
Siberian Platform crust underlying the rift is consistent with
these observations, although we cannot determine whether
the faults are listric and whether upper mantle extension
occurs under the lake.
[24] The van der Beek [1997] model assumes that upper
mantle extension occurs under the rift. Alternatively, the
locus of upper mantle extension could be offset from the
rift (Figure 10), as was proposed for the Atlantic margin of
North America [Dunbar and Sawyer, 1989; Braun and
Beaumont, 1989; Harry and Sawyer, 1992]. Harry and
Sawyer [1992] suggested that upper crustal extension
during the first 50 m.y. of rifting resulted from reactivation
of the thin-skinned Appalachian fold-and-thrust belt over a
cold lithosphere, while upper mantle extension took place
200–300 km seaward. During the 10 m.y. prior to the
initiation of seafloor spreading in the Atlantic Ocean,
crustal extension shifted to the location of upper mantle
extension as asthenospheric upwelling thermally weakened
the overlying crust.
[25] The Baikal rift system is similar in many respects to
the Triassic-Jurassic Atlantic rift system.
1. The Atlantic rift system followed the sinuous
Paleozoic suture between the African and North American
continents far from active plate boundaries at that time
(Figure 11) [e.g., Manspeizer and Cousminer, 1988]. This
distribution is reminiscent of the distribution of Baikal rift
basins along the sinuous Sayan-Baikal fold-and-thrust
belt.
2. The Atlantic rift system consists of numerous rift
basins with a mixture of shapes and sizes. They are not
arranged along a narrow axis (i.e., the rift valley) and some
basins are even sub-parallel to one other within the suture
zone. A similar mixture of shapes and size of basins and a
similar non-axial arrangement characterize the Baikal rift
system. Barguzin Basin is subparallel to the North Basin as
are some of the smaller basins north of Lake Baikal. In both
rift systems the location of the basins probably follows
upper crustal weaknesses in the form of pre-existing thrust
sheets within the Paleozoic fold-and-thrust belt.
3. The Newark basin (Figure 11), a large (220 by 50 km)
and deep (>7.5 km) basin within the Atlantic rift system,
was connected to the Gettysburg and Culpepper Basins to
the south via narrow necks creating an elongated lake much
like Lake Baikal [Olsen et al., 1996]. All three basins have
their major border fault on the NW side.
4. The first 25–30 m.y. of rifting in the Newark Basin
were nonvolcanic [Manspeizer and Cousminer, 1988].
5. Deep seismic reflection and refraction profiles under
other Triassic-Jurassic basins of the Atlantic continental
margin show that Moho depth is not elevated under those
basins [Johnson et al., 1998; Nelson et al., 1985].
6. A pervasive hydrothermal convection system circu-
lated relatively high-temperature fluids within the basin
during rifting despite the lack of a shallow mantle source
[Steckler et al., 1993]. High geothermal gradients in Lake
Baikal initially interpreted to reflect shallow mantle
sources [Zorin and Lepina, 1985] can also be explained
by hydrothermal circulation of meteoric water [Golubev,
1990].
[26] The architecture of the Baikal rift system is, in fact,
different from many other continental rift systems to which
it had been previously compared [Kazmin, 1991; Lipman
et al., 1989; Lysak, 1987; Zorin and Lepina, 1985]. Basins
within the East Africa rift system, the Ethiopian Rift, the
Rio Grande Rift and the Gulf of Suez are arranged along a
narrow long axis (the rift valley). The differences between
the Baikal rift system and the above rift systems are
perhaps related to the location of upper mantle extension
and thermal perturbation relative to the location of the
crustal rift in light of suggestions for mantle upwelling
southeast [Kazmin, 1991; Kiselev and Popov, 1992; Wind-
ley and Allen, 1993] or northwest [Petit et al., 1998] of
Baikal rift zone. Alternatively, the differences may be
attributed to the alignment of the former compression
and present extension directions [e.g., Zonenshain et al.,
1990], which enables extension by fault reactivation across
the width of the fold-and-thrust belt. In summary, we
believe that a thin-skinned extension model for the upper
crust with an offset separating the loci of upper crust and
upper mantle extension may apply to the development of
the Baikal rift.
[27] Other mechanical models have recently been sug-
gested to explain the gravity and topography fields across
Lake Baikal and, in particular, its uplifted margins. Petit et
al. [1997] modeled the gravity field by a rigid plate with
discontinuities at the NW and SE sides of the Central Basin.
The Siberian and Transbaikal plates in their model are
downwarped toward the rift, and the crust beneath the
uplifted margins is thickened. The basin is thereby effec-
tively in local isostatic equilibrium, and Moho under the rift
is predicted to warp up by 7 km relative to the Siberian
Platform [Petit et al., 1997]. The difference between the
Moho depth of our model and of the Siberian Platform
[Pavlenkova, 1996] is only 3 km or less. Moreover, there is
no geological evidence for a lithospheric-scale discontinuity
SE of the lake [Petit et al., 1997].
[28] Lesne et al. [2000] fit the basin subsidence and the
surrounding uplifted shoulders by a mechanical finite ele-
ment model with layered rheology, in which extension takes
place on a normal fault cutting the entire lithosphere. Their
model generates the observed subsidence and uplift in only
2.5–3.5 m.y, which they attribute to the ‘‘fast rifting stage.’’
In section 5.2 we argue that the Central Basin and probably
also the Northern Basin are much older than 3 Ma and that
there is no clear evidence for a fast rifting stage starting at
3 Ma. In addition, their model generates a secondary basin
SE of the main basin, which they interpret to represent the
Barguzin Basin. However, the bounding fault in their model
dips in the opposite direction to the bounding fault of the
TEN BRINK AND TAYLOR: CRUSTAL STRUCTURE OF BAIKAL RIFT ETG 2 - 11
Barguzin Basin. There is no secondary basin SE of the
Central Basin.
5.2. Evolution of the Central Basin
[29] Rifting of Baikal basins probably began during the
middle Oligocene (
35 Ma), although earlier (Paleocene to
early Eocene) lacustrine and alluvial sediments were found
at the bottom of boreholes in the Selenga Delta [Mats,
1993]. Where exposed or drilled, the Oligocene to Pliocene
sediments often comprise alternating beds of clay, siltstones,
and sand deposited in a lacustrine environment [Mats,
1993]. Rifting is generally considered to have intensified
since the middle Pliocene (
3 Ma) because Pliocene and
younger sediments consist of coarser sands, argillites, and
silts, indicating the intensification of the topographic relief,
and because of a perceived increase in sedimentation rate
[Logatchev and Zorin, 1987; Delvaux et al., 1997; Kazmin
et al., 1995; Mats, 1993]. Here we argue against the
existence of rift-wide stages and for episodic subsidence
and enlargement of individual subbasins, as is observed in
other rift basins [e.g., Ebinger et al., 1989].
[30] Borehole BDP96 on Academic Ridge (Figure 3)
yielded an age of 5 Ma at a depth of 192 m subbottom
based on a magnetic polarity stratigraphy and a nearly
constant sedimentation rate of 4 cm/kyr without major
hiatuses or disconformities [Williams et al., 1997]. The
composition and texture of sediments are remarkably uni-
form throughout the borehole, indicating that the present
hemipelagic lacustrine facies has persisted during the past
5 m.y. on Academic Ridge [Kuzmin et al., 2000]. An
adjacent hole, BDP98, yielded a paleomagnetic age of
11 Ma at a depth of 650 m subbottom and did not appear
Figure 11. (a, b) Location of Atlantic rift system (black) within Pangea at 
210 Ma and the location of
surface and subsurface basins of the Atlantic rift system along the Atlantic coast of the United States and
Canada [after Olsen et al., 1996]. (c, d) Simplified geological map view and vertical cross sections of
Newark Basin [after Schlische, 1992]. P and L are shallow and deepwater lacustrine strata, respectively,
and S is fluvial sedimentary strata.
ETG 2 - 12 TEN BRINK AND TAYLOR: CRUSTAL STRUCTURE OF BAIKAL RIFT
to have major hiatuses (J. King, personal communication,
2000). This depth is equivalent to 795 ms two-way travel
time (TWTT), using an average velocity of 1.67 km/s (C.
Scholz, personal communication, 2000). An MCS profile
adjacent to these boreholes (Figure 9) shows reflective
sedimentary section down to 
800 ms subbottom, where
a horizon, labeled B4 by Moore et al. [1997], is observed.
The underlying 420 ms to basement in this profile lack
continuous reflections. A horizon interpreted to be equiv-
alent to B4 in the Central Basin was identified close to the
water multiple at TWTT of 4–4.5 s (Figure 8) [Moore et al.,
1997], which corresponds to a depths of 2.4–2.9 km
subbottom assuming average velocities of 2.25–2.5 km/s
(Figure 7b).
[31] Hutchinson et al. [1992] interpreted the reflective
sedimentary section above 4 s TWTT to have been depos-
ited since middle Pliocene. If the stratigraphic correlation
between the Academic Ridge and the Central Basin is
correct, then the reflective upper sedimentary unit in the
Central Basin was deposited over much longer time period
than previously assumed [Hutchinson et al., 1992], at least
since 11 Ma. In addition, changes in reflectivity with depth
were shown in later work to be gradual and explained by
diagenesis [Lee et al., 1996]. Our wide-angle reflection data
show no evidence for a regional reflecting horizon within
the sedimentary section.
[32] Assuming an age of 11 Ma at this depth, the average
sedimentation rate for the northern part of the Central Basin
(uncorrected for compaction) is 22–26 cm/kyr. Slow con-
tinuous deposition is also observed in the Newark Basin,
where 6700 m of lacustrine sediments were deposited over
24 m.y. [Olsen et al., 1996] an average of 28 cm/kyr. On
Academic Ridge the average sedimentation rate for the past
11 m.y. is 1.5 times faster than the average rate for the past
5 m.y., and sedimentation rate for the past 5 m.y. has been
almost uniform [Williams et al., 1997]. Sedimentation rate in
the Selenga Delta during the Holocene and during past
650 kyr is similar, 15–20 cm/kyr [Scholz and Hutchinson,
2000; Colman et al., 1996]. Sedimentation rate in the
southwestern part of the Northern Basin during the last
2.5 m.y. was estimated to be much higher, 60–70 cm/kyr
[Kuzmin et al., 2000] based on the correlation of a prominent
unconformity there with a reflector near borehole BDP96
(Horizon B-10 [Moore et al., 1997]). However, this corre-
lation was based solely on the assumption that prograding
clinoforms in these two areas, which are 50 km apart, have
the same age [Moore et al., 1997]. More generally, equating
changes in sedimentation rate with changes in subsidence
rate, can be problematic, because sedimentation rate depends
on sediment supply, which is spatially and temporaly vari-
able within the lake environment.
[33] We also see no clear indication for a fast rifting stage
starting 3 m.y. ago in the initiation of subsidence of new
basins. The age of the Northern Basin is estimated to be 6.6
Ma [San’kov et al., 2000], late Miocene [Kazmin et al.,
1995], 14 Ma (S. M. Colman et al., Quaternary depositional
patterns and environments in Lake Baikal from high-reso-
lution seismic stratigraphy and coring, unpublished manu-
script, 2001), or the same as the age of the Central Basin
[Moore et al., 1997]. The estimated age of different basins
north of Lake Baikal ranges between 1 and 7 Ma [San’kov
et al., 2000]. More generally, equating changes in subsi-
dence rate with changes in extension rate can also be
problematic.
[34] Growth of normal faults by linking isolated fault
segments and the cessation of movement on other fault in
stress shadow zones of the newly linked faults can lead to
episodes of rapid local subsidence and to the enlargement
and coalescence of existing basins [Gupta et al., 1998]. This
can lead to changes in sediment transport paths. Indeed,
where documented, increasing subsidence rate has been
episodic and has not been coeval even across short dis-
tances. For example, an episode of rapid subsidence started
650 kyr ago in the Selenga Delta (S. M. Colman et al.,
Quaternary depositional patterns and environments in Lake
Baikal from high-resolution seismic stratigraphy and coring,
unpublished manuscript, 2001) and rapid subsidence of the
Selenga Basin occurred 1.7–1.4 m.y. ago [Scholz and
Hutchinson, 2000]. Both locations are underlain by thick
(4–8 km) sedimentary sections, which have been subsiding
long before the late Pleistocene. Seismic stratigraphy
(Figure 8) [Moore et al., 1997] indicates that the Central
Basin grew northward with time, probably during the Late
Miocene [Kazmin et al., 1995; Mats, 1993]. In our velocity
model, high-velocity sediments are confined to the deep
portion of the basin, whereas lower-velocity sediments
extend over the entire basin length (Figure 7b). Agar and
Klitgord [1995] documented a southward propagation of
Primorskiy Fault, probably within the last 1 m.y. An
embayment, Maloy More, and a subaerial graben, which
connect to the Northern Basin, were formed in the wake of
this propagation [Agar and Klitgord, 1995].
6. Conclusions
[35] Results from a wide-angle seismic reflection and
refraction profile along the Central Basin of Baikal rift
show that the base of the sedimentary section reaches a
depth of 8–9 km beneath the Central Basin and shallows
toward Barguzin Bay and under the Selenga Delta. The
crust includes an 8-km-thick basal high-velocity (
7.2–
7.3 km/s) layer. The Moho is 
40 km deep, and upper
mantle velocity is probably normal. We believe that the
high-velocity basal layer does not represent rift-related
magmatic addition to the lower crust because there is no
geological evidence for volcanism in the Central Basin, in
contrast to other rifts with a magmatic layer (Oslo Graben,
West Antarctica). In fact, Baikal rift volcanism occurs
mainly outside the rift basins, often in areas of high top-
ography, and it extends hundreds of kilometers to the south.
In addition, the largest volume of Baikal volcanism
occurred during the Miocene, and xenoliths do not show
progressive shallowing of the melt region with time. The
inferred mantle temperature is not particularly high, and the
elastic thickness of the lithosphere is indicative of a cold
lithosphere. We interpret the basal layer as a relic of the
prerift crust, most likely a relic of the Siberian Platform.
Crustal structure of the Siberian Platform is characterized by
a 10-km-thick basal layer with velocities of 7.2–7.4 km/s
and a Moho depth of 42 km. World-wide compilations of P
wave velocities in continental crust shows that such a high-
velocity basal layer is typical of shields and platforms.
[36] This interpretation implies that rifting of the Central
Basin is confined to the upper and middle crust, whereas
TEN BRINK AND TAYLOR: CRUSTAL STRUCTURE OF BAIKAL RIFT ETG 2 - 13
the lower crust is not modified to an extent that can be
detected seismically. This style of rifting can be achieved by
normal faults soling into a midcrustal detachment. Krylov
et al. [1993] present a line drawing from a deep seismic
reflection line in the upper Angara region showing a band
of reflectivity extending from the upper crust and flattening
at depths of 18–22 km, which may emanate from this
detachment.
[37] The locus of upper mantle extension is possibly
shifted to the southeast of Baikal rift zone where high
topography, volcanism, and, according to some, a thin
asthenosphere are observed. A similar rifting model was
proposed for the Triassic-Jurassic rift system along the
Atlantic margin of North America. In this model, upper
crustal extension utilized preexisting zones of weakness,
whereas upper mantle extension occurred 200–300 km
seaward. Many aspects of the geometry, distribution, and
crustal structure of the Atlantic rift system are similar to the
Baikal rift system.
[38] We question the common division of Baikal rift
evolution into a fast rifting stage starting 
3 Ma and an
earlier slower phase. Our seismic refraction model shows
that sediment velocity increases gradually with depth and
that there are no regional wide-angle reflections within the
sediments to support an abrupt and major change in
sedimentation. In addition, using the stratigraphic correla-
tions of Moore et al. [1997], we suggest that the upper 2.4–
2.9 km of sediments in the basin was deposited over at least
11 m.y., a much slower rate than previously assumed. We
propose instead that rift evolution may be characterized by
propagation and coalescence of boundary faults leading to
local episodes of subsidence and to areal enlargement of
basins.
[39] Acknowledgments. This project was jointly funded by the U.S.
Geological Survey Coastal and Marine Program and the Russian Academy
of Sciences. Debbie Hutchinson and Alexander Golmshtok coordinated the
project. Logistical support was provided by Institutes of Limnology and of
the Earth Crust in Irkutsk, the Listviyanka field office, and the Shirshov
Institute of Oceanology in Gelenzhik. We thank Greg Miller, Alik Badar-
dinov, Dwight Coleman, Mark Behrendt, and Captain A. Sakhalov and the
crew of the R/V Titov for their dedicated field support of the OBS
operation. We thank Alexander Golmshtok, Kim Klitgord, Chris Scholz,
Liyosha Akentiev, and Dave Nichols for coordinating the shooting. Dis-
cussions with Steve Colman, Chris Scholz, John King, Kim Klitgord, and
Debbie Hutchinson and helpful reviews by Steve Colman, Debbie Hutch-
inson, Cindy Ebinger, Jacques Deverche`re, and an anonymous reviewer are
gratefully acknowledged. Phil Molzer provided computer and graphical
support.
References
Agar, S. M., and K. K. Klitgord, Rift flank segmentation, basin initiation
and propagation: A neotectonic example of Lake Baikal, J. Geol. Soc.
London, 152, 849–860, 1995.
Agena, W. F., M. W. Lee, J. J. Miller, and D. R. Hutchinson, Lake Baikal,
1992: Processing of multichannel seismic reflection data, U.S. Geol. Surv.
Open File Rep., 94-0263, 37 pp., 1994.
Braun, J., and C. Beaumont, Dynamical models of the role of crustal shear
zones in asymmetric continental extension, Earth Planet. Sci. Lett., 93,
405–423, 1989.
Burov, E. B., F. Houdry, M. Diament, and J. Deverchere, A broken plate
beneath the Baikal rift zone revealed by gravity modelling, Geophys. Res.
Lett., 21, 129–132, 1994.
Calais, E., and S. Amarjargal, New constraints on current deformation in
Asia from continuous GPS measurements at Ulan Baatar, Mongolia,
Geophys. Res. Lett., 27, 1527–1530, 2000.
Calais, E., O. Lesne, J. Deverchere, V. San’kov, A. Lukhnev, A. Mirosh-
nitchenko, V. Buddo, K. Levi, V. Zalutzky, and Y. Bashkuev, Crustal
deformation in the Baikal rift from GPS measurements, Geophys. Res.
Lett., 25, 4003–4006, 1998.
Christensen, N. I., and W. D. Mooney, Seismic velocity structure and
composition of the continental crust: A global view, J. Geophys. Res.,
100, 9761–9788, 1995.
Colman, S. M., G. A. Jones, M. Rubin, J. W. King, J. A. Peck, and W. H.
Orem, AMS radiocarbon analyses from Lake Baikal, Siberia: Challenges
of dating sediments from a large oligotrophic lake, Quat. Sci. Rev., 15,
669–684, 1996.
Delvaux, D., R. Moeys, G. Stapel, C. Petit, K. Levi, A. Miroshnichenko,
V. Ruzhnich, and V. San’kov, Paleostress reconstructions and geody-
namics of the Baikal region, central Asia, part 2, Cenozoic rifting, Tec-
tonophysics, 282, 1–38, 1997.
Deverche`re, J., F. Houdry, N. V Solonenko, A. V. Solonenko, and V. S.
Sankov, Seismicity, active faults and stress field of the North Muya
region, Baikal rift: New insights on the rheology of extended continental
lithosphere, J. Geophys. Res., 98, 19,895–19,912, 1993.
Doser, D. I., Faulting within the eastern Baikal rift as characterized by
earthquake studies, Tectonophysics, 196, 109–139, 1991.
Dunbar, J. A., and D. S. Sawyer, How preexisting weaknesses control the
style of continental breakup, J. Geophys. Res., 94, 7278–7292, 1989.
Durrheim, R. J., and W. D. Mooney, Archean and Proterozoic crustal evo-
lution: Evidence from crustal seismology, Geology, 19, 606–609, 1991.
Ebinger, C. J., A. L. Deino, R. E. Drake, and A. L. Tesha, Chronology of
volcanism and rift basin propagation: Rungwe volcanic province, East
Africa, J. Geophys. Res., 94, 15,785–15,803, 1989.
Gao, S., P. M. Davis, H. Liu, P. D. Slack, Y. A. Zorin, N. A. Logatchev,
M. G. Kogan, P. D. Burkholder, and R. P. Meyer, Asymmetric upwarp of
the asthenosphere beneath the Baikal rift zone, Siberia, J. Geophys. Res.,
99, 15,319–15,330, 1994.
Golubev, V.A., Hydrothermal systems, permeability and thermal models
of the crust in the Baikal rift zone, Izv. Earth Phys., 26, 916–924,
1990.
Gupta, S., P. A. Cowie, N. H. Dawers, and J. R. Underhill, A mechanism to
explain rift-basin subsidence and stratigraphic patterns through fault-
array evolution, Geology, 26, 595–598, 1998.
Harry, D. L., and D. S. Sawyer, A dynamic model of lithospheric extension
in the Baltimore Canyon Trough region, Tectonics, 11, 420–436, 1992.
Holbrook, W. S., and P. B. Kelemen, Large igneous province on the U.S.
Atlantic margin and implications for magmatism during continental
breakup, Nature, 364, 433–436, 1993.
Holbrook, W. S., T. M. Brocher, U. S. ten Brink, and J.A. Hole, Crustal
structure of a transform plate boundary: San Francisco Bay and the
central California continental margin, J. Geophy. Res., 101, 22,311–
22,334, 1996.
Hutchinson, D. R., A. J. Golmshtok, L. P. Zonenshain, T. C. Moore, C. A.
Scholz, and K. D. Klitgord, Depositional and tectonic framework of the
rift basins of Lake Baikal from multichannel seismic data, Geology, 20,
589–592, 1992.
Ionov, D. A., S. Y. O’Reilly, and I. V. Ashchepkov, Feldspar-bearing lher-
zolite xenoliths in alkali basalts from Hamar-Daban, southern Baikal,
Russia, Contrib. Mineral. Petrol., 122, 174–190, 1995.
Johnson, T. L., B. E. Wright, D. B. Stewart, and F. N. Zihlman, Global
Geoscience Transect 8: Quebec-Maine-Gulf of Maine Transect, south-
eastern Canada, northern U.S.A., U.S. Geol. Surv. Misc. Invest. Ser.
Map, I-2329, 1998.
Kazmin, V. G., The position of continental flood basalts in rift zones and its
bearing on models of rifting, Tectonophysics, 199, 375–387, 1991.
Kazmin, V. G., A. Y. Golmshtok, K. Klitgord, T. Moore, D. Hutchinson,
C. Scholz, and E. Weber, Structure and development of the Akademi-
chesky Ridge area (Baikal rift) according to seismic investigations, Russ.
Geol. Geophys., 36, 155–167, 1995.
Keller, G. R., M. H. P. Bott, R. F. Wendlandt, D. I. Doser, and P. Morgan,
The Baikal rift system, in Continental Rifts: Evolution, Structure, Tec-
tonics, edited by K. H. Olson, pp. 325–339, Elsevier Sci., New York,
1995.
Kiselev, A. I., Volcanism of the Baikal rift zone, Tectonophysics, 143, 235–
244, 1987.
Kiselev, A. I., and A. M. Popov, Asthenospheric diapir beneath Baikal rift:
Petrological constraints, Tectonophysics, 208, 287–295, 1992.
Krylov, S. V., Nedra Bajkala, the Baikal Interior From Seismic Data, 105
pp., Trudy Inst. Geol. i Geofiz., Novosibirsk, Russia, 1981.
Krylov, S. V., Z. R. Mishenkina, Y. V. Kulchinskii, E. N. Ten, and I. F.
Sheludko, Characteristics of seismoactive lithosphere for the northeast of
Baikal region from data of detailed studies by DSS technique on P- and
S- waves, Russ. Geol. Geophys., 34, 110–119, 1993.
Kuzmin, M. I., E. B. Karabanov, A. A. Prokopenko, V. F. Gelety, V. S.
Antipin, D. F. Williams, and A. N. Gvozdkov, Sedimentation processes
and new age constraints on rifting stages in Lake Baikal: Results of deep
water drilling, Int. J. Earth Sci., 89, 183–192, 2000.
ETG 2 - 14 TEN BRINK AND TAYLOR: CRUSTAL STRUCTURE OF BAIKAL RIFT
Lee, M. W., W. F. Agena, and D. R. Hutchinson, Amplitude blanking in
seismic profiles from Lake Baikal, Mar. Pet. Geol., 13, 549–563, 1996.
Lesne, O., E. Calais, and J. Deverchere, Finite element modelling of crustal
deformation in the Baikal rift zone: New insights into the active-passive
debate, Tectonophysics, 289, 327–340, 1998.
Lesne, O., E. Calais, J. Deverchere, J. Che’ry, and R. Hassani, Dynamics of
intracontinental extension in the north Baikal rift from two-dimensional
numerical deformation modeling, J. Geophys. Res., 105, 21,727–21,744,
2000.
Lipman, P. W., N. A. Logatchev, Y. A. Zorin, E. Chapin, V. Kovalenko, and
P. Morgan, Intracontinental rift comparisons: Baikal and Rio Grande Rift
systems, Eos Trans. AGU, 70, 578–579, 586–587, 1989.
Logatchev, N. A., and Y. A. Zorin, Evidence and causes of the two-stage
development of the Baikal rift, Tectonophysics, 143, 225–234, 1987.
Logatchev, N. A., and Y. A. Zorin, Baikal rift zone: Structure and geody-
namics, Tectonophysics, 208, 273–286, 1992.
Loss, J., I. A. Pecher, and U. S. ten Brink, RayGUI: A graphical user
interface for interactive ray-tracing (rayinvr), U.S. Geol. Surv. Open File
Rep., 98–203, 1998.
Lysak, S. V., Terrestrial heat flow of continental rifts, Tectonophysics, 143,
31–41, 1987.
Manspeizer, W., and H. L. Cousminer, Late Triassic-Early Jurassic synrift
basins of the U.S. Atlantic margin, in The Geology of North America,
Vol. I-2, The Atlantic Continental Margin, U.S., edited by R. E. Sheridan
and J. A. Grow, pp. 197–216, Geol. Soc. of Am., Boulder, Colo., 1988.
Mats, V., The structure and development of Baikal rift depression, Earth
Sci. Rev., 34, 81–118, 1993.
Moore, T. C., Jr., K. D. Klitgord, A. J. Golmshtok, and E. T. Weber,
Sedimentation and subsidence patterns in the central and north basins
of Lake Baikal from seismic stratigraphy, Geol. Soc. Am. Bull., 109,
746–766, 1997.
Mutter, J. C., W. R. Buck, and C. M. Zehnder, Convective partial melting,
1, A model for the formation of thick basaltic sequences during the
initiation of spreading, J. Geophys. Res., 93, 1031–1048, 1988.
Nelson, K. D., J. A. Arnow, J. H. McBride, J. H. Willemin, J. Huang,
L. Zheng, J. E. Oliver, L. D. Brown, and S. Kaufman, New COCORP
profiling in the southeastern United States, part I, Late Paleozoic suture
and Mesozoic rift basin, Geology, 13, 714–718, 1985.
Newmann, E. R., K. H. Olsen, and W. S. Baldridge, The Oslo rift, in
Continental Rifts: Evolution, Structure, Tectonics, edited by K. H. Olsen,
pp. 345–373, Elsevier Sci., New York, 1995.
Olsen, P. E., D. V. Kent, B. Cornet, W. K. Witte, and R. W. Schlische, High-
resolution stratigraphy of the Newark rift basin, Geol. Soc. Am. Bull.,
108, 40–77, 1996.
Pavlenkova, N. I., Crust and upper mantle structure in northern Eurasia
from seismic data, Adv. Geophys., 37, 1–129, 1996.
Petit, C., E. B. Burov, and J. Deverchere, On the structure and mechanical
behavior of the extending lithosphere in the Baikal rift from gravity
modelling, Earth Planet. Sci. Lett., 149, 29–42, 1997.
Petit, C., I. Koulakov, and J. Deverchere, Velocity structure around the
Baikal rift zone from teleseismic and local earthquake traveltimes and
geodynamic implications, Tectonophysics, 296, 125–144, 1998.
Poort, J., P. van der Beek, and M. ter Voorde, An integrated modelling study
of the central and northern Baikal rift: Evidence for non-uniform litho-
spheric thinning, Tectonophysics, 291, 101–122, 1998.
Rogozhina, V. A., and V. M. Kozhevnikov, Anomalous Mantle Area Be-
neath the Baikal Rift, 103 pp., Nauka, Novosibirsk, Russia, 1979.
Ruppel, C., Extensional processes in continental lithosphere, J. Geophys.
Res., 100, 24,187–24,215, 1995.
Ruppel, C., M. G. Kogan, and M. K. McNutt, Implications of new gravity
data for the Baikal rift zone structure, Geophys. Res. Lett., 20, 1635–
1638, 1993.
San’kov, V., J. Deverchere, Y. Gaudemer, F. Houdry, and A. Filippov,
Geometry and rate of faulting in the north Baikal rift, Siberia, Tectonics,
19, 707–722, 2000.
Schlische, R. W., Structural and stratigraphic development of the Newark
extensional basin: Evidence for the growth of the basin and its bounding
structures, Geol. Soc. Am. Bull., 104, 1246–1263, 1992.
Scholz, C. A., and D. H. Hutchinson, Stratigraphic and structural evolution
of the Selenga Delta accommodation zone, Lake Baikal rift, Siberia, Int.
J. Earth Sci., 89, 212–228, 2000.
Scholz, C. A., K. D. Klitgord, D. H. Hutchinson, U. S. ten Brink, L. P.
Zonenshain, A. Y. Golmschtok, and T. C. Moore, Results of 1992 seismic
reflection experiment in Lake Baikal, Eos Trans. AGU, 74, 465, 469–
470, 1993.
Steckler, M. S., G. I. Omar, G. D. Karner, and B. P. Kohn, Pattern of
hydrothermal circulation within the Newark basin from fission-track
analysis, Geology, 21, 735–738, 1993.
Tapponnier, P., and P. Molnar, Active faulting and Cenozoic tectonics of the
Tien Shan, Mongolia and Baykal regions, J. Geophys. Res., 84, 3425–
3459, 1979.
van der Beek, P., Flank uplift and topography at the central Baikal rift (SE
Siberia): A test of kinematic models for continental extension, Tectonics,
16, 122–136, 1997.
Vanneste, M., M. De Batist, A. Golmshtok, A. Kremlev, and W. Versteeg,
Multi-frequency seismic study of gas hydrate-bearing sediments in Lake
Baikal, Siberia, Mar. Geol., 172, 1–21, 2001.
Williams, D. F., J. Peck, E. B. Karabanov, A .A. Prokopenko, V. Krav-
chinsky, J. King, and M. I. Kuzmin, Lake Baikal record of continental
climate response to orbital insolation during the past 5 million years,
Science, 278, 1114–1117, 1997.
Windley, B. F., and M. B. Allen, Mongolian plateau: Evidence for a late
Cenozoic mantle plume under central Asia, Geology, 21, 295–298, 1993.
Yin, A., Mode of Cenozoic east-west extension in Tibet suggesting a com-
mon origin of rifts in Asia during the Indo-Asian collision, J. Geophys.
Res., 105, 21,745–21,759, 2000.
Zamarayev, S. M., and V. V. Ruzhich, On relationships between the Baikal
rift zone and ancient structures, Tectonophysics, 45, 41–47, 1978.
Zelt, C. A., and R. B. Smith, Seismic traveltime inversion for 2-D crustal
velocity structure, Geophys. J. Int., 108, 16–34, 1992.
Zonenshain, L. P., and L. A. Savostin, Geodynamics of the Baikal rift zone
and plate tectonics of Asia, Tectonophysics, 76, 1–45, 1981.
Zonenshain, L. P., M. I. Kuzmin, and L. M. Natapov, Geology of the USSR:
A Plate-Tectonic Synthesis, Geodyn. Ser., vol. 21, 242 pp., AGU,
Washington, D. C., 1990.
Zorin, Y. A., and S. V. Lepina, Geothermal aspects of development of
asthenospheric upwelling beneath continental rift zones, J. Geodyn., 3,
1–22, 1985.
Zorin, Y. A., V. M. Kozhevnikov, M. R. Novoselova, and E. K. Turutanov,
Thickness of the lithosphere beneath the Baikal rift zone and adjacent
regions, Tectonophysics, 168, 327–337, 1989.

M. H. Taylor, Department of Earth and Space Sciences, University of
California, 3806 Geology Bldg., Los Angeles, CA 90095, USA. (mtaylor@
ess.ucla.edu)
U. S. ten Brink, U.S. Geological Survey, Woods Hole Field Center,
Woods Hole, MA 02543, USA. (utenbrink@usgs.gov)
TEN BRINK AND TAYLOR: CRUSTAL STRUCTURE OF BAIKAL RIFT ETG 2 - 15