Crustal structure of the ocean-continent transition at Flemish Cap: Seismic refraction results

Abstract

We conducted a seismic refraction experiment across Flemish Cap and into the deep basin east of Newfoundland, Canada, and developed a velocity model for the crust and mantle from forward and inverse modeling of data from 25 ocean bottom seismometers and dense air gun shots. The continental crust at Flemish Cap is 30 km thick and is divided into three layers with P wave velocities of 6.0-6.7 km/s. Across the southeast Flemish Cap margin, the continental crust thins ? 90-km-wide zone to only 1.2 km. The ocean-continent boundary is near the base of Flemish Cap and is marked by a fault between thinned continental crust and 3-km-thick crust with velocities of 4.7-7.0 km/s interpreted as crust from magma-starved oceanic accretion. This thin crust continues seaward for 55 km and thins locally to ∼1.5 km. Below a sediment cover (1.9-3.1 km/s), oceanic layer 2 (4.7-4.9 km/s) is ∼1.5 km thick, while layer 3 (6.9 km/s) seems to disappear in the thinnest segment of the oceanic crust. At the seawardmost end of the line the crust thickens to ∼6 km. Mantle with velocities of 7.6-8.0 km/s underlies both the thin continental and thin oceanic crust in an 80-km-wide zone. A gradual downward increase to normal mantle velocities is interpreted to reflect decreasing degree of serpentinization with depth. Normal mantle velocities of 8.0 km/s are observed ∼6 km below basement. There are major differences compared to the conjugate Galicia Bank margin, which has a wide zone of extended continental crust, more faulting, and prominent detachment faults. Crust formed by seafloor spreading appears symmetric, however, with 30-km-wide zones of oceanic crust accreted on both margins beginning about 4.5 m.y. before formation of magnetic anomaly MO (∼118 Ma).

Full text

Crustal structure of the ocean-continent transition at Flemish Cap:
Seismic refraction results
Thomas Funck,
1
John R. Hopper,
1
Hans Christian Larsen,
1
Keith E. Louden,
2
Brian E. Tucholke,
3
and W. Steven Holbrook
4
Received 6 February 2003; revised 29 July 2003; accepted 11 August 2003; published 19 November 2003.
[1]We conducted a seismic refraction experiment across Flemish Cap and into the deep
basin east of Newfoundland, Canada, and developed a velocity model for the crust
and mantle from forward and inverse modeling of data from 25 ocean bottom
seismometers and dense air gun shots. The continental crust at Flemish Cap is 30 km thick
and is divided into three layers with Pwave velocities of 6.0–6.7 km/s. Across the
southeast Flemish Cap margin, the continental crust thins over a 90-km-wide zone to only
1.2 km. The ocean-continent boundary is near the base of Flemish Cap and is marked
by a fault between thinned continental crust and 3-km-thick crust with velocities of
4.7–7.0 km/s interpreted as crust from magma-starved oceanic accretion. This thin crust
continues seaward for 55 km and thins locally to 1.5 km. Below a sediment cover
(1.9–3.1 km/s), oceanic layer 2 (4.7– 4.9 km/s) is 1.5 km thick, while layer 3 (6.9 km/s)
seems to disappear in the thinnest segment of the oceanic crust. At the seawardmost end of
the line the crust thickens to 6 km. Mantle with velocities of 7.6 –8.0 km/s underlies
both the thin continental and thin oceanic crust in an 80-km-wide zone. A gradual
downward increase to normal mantle velocities is interpreted to reflect decreasing degree
of serpentinization with depth. Normal mantle velocities of 8.0 km/s are observed 6km
below basement. There are major differences compared to the conjugate Galicia Bank
margin, which has a wide zone of extended continental crust, more faulting, and
prominent detachment faults. Crust formed by seafloor spreading appears symmetric,
however, with 30-km-wide zones of oceanic crust accreted on both margins beginning
about 4.5 m.y. before formation of magnetic anomaly M0 (118 Ma). INDEX TERMS:
3025 Marine Geology and Geophysics: Marine seismics (0935); 8105 Tectonophysics: Continental margins
and sedimentary basins (1212); 9325 Information Related to Geographic Region: Atlantic Ocean; KEYWORDS:
refraction seismics, ocean-continent transition, serpentinized mantle
Citation: Funck, T., J. R. Hopper, H. C. Larsen, K. E. Louden, B. E. Tucholke, and W. S. Holbrook, Crustal structure of the ocean-
continent transition at Flemish Cap: Seismic refraction results, J. Geophys. Res.,108(B11), 2531, doi:10.1029/2003JB002434, 2003.
1. Introduction
[2] Rifted margins are created by extension and breakup
of continental crust to form intervening ocean basins. They
are often classified as volcanic or nonvolcanic based on the
amount of extrusive and intrusive magmatic activity during
the rifting [Louden and Lau, 2001]. The mechanical behav-
ior of the lithosphere under extension is well studied on
nonvolcanic margins where the extensional fabric has not
been modified by large volumes of synrift or postrift
volcanism. The best studied nonvolcanic margin is the
Iberia margin, which was drilled at Galicia Bank during
leg 103 of the Ocean Drilling Program (ODP) and in the
Iberia Abyssal Plain during legs 149 and 173. The Iberia
margin is characterized by the exhumation of continental
mantle during the final stages of breakup. Shallow mantle
rocks were then transformed to serpentinite by high-tem-
perature interaction with seawater [Sawyer et al., 1994;
Whitmarsh et al., 1996; Discovery 215 Working Group,
1998; Whitmarsh et al., 2000]. The extreme lithospheric
thinning was accompanied by little or no decompressional
melting of the asthenospheric mantle, which poses a
problem for melting models at rifted margins [Minshull et
al., 2001]. The width of the transition zone exhibiting
serpentinized mantle varies along strike off Iberia, with a
wider zone in the southern Iberia Abyssal Plain [Dean et al.,
2000] and a narrow zone off Galicia Bank [Whitmarsh et
al., 1996].
[3] The Newfoundland continental margin is conjugate to
the Iberia margin (Figure 1) and geophysical studies indi-
cate significant cross-rift asymmetries. Previous refraction
seismic work on the Newfoundland margin is sparse, but
JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 108, NO. B11, 2531, doi:10.1029/2003JB002434, 2003
1
Danish Lithosphere Centre, Copenhagen, Denmark.
2
Department of Oceanography, Dalhousie University, Halifax, Nova
Scotia, Canada.
3
Department of Geology and Geophysics, Woods Hole Oceanographic
Institution, Woods Hole, Massachusetts, USA.
4
Department of Geology and Geophysics, University of Wyoming,
Laramie, Wyoming, USA.
Copyright 2003 by the American Geophysical Union.
0148-0227/03/2003JB002434$09.00
EPM 10 -1

there is no indication of exhumed mantle in the transition
zone between the extended continental crust and normal
oceanic crust. However, there is evidence for serpentinized
mantle underneath oceanic and thinned continental crust
(Reid’s [1994] reinterpretation of Todd and Reid [1989];
Louden and Chian’s [1999] reinterpretation of Reid and
Keen [1990]; Reid [1994]).
[4] Consistent data along both margin conjugates are
necessary to understand the asymmetries and to advance
the interpretation of existing data sets. Since the Newfound-
land margin is not as well sampled as the Iberia counterpart,
a reflection and refraction seismic study was initiated to
improve the image of the crustal structure along the New-
foundland margin. The experiment was carried out in 2000
as part of the SCREECH (Study of Continental Rifting and
Extension on the Eastern Canadian Shelf) project with three
major transects (Figure 2). This paper presents the results of
the northernmost refraction seismic line (line 1) across
Flemish Cap. Record sections of the coincident reflection
seismic line are shown by Hopper et al. [2003]. Line 1 is
conjugate to the ODP drilling transect of leg 103 [Boillot et
al., 1987] and refraction line 6 [Whitmarsh et al., 1996] on
the Galicia Bank continental margin. More recent reflection
and refraction seismic data from Galicia Bank and the
Galicia Interior Basin are presented by Pe´rez-Gussinye´et
al. [2003] and Zeltetal.[2003].
Figure 1. Location map of the conjugate Newfoundland
and Iberia margins. Thin solid line shows the 2000-m
bathymetric contour. Thick solid lines indicate fracture
zones (FZ). Magnetic anomalies M0 and M34 are marked
by dashed lines. Abbreviations are FC, Flemish Cap; GB,
Grand Banks; Nfl, Newfoundland; G, Galicia Bank; IA,
Iberia Abyssal Plain; GS, Goban Spur.
Figure 2. Bathymetric map of the study area showing the location of the three lines of the SCREECH
experiment. Solid bold lines show the shot line. Circles along line 1 indicate the locations of ocean
bottom seismographs (annotations show the station number). Contour interval of the bathymetry is
1000 m (solid lines). In addition, the 200- and 500-m-depth contours are plotted (thin solid lines). Rift
basins on the Grand Banks and Flemish Cap (modified from Grant and McAlpine [1990]) are shown as
shaded regions with dotted lines. Magnetic anomalies M0 and M3 (gray lines) are taken from Srivastava
et al. [2000]. Seismic lines relevant to this study are shown by dashed lines: line 77-3 [Keen and Barrett,
1981]; line 85-3 [Keen et al., 1987]; line 7 [Reid, 1994]; line 91-2 [Marillier et al., 1994]; T1-18 [Todd
and Reid, 1989].
EPM 10 -2 FUNCK ET AL.: SEISMIC REFRACTION STUDY AT FLEMISH CAP MARGIN

[5] The new seismic data on the Newfoundland margin
were collected to determine the exact character of the
perceived asymmetry between the conjugate margins and
the mechanisms that caused the deformation. Some of the
key questions to be addressed by the SCREECH refraction
seismic data include the following: (1) Is the continental
mantle on the Newfoundland margin unroofed and ser-
pentinized in a manner similar to that on the Iberia side?
(2) What is the amount and distribution of crustal exten-
sion? (3) Is there any evidence for significant magmatism?
2. Geological Setting
[6] The Flemish Cap is a roughly circular fragment of
continental crust located some 600 km east of Newfound-
land (Figure 2). The crustal thickness has been estimated to
be 28 km based on reflection seismic data [Keen et al.,
1987, 1989]. Flemish Cap consists of a central core of
rocks of Hadrynian age (750 –830 Ma), with an onlapping
sequence of undisturbed to disturbed Mesozoic-Cenozoic
sediments [King et al., 1985]. Basement in the core zone
consists mainly of weakly metamorphosed granodiorite or
granite, and some volcanic rocks. King et al. [1985]
concluded that Flemish Cap is part of the Avalon Zone of
the Appalachian orogen.
[7] In the southwestern segment of Flemish Cap more
than 3 km of sediments are present in a north trending half-
graben basin bounded to the west by the Beothuk Ridge
[Grant and McAlpine, 1990]. Sediments in this graben are
probably of Early to Late Cretaceous age [Grant and
McAlpine, 1990]. To the west, Flemish Cap is separated
from the Grand Banks by the Flemish Pass rift basin. Moho
depth at Flemish Pass is 22 km and the sediment infill is
around 6 km [Keen and Barrett, 1981], indicating basement
thickness of about 16 km.
[8] Rifting at the Newfoundland margin occurred in three
major phases [Tucholke et al., 1989; Grant and McAlpine,
1990]. During the first phase in the Late Triassic, northeast
trending grabens on the Grand Banks were formed. The
second phase of rifting began in the Late Jurassic and
continued until the Grand Banks/Flemish Cap were sepa-
rated from Iberia, with the rift propagating from south to
north. According to the plate reconstruction of Srivastava
et al. [2000], magnetic anomaly M3 (early Barremian,
126 Ma) is the oldest anomaly recognizable in the north
at Flemish Cap. A third, Late Cretaceous phase of rifting led
to the opening of Labrador Sea and the separation of Orphan
Knoll and the northern part of Flemish Cap from NW
Europe and Rockall Plateau. Seafloor spreading between
northern Flemish Cap and Goban Spur (Figure 1) was
initiated 110 Ma [Srivastava et al., 1988].
3. Wide-Angle Seismic Experiment
3.1. Data Acquisition
[9] Line 1 of the SCREECH experiment is a 320-km-long
NW-SE transect across Flemish Cap into Newfoundland
Basin (Figure 2). The refraction/wide-angle reflection
(R/WAR) seismic work along the line was a two-ship oper-
ation. Ocean bottom seismometer (OBS) operations were
carried out by R/V Oceanus, while the air gun array was
towed by R/V Maurice Ewing. The tuned array consisted of
20 air guns with a total volume of 140 L. Individual gun sizes
ranged from 2.4 L to 14.3 L. The shot spacing was 200 m.
[10] 29 OBS were deployed along the transect. 14 instru-
ments (owned by Dalhousie University and the Geological
Survey of Canada) were equipped with three-component
4.5-Hz geophones and a hydrophone, while the other 15
recorders (owned by Woods Hole Oceanographic Institu-
tion) had a hydrophone component only. The station spac-
ing varied from 21 km in the shallow water of Flemish Cap
to 8 km in Newfoundland Basin. One instrument could not
be recovered after the experiment (OBS 9), and three
instruments did not record any data due to technical prob-
lems (OBS 1, 24, and 29).
[11] While the OBS were being recovered by R/V
Oceanus,R/VMaurice Ewing shot the line a second time
in order to collect coincident multichannel seismic (MCS)
data [Hopper et al., 2003]. Segments of these shots with a
spacing of 50 m were recorded by some of the OBS.
[12] For navigation (OBS and shot locations) and shot
timing, the Global Positioning System (GPS) was used.
Water depths along the transect were obtained from the R/V
Maurice Ewing’s Hydrosweep center beam using a depth-
velocity function from a CTD (conductivity, temperature,
depth) measurement at the southeast end of the transect
down to a depth of 4300 m.
3.2. Data Processing
[13] OBS data were converted to SEGY format, and time
corrections for the drift of the OBS clock were applied. Data
were debiased and resampled to 10 ms. Travel times of the
direct wave were used to determine the exact position of the
OBS at the seafloor, from which the shot-receiver ranges
were calculated. The maximum offset between OBS deploy-
ment position and location on the seafloor was 1350 m.
[14] Amplitude spectra show that the main seismic energy
is in the frequency range from 5 to 10 Hz. The record
sections were band-pass filtered from 4.5 to 11 Hz. Seven
instruments were affected by 6-Hz noise of unknown origin
on the geophone components. Some OBS recorded this
noise continuously while other instruments were only
periodically exposed to the signal. Seismic traces that were
affected by the noise were notch filtered. To preserve the
seismic energy, it was necessary to use steep filter slopes,
which sometimes created ripples around the wavelet.
[15] The seismic data were also deconvolved to sharpen
the wavelet and on some OBS an fk filter was applied to
reduce steeply dipping coherent noise. Traces in the record
sections in Figures 3 –7 are weighted by their distance to the
OBS to increase the amplitudes for large offsets.
3.3. Methodology
[16] Deployment positions of the OBS were located along
a great circle arc, which defines the baseline for the two-
dimensional velocity modeling. OBS 1 defines the origin of
the horizontal axis (x= 0 km) and the southeasternmost
instrument (OBS 29) was located at a range of 319.84 km.
After recalculation of the OBS locations, instrument posi-
tions were projected onto this baseline.
[17] The velocity model for the crust and uppermost
mantle was developed using the programs RAYINVR and
TRAMP [Zelt and Smith, 1992; Zelt and Forsyth, 1994].
First, observed travel times were fitted by forward model-
FUNCK ET AL.: SEISMIC REFRACTION STUDY AT FLEMISH CAP MARGIN EPM 10 -3

ing, with the model being developed from top to bottom.
Additional constraints were obtained from the coincident
reflection seismic record [Hopper et al., 2003] from km 202
to the seaward end of the line. These data define the
sedimentary structure and the basement morphology in
more detail than modeling of the wide-angle seismic data
permits. After an initial model was obtained, velocities and
the geometry of layer boundaries were refined using the
inversion algorithm in RAYINVR. Amplitude information
was considered in the modeling process by qualitative
comparison of ray theoretical synthetic seismograms with
the field records. The source signal used for the synthetic
seismograms was extracted from a shot with a high signal-
to-noise ratio at an offset of 28 km.
3.4. Seismic Data
[18] The seismic data are generally of good quality.
Record sections (Figures 3 –7) are shown together with
the calculated travel times through the final velocity model
to allow for a comparison how well the model fits the data.
Names of individual seismic phases are based on the later
interpretation of the velocity model.
[19] Record sections from stations deployed on Flemish
Cap (Figures 3 and 4) show two internal crustal reflectors
(P
c1
Pand P
c2
P), subdividing the crust into three layers
(upper, middle and lower crust). Refractions within these
layers are labeled P
c1
,P
c2
, and P
c3
, respectively. Ampli-
tudes of the P
c1
Pand P
c2
Preflections are low and indicate a
small impedance contrast. However, for OBS 2 (Figure 3) a
strong postcritical P
c2
Preflection is observed between off-
sets of 125 and 135 km. These high amplitudes fit well with
results from amplitude modeling. The Moho reflection
(P
m
P) from the base of the crust has a higher amplitude.
Figure 4 shows a rather complicated signature of the P
m
P
phase to the southeast on the outer continental margin when
compared to the smoother shape of the phase to the
northwest, where the rays are reflected from a more or less
flat crust-mantle boundary.
[20] Sediments are identified in three separate areas along
the line. One is located in the northwestern part of Flemish
Cap, one extends from the slope break landward on the
southeast margin, and the last one is the deep Newfound-
land Basin. A maximum of three layers is sufficient to
describe the velocity structure of the sediments and the
according refracted phases are named P
S1
,P
S2
, and P
S3
from top to bottom. Reflections from the base of sediment
layers 1 and 2 are called P
S1
P, and P
S2
P, respectively, unless
they are identical with the basement reflection (P
B
P).
[21] OBS 19 (Figure 5) marks the seaward transition to a
different crustal domain. This is the most seaward instru-
ment that records a refraction (P
c1
) with a phase velocity of
6 km/s, which is indicative of upper continental crust. To
Figure 3. (top) Record section with computed travel times and (bottom) ray path diagram for the
vertical geophone of OBS 2. Horizontal scale in the record section is shot-receiver distance (offset), and
the vertical scale is the travel time using a reduction velocity of 6.5 km/s. A triangle indicates the receiver
location. See text for description of phases and processing. The horizontal scale of the ray path diagram is
distance along the velocity model (Figure 10). Abbreviations are MCB, midcrustal boundary; Sed.,
sediments.
EPM 10 -4 FUNCK ET AL.: SEISMIC REFRACTION STUDY AT FLEMISH CAP MARGIN

the southeast, two new seismic phases appear. The slower
phase has a velocity slightly lower than 5.0 km/s, whereas
the faster one is modeled with a velocity close to 7.0 km/s.
These two refractions are named P
L2
and P
L3
, respectively.
Phase P
L3
is not observed on OBS 23 (Figure 6). Instead the
lower-velocity layer appears to sit directly on top of mantle,
creating a strong P
m
Preflection at an offset of 10 km with
a triplication (P
L2
,P
m
P, and P
n
, where P
n
is the refraction
within the upper mantle). P
n
phases in Newfoundland Basin
have high amplitudes over large ranges (Figure 7) indicating
a higher velocity gradient than expected for normal mantle.
Phase velocities are also <8.0 km/s. OBS 25 (Figure 7), for
example has a P
n
phase velocity of 7.7 km/s to the
northwest. In summary, the uppermost mantle velocity
structure beneath Newfoundland Basin appears to be dif-
ferent from ‘‘normal’’ mantle that is characterized by
velocities around 8.0 km/s and a low-velocity gradient.
[22] Several stations show some surprisingly high-ampli-
tude arrivals interpreted as diffractions based on their steep
dips (e.g., Figures 5 and 6). These diffractions cannot be
modeled with the ray-tracing program used in this study.
4. Results
4.1. Velocity Model
[23] The fit of the calculated to the observed travel times
is illustrated in Figures 8 and 9. The corresponding Pwave
velocity model is shown in Figure 10. The central part
of Flemish Cap from km 80 to km 150 is characterized by
30-km-thick crust that is subdivided into three layers.
Velocities in the 7-km-thick upper crust vary between 6.0
and 6.2 km/s. The base of the upper crust is mapped almost
continuously by weak wide-angle reflections (P
c1
P). Mid-
crustal velocities range from 6.3 to 6.4 km/s with the
midcrustal boundary located at a depth between 17 and
19 km. Lower crustal velocities are 6.6 to 6.7 km/s. Clear
shear wave observations along the line are limited to the
OBS with prominent upper crustal refractions S
c1
in the
central part of Flemish Cap (Figure 4). Computation of
the phase velocities and comparison with the velocities of
the corresponding P
c1
phases yields Pto Swave velocity
ratios between 1.71 and 1.73 (Poisson’s ratio between 0.24
and 0.25).
[24] No major sediment cover was detected on the central
part of Flemish Cap. To the northwest, sediments with a
velocity of 2.8 km/s and a maximum thickness of 900 m
were found. They are deposited on top of a layer up to 6 km
thick with a velocity of 5.4 km/s. The shape of this layer
resembles a sedimentary basin although the velocities could
well be within the range of igneous crust. The possible
origin of this unit will be discussed later.
[25] Southeast of km 150, the continental crust of
Flemish Cap thins abruptly. Close to the slope break there
is a 25-km-wide basin with up to 4 km of sediment. This
Figure 4. (top) Record section with computed travel times and (bottom) ray path diagram for the
vertical geophone of OBS 7. Horizontal scale in the record section is shot-receiver distance (offset) and
the vertical scale is the travel time using a reduction velocity of 6.5 km/s. A triangle indicates the receiver
location. See text for description of phases and processing. The horizontal scale of the ray path diagram is
distance along the velocity model (Figure 10). MCB is midcrustal boundary.
FUNCK ET AL.: SEISMIC REFRACTION STUDY AT FLEMISH CAP MARGIN EPM 10 -5

basin is subdivided into three individual layers with veloci-
ties of 2.1, 3.1, and 4.6 km/s from top to bottom. At km 215,
the crust is only 6 km thick and the middle crust has
disappeared. Velocities similar to middle or lower crust
(6.6 km/s) are found in a small sliver that extends from km
215 to km 231 with a thickness of up to 2 km. Farther
seaward, a 1.5-km-thick layer with velocities similar to upper
crust (6.1 km/s) extends up to km 242. The thin crust between
km 210 and km 240 is covered with three sediment layers.
The upper two layers are stratified units with velocities
between 2.0 and 3.8 km/s, the lower layer consists of two
distinctive features with a dome-like shape and a more
chaotic reflection pattern [Hopper et al., 2003]. Their
velocities are 3.9 to 5.2 km/s in the northwest and 4.9 to
5.2 km/s in the southeast.
[26] Seaward of km 242 is a significant change of crustal
velocity structure. Here the crust consists of two layers, an
upper layer with a thickness between 1 and 2 km and
velocities of 4.7 to 4.9 km/s, and a lower layer with a
maximum thickness of 2 km and velocities of 6.8 to
7.0 km/s. No reflections are observed between these two
layers, indicating a smooth velocity transition at the bound-
ary. The lower layer disappears between km 274 and km 288.
Farther seaward, the lower layer thickens again up to
4 km, while velocities in the upper layer are reduced to
4.3–4.5 km/s. Here the crust resembles normal thickness
oceanic crust with oceanic layers 2 and 3.
[27] Sediments in the deep Newfoundland Basin are
seismically divided into two units, both of them with
stratified layering [Hopper et al., 2003]. The upper sedi-
mentary unit has velocities of 1.9 to 2.2 km/s and a
thickness of 1.2 km. The lower unit, where it is not missing
in the areas of basement highs, has velocities around
3.0 km/s and is generally less than 1 km thick.
[28] Normal mantle velocities along the line are 8.0 km/s.
However, beneath the thin crust (<6 km thickness) from km
217 to km 299, uppermost mantle velocities are significantly
reduced. They are as low as 7.6 km/s directly beneath
the thin crust and increase downward to a normal mantle
velocity of 8.0 km/s at a depth of 6 km below the top of the
basement. This zone of reduced velocities is interpreted as
serpentinized mantle.
4.2. Model Resolution and Uncertainty
[29] Travel time residuals, number of observations, and
normalized c
2
for individual phases are summarized in
Table 1. The total misfit is 105 ms. The estimated pick
uncertainty varied between 30 ms for high-amplitude obser-
vations close to the OBS and 250 ms for weak P
n
arrivals.
Using these uncertainties, the normalized c
2
for the exper-
iment is 1.394, which is close to the optimum value of 1. It
should be noted that the spacing of the travel time picks was
set to 500 m, to be able to separate individual picks during
the modeling process and to avoid bias toward stations
Figure 5. (top) Record section with computed travel times and (bottom) ray path diagram for the
vertical geophone of OBS 19. Horizontal scale in the record section is shot-receiver distance (offset) and
the vertical scale is the travel time using a reduction velocity of 6.5 km/s. A triangle indicates the receiver
location. See text for description of phases and processing. The horizontal scale of the ray path diagram is
distance along the velocity model (Figure 10). Abbreviations are Sed., sediments; Lay., layer.
EPM 10 -6 FUNCK ET AL.: SEISMIC REFRACTION STUDY AT FLEMISH CAP MARGIN

where travel times were picked from the coincident MCS
line with a shot spacing of only 50 m.
[30] Figure 10 shows the values of the diagonal of the
resolution matrix for the velocity nodes of the model. Ideally,
these values are 1, with values less than 1 indicating spatial
averaging of the true Earth’s structure by a linear combination
of model parameters [Zelt, 1999]. Resolution matrix diago-
nals of greater than 0.5 –0.7 indicate reasonably well resolved
model parameters [Lutter and Nowack, 1990]. Resolution is
excellent within the thick continental crust of Flemish Cap
with diagonal values of 0.9 in the central part. Toward the
northwest the resolution decreases as the model boundary is
approached and the ray coverage is less dense. The velocity
model is characterized by a number of layer pinch-outs.
Velocity nodes at the pinch-out locations are singularities
where the resolution is zero. This explains a generally low
crustal resolution between km 200 and km 250.
[31] The small crustal block around km 225 with a
velocity of 6.6 km/s is poorly resolved (resolution <0.1).
However, it is well imaged in the reflection seismic record
[Hopper et al., 2003] where good picks from the reflection
travel time modeling of focusing analysis indicate a velocity
of 6.6 km/s. Changing this velocity by more than 0.3 km/s
results in a blurred image of the reflectors. Moreover, the
boundaries of the block are characterized by strong reflec-
tivity, which requires some impedance contrast to the
neighboring structure. This accords with the velocity model
(Figure 10) where the crust adjacent to the block has
velocities between 6.0 and 6.4 km/s and the underlying
mantle is modeled with 7.6 to 8.0 km/s.
[32] Resolution within the sediments is often reduced
(Figure 10), particularly in the basin northwest of the slope
break. Some of this can be attributed to the effects of layer
pinch-outs, but the main cause is a lack of overlapping ray
coverage. The lowermost sediment layer in Newfoundland
Basin between km 200 and km 240 has a resolution of
<0.2, which according to Lutter and Nowack [1990] has to
be considered as not constrained. However, the 42 obser-
vations of P
S3
phases within that layer have a travel time
residual of 102 ms (Table 1), which is similar to the
overall error. Again, velocities and geometry obtained
from the coincident MCS line [Hopper et al., 2003] guided
the modeling here and is in accord with the few observa-
tions in the R/WAR seismic data. A similar argument can
be used for the upper crustal layer (layer 2) southwest of
km 242 where the resolution is <0.5 except for two small
isolated areas around km 270 and km 295. Refractions
(P
L2
) within this layer are often not reversed, and gener-
ally, the phase cannot be correlated very far from the OBS.
This results formally in a low resolution. However, the
basement, which is guiding the refraction, is well known
from the reflection seismic record [Hopper et al., 2003],
and it eliminates the ambiguity resulting from nonreversed
observations.
[33] Velocity resolution within the mantle is good with
values around 0.8 (Figure 10). Lower resolution at the
Figure 6. (top) Record section with computed travel times and (bottom) ray path diagram for the
hydrophone of OBS 23. Horizontal scale in the record section is shot-receiver distance (offset) and the
vertical scale is the travel time using a reduction velocity of 6.5 km/s. A triangle indicates the receiver
location. See text for description of phases and processing. The horizontal scale of the ray path diagram is
distance along the velocity model (Figure 10). Abbreviations are Sed., sediments; Cont., continental.
FUNCK ET AL.: SEISMIC REFRACTION STUDY AT FLEMISH CAP MARGIN EPM 10 -7

northwestern and southeastern limit of the serpentinized
mantle can be related to layer pinch-out and proximity to the
model boundary.
[34] The only area that is of some concern with respect to
resolution is located southeast of OBS 26 (297 km). Here
the shooting ship had to make way for fishing long lines and
was forced 6.5 km to the north of the baseline on which the
OBS were deployed. The MCS record along the baseline
[Hopper et al., 2003] shows two basement highs in this
segment, around km 309 and km 318, respectively (see also
Figure 10). Much of the data misfit is obviously related to
the possibility that these highs are local features and not
ridges that extend off the line. In addition, instrument
relocations of OBS 27 and 28 have higher uncertainties
due to the larger distances to the shots. It is difficult to
quantify the possible error in velocity in the southeast. The
largest impact should be on the upper crustal layer because
it is affected most by the basement morphology. Some
of the reduced velocities in the upper crustal layer (4.3 –
4.5 km/s), compared to 4.7–4.9 km/s farther to the NW, are
likely related to these geometry problems.
[35] The steep southeastern flank of the serpentinized
mantle around km 298 (Figure 10) is mostly required by
OBS 28, where the P
n
phase would otherwise arrive too
early. Other OBS are less sensitive and can be modeled with
the flank dipping less steeply. Irrespective of the dip, crustal
thickening is required at the southeastern end of the line.
[36] Most layer boundaries are well defined by either
wide-angle reflections (Figure 10), in the cases of subbase-
ment structures, or by the coincident MCS record [Hopper
et al., 2003] in the cases of sedimentary layers and
basement. The least constrained is the transition from
serpentinized mantle to normal mantle. In the velocity
model (Figure 10), the 8.0 km/s contour, which marks
the boundary between serpentinized mantle with a high-
velocity gradient and the normal mantle with a low gradient,
is put at a depth of 12.75 km. At this depth, the travel time
residual of the P
n
phase is a minimum (124 ms). Variations
of 1 km in the depth result in an increase of 6 ms in the
residual, which is not a very significant variation. Stronger
evidence in support of this depth is the amplitude variation.
Once the P
n
rays bottom in the low-gradient normal mantle,
their amplitudes decrease. The P
n
phase for OBS 26
(Figure 11) is relatively strong up to km 234 (corresponding
to an offset of 63 km) and then it becomes so weak that
travel time picking becomes difficult. With the base of the
serpentinized mantle at 12.75 km/s, strong P
n
arrivals can
be expected up to km 238, which is a good match with the
observation. Changing the depth to 11.75 km or 13.75 km
would move the distance to km 251 and km 225, respec-
tively. This is a variation that is well resolvable and it
provides confidence that the depth error of the boundary
(second-order velocity discontinuity) is well within 1 km.
4.3. Gravity Modeling
[37] Two-dimensional gravity modeling (algorithm of
Talwani et al. [1959]) was performed along the line to
verify how consistent the velocity model is with the gravity
Figure 7. (top) Record section with computed travel times and (bottom) ray path diagram for the
vertical geophone of OBS 25. Horizontal scale in the record section is shot-receiver distance (offset) and
the vertical scale is the travel time using a reduction velocity of 6.5 km/s. A triangle indicates the receiver
location. See text for description of phases and processing. The horizontal scale of the ray path diagram is
distance along the velocity model (Figure 10). Abbreviation is Sed., sediments.
EPM 10 -8 FUNCK ET AL.: SEISMIC REFRACTION STUDY AT FLEMISH CAP MARGIN

Figure 8. Comparison of observed and calculated travel times for OBS 2 – 16, shown together with the
corresponding ray paths. Observed data are indicated by vertical bars, with heights representing pick
uncertainty; calculated data are indicated by solid lines. Triangles mark the receiver locations. Horizontal
scale is the model position; a reduction velocity of 6.5 km/s has been applied for the travel times.
Abbreviation Dir. is the direct wave.
FUNCK ET AL.: SEISMIC REFRACTION STUDY AT FLEMISH CAP MARGIN EPM 10 -9

data. Free-air gravity anomalies were obtained from the
shipboard gravimeter on R/V Maurice Ewing (Figure 12).
The initial density model was derived from conversion of P
wave velocities to density using the empirical formula of
Ludwig et al. [1970]
r¼0:00283v4þ0:0704v30:598v2þ2:23v0:7
with rthe density in g/cm
3
and vthe Pwave velocity in
km/s. When the gravity model was extended to infinity to
avoid edge effects, there was a significant misfit in the
northwest due to the proximity of the end of the line to the
rift basin in Flemish Pass. This lateral change in crustal
structure immediately northwest of line 1 is illustrated by
the decrease of the free-air gravity anomaly from Flemish
Cap to Flemish Pass (Figure 13b). In the absence of
detailed velocity information in Flemish Pass from other
data, we raised the Moho depth to 23 km in the northwest
extension of the model to take some of the basin structure
into account. This value is close to the 22 km reported by
Keen and Barrett [1981].
[38] Another necessary change at the northwestern end of
the line was an increase in density from 2.58 to 2.66 g/cm
3
in the 5.4 km/s near-surface layer. Even with this adjustment
there remains a considerable misfit of up to 21 mGal
between the observed and calculated gravity in this area.
The 5.4 km/s layer may not be a truly two-dimensional
feature. At least to the west the layer is intercepted by the
Flemish Cap Basin (Figures 2 and 13b).
[39] To fit the long-wavelength pattern of the gravity
signal, lateral density changes within the mantle had to be
introduced. With the base of the model at a depth of 65 km,
Figure 9. Same as Figure 8 for OBS 17 –28.
EPM 10 -10 FUNCK ET AL.: SEISMIC REFRACTION STUDY AT FLEMISH CAP MARGIN

mantle densities beneath the thinned continental crust and
farther seaward were set to 3.27 –3.29 g/cm
3
. Beneath the
full thickness continental crust of Flemish Cap, a density of
3.32–3.33 g/cm
3
was used. Finally, the density within the
serpentinized mantle from km 231 to km 244 was reduced
from 3.24 to 3.20 g/cm
3
compared to our initial model,
while densities in the lower crustal layer (layer 3) seaward
of km 242 were increased by 0.10 g/cm
3
.
[40] The adjusted density model (Figure 12) fits the slope
anomaly very well, both in terms of the amplitude and shape
of the anomaly. Some misfits can probably be attributed to
deviations from the two dimensionality of the model, as
inferred for the northwestern end of the line. Another such
zone is the sedimentary basin between km 160 and km 185,
which is associated with a local reduction in the calculated
gravity by about 10 mGal. An equivalent signal is missing
in observed gravity, which may be an indication of limited
extent of the basin perpendicular to the line. The density
model is approximately isostatically balanced at its base
(65 km), where the lithostatic pressure is 1950 MPa with
variations of <20 MPa (Figure 12).
5. Discussion
5.1. Continental Crust
[41] The velocity model obtained for the central part of
Flemish Cap is consistent with other observations within the
Avalon Zone of the Appalachians, to which King et al.
[1985] correlates Flemish Cap. On a line 10 km east of
Newfoundland (line 91-2, Figure 2), Marillier et al. [1994]
report velocities between 5.7 and 6.3 km/s in the upper
Avalon crust and 6.8 km/s at its base near 35 km depth.
Within other zones of the Newfoundland Appalachians,
velocities at the top of the crust are also close to 6.0 km/s
and those at the base normally do not exceed 7.0 km/s (see
summary of Hall et al. [1998]). In the Appalachians north
of Newfoundland, Funck et al. [2001] found a velocity
structure almost identical to line 1 across Flemish Cap. They
Figure 10. (top) Pwave velocity model with a contour interval (thin solid lines) of 0.1 km/s between
5.4 and 8.0 km/s. Numbers indicate velocity, in km/s. White numbers indicate areas that are primarily
constrained by focusing analysis of the coincident reflection seismic data [Hopper et al., 2003]. The outer
perimeter of the model with no ray coverage is omitted. Layer boundaries constrained by reflections are
drawn with bold solid lines. Red circles mark the location of OBS used for the modeling, gray circles
show the location of OBS with no data recovery. Station numbers are indicated at the top of the image.
(bottom) Diagonal values of the resolution matrix of the Pwave velocity model displayed in contour
format (0.1 contour interval). Abbreviations are Sed., sediments; MCB, midcrustal boundary; OCB,
ocean-continent boundary.
FUNCK ET AL.: SEISMIC REFRACTION STUDY AT FLEMISH CAP MARGIN EPM 10 -11

also report a reflective boundary between the upper and
middle crust at a depth of 10 km, similar to line 1. Upper
crustal Pwave velocities of 6.0 km/s and a Poisson’s ratio
of 0.24–0.25 are indicative of granite, granodiorite or
gneiss [Holbrook et al., 1992]. These results are consistent
with rock samples from Flemish Cap, which indicate
granodiorite as the major component and minor granite
[King et al., 1985].
[42] The 6-km-thick layer with a velocity of 5.4 km/s at
the northwestern end of line 1 is difficult to interpret. The
velocity could indicate either high-velocity sedimentary
rocks or rather low-velocity igneous crust. The shape of
the unit has similarity to a sedimentary basin, but there is
no recognizable stratification on the coincident reflection
seismic record or along MCS line 85-3 which crosses it
(Figure 2) [Keen et al., 1987]. The highest sedimentary or
possibly basement velocity in the adjacent Flemish Pass
basin is 4.8 km/s [Keen and Barrett, 1981], and there is
no equivalent to the 5.4 km/s layer within Flemish Pass or
farther seaward along line 1. Hence it seems unlikely
that the 5.4 km/s layer is a sediment or sediment/basalt
fill associated with Late Triassic rifting along the New-
foundland margin. Either the layer consists of older
(meta)sediments or it is indeed basement. Basement
velocities of 5.4 km/s are absent elsewhere in the New-
foundland Appalachians, which suggests that the unit may
be prerift sediments or mixed sedimentary/igneous rocks.
Age and thickness of the unit must have resulted in
compaction or low-grade metamorphism, which could con-
tribute to the lack of reflectivity within the unit. Refraction
seismic line 91-2 (Figure 2) parallel to the coast off
Newfoundland shows velocities increasing from 4.6 km/s
through 5.2 to 5.7 km/s over the top 5 km [Marillier et al.,
1994; Hall et al., 1998] and probably correspond to the
Precambrian sediments and volcanics of the Avalon terrane.
Similar ‘‘basement’’ velocities (5.0– 5.3 km/s) are observed
in the Galicia Interior Basin on the conjugate Iberia margin
[Pe´rez-Gussinye´etal., 2003], but the composition of these
rocks is unknown.
[43] The weak wide-angle reflections observed at mid-
crustal levels at Flemish Cap appear to be related to prerift
structures within the continental crust. Reflection seismic
line 85-3 across Flemish Cap (Figure 2) shows midcrustal
reflectivity that can be correlated over wide zones [Keen et
al., 1987]. On the southeast side of Flemish Cap, thinning of
the continental crust starts at km 140 and includes all crustal
layers. The midcrustal layer disappears first at km 220,
while a thin sliver with velocities of 6.6 km/s continues as
far as km 232. From there, only thin upper crust extends
seaward for another 10 km, with crustal thickness around
Table 1. Number of Observations (n), RMS Misfit Between
Calculated and Picked Travel Times (t
rms
), and Normalized c
2
for
Individual Phases
Phase nt
rms
,s c
2
Direct wave 854 0.075 1.858
P
S1
(Basin NW Flemish Cap) 20 0.078 0.708
P
S2
(Basin NW Flemish Cap) 196 0.065 1.376
P
B
P(Basin NW Flemish Cap) 44 0.063 0.637
P
S1
(Basin slope break) 8 0.041 0.298
P
S2
(Basin slope break) 16 0.039 0.177
P
S3
(Basin slope break) 51 0.066 1.760
P
S1
(Newfoundland Basin) 89 0.108 2.181
P
S1
P(Newfoundland Basin) 124 0.101 1.918
P
S2
(Newfoundland Basin) 80 0.149 4.128
P
S2
P(Newfoundland Basin) 34 0.138 3.158
P
S3
(Newfoundland Basin) 42 0.102 3.432
P
B
P(Newfoundland Basin) 100 0.107 2.479
P
c1
1623 0.065 1.015
P
c1
P257 0.088 1.053
P
c2
1174 0.132 1.630
P
c2
P588 0.105 0.642
P
c3
40 0.129 1.993
P
L2
212 0.087 1.281
P
L3
220 0.094 2.643
P
m
P1652 0.113 1.014
P
n
1780 0.124 1.487
All phases 9204 0.105 1.394
Figure 11. (top) Record section of OBS 26 (middle)
compared to the corresponding synthetic seismogram.
Horizontal scale in the record section is shot-receiver
distance (offset) and the vertical scale is the travel time
using a reduction velocity of 6.5 km/s. A triangle indicates
the receiver location. Processing of the record section
includes a 6-Hz notch filter on some traces, deconvolution,
band-pass filter (4.5 –11 Hz), and trace scaling by range.
The synthetic seismogram is band-pass filtered and scaled
in the same way with some Gaussian noise added. (bottom)
The horizontal scale of the ray path diagram is distance
along the velocity model (Figure 10). Rays with bottom
points in the partially serpentinized mantle are drawn as
solid lines, and rays that reach into the normal mantle are
shown as dashed lines. Abbreviation Serp. M., serpentinized
mantle.
EPM 10 -12 FUNCK ET AL.: SEISMIC REFRACTION STUDY AT FLEMISH CAP MARGIN

1.2 km, compared to 30.4 km maximum crustal thickness at
Flemish Cap. This yields a crustal thinning factor bof 25.
However, even with these extreme amounts of crustal
extension there is no indication of major magmatic activity.
5.2. Ocean-Continent Transition
[44] Between the thinned continental crust and what
appears to be close to normal thickness oceanic crust is a
55-km-wide zone of transitional crust with a thickness
between 2 and 3 km (Figure 10). The boundary between
the continental and transitional crust is located just seaward
of a sediment-capped block at km 240 with velocities
around 5.0 km/s. This feature is interpreted as continental
crust of unknown origin based on the velocities and the
reflection pattern [Hopper et al., 2003]. The coincident
reflection seismic line shows a faulted contact between
the two crustal types around km 242 [Hopper et al., 2003].
[45] To evaluate the character of the transitional crust,
three possible models are considered here: (1) thinned
continental crust, (2) exhumed and serpentinized upper
mantle, and (3) igneous crust generated by a slow spread-
ing ridge. There are clear arguments against continental
crust. Velocities in the transitional crust (4.8 km/s in the
upper crust and 6.9 km/s below) are not compatible with
velocities found in the continental crust at Flemish Cap
(6.0 to 6.7 km/s). The 4.8 km/s layer might be attributed to
crustal modifications by faulting, brecciation and hydro-
thermal alterations; however, the deeper 6.9 km/s layer has
significantly higher velocity than any lower continental
crust to the west, and it is distinctly different from the very
thin continental crust immediately to the west. Only if this
crust was heavily intruded by igneous mafic material
would it be likely to have these high velocities.
[46] The second alternative is that the transitional crust is
exhumed and highly serpentinized upper mantle continuous
with the underlying layer with velocities of 7.6 to 8.0 km/s
representing less deeply serpentinized mantle. Hence veloc-
ities in the transitional crust would range from 4.7 to 8.0 km/s
and would be comparable to values obtained from the
conjugate Galicia Bank (3.5–7.8 km/s) [Whitmarsh et al.,
1996] where the velocity gradient has been related to
decreasing serpentinization with depth. The largest problem
in explaining the transitional crust in this manner is that
velocity is layered within three distinct units. In addition,
large sections of the top of the lowermost serpentinized unit
(7.6–8.0 km/s) are characterized by wide-angle reflections,
and an even wider zone exhibits high-amplitude normal
incidence reflectivity suggesting a concordantly stratified
upper crust [Hopper et al., 2003]. Such prominent and
laterally continuous reflections within the serpentinized
mantle are not known from other studies and are much
more likely to represent volcanic strata.
[47] In the transition zone of the southern Iberia Abyssal
Plain (IAP), the velocity model of Dean et al. [2000] shows
two layers (velocities between 4.3 and 7.3 km/s) overlying
serpentinized mantle with velocities between 7.3 and
7.9 km/s. The preferred model for the two upper layers is
that they consist of mantle peridotite serpentinized between
Figure 12. Two-dimensional gravity modeling for line 1. The Pwave velocity model (Figure 10) was
converted to density using the velocity-density relationship of Ludwig et al. [1970] with additional
modification as discussed in the text. Densities in the model (middle) are given in g/cm
3
. (top) Observed
and calculated gravity anomalies are shown by dashed and solid lines, respectively. (bottom) The
lithostatic pressure at the base of the model (depth of 65 km) is shown as solid line.
FUNCK ET AL.: SEISMIC REFRACTION STUDY AT FLEMISH CAP MARGIN EPM 10 -13

Figure 13. (a) Magnetic anomaly map and (b) free-air gravity anomaly map. Magnetic data are taken
from Verhoef et al. [1996], and gravity data are taken from Sandwell and Smith [1997]. Bold solid lines
show the shot lines of the SCREECH experiment with circles indicating the location of OBS. Reflection
seismic line 85-3 [Keen et al., 1987] and refraction lines 7 [Reid, 1994] and T1-T18 [Todd and Reid,
1989] are drawn with dashed lines. Blue lines indicate the location of magnetic anomalies M0 and M3
[Srivastava et al., 2000]. The 200- and 1000-m-depth contours of the bathymetry are drawn as solid lines.
The shading of the magnetic anomalies is by illumination from the east. Abbreviations are Nfld.,
Newfoundland; OCB, interpreted ocean-continent boundary; FCB, Flemish Cap Basin; FPB, Flemish
Pass Basin; JDB, Jeanne d’Arc Basin; WB, Whale Basin; HB, Horseshoe Basin; R., Ridge.
EPM 10 -14 FUNCK ET AL.: SEISMIC REFRACTION STUDY AT FLEMISH CAP MARGIN

25 and 100%. While the velocities are not too dissimilar to
those found in the transition zone of line 1 (4.7 to 7.0 km/s),
there are some differences that may indicate another com-
position. The combined thickness of the two upper layers in
the southern IAP stays fairly constant at about 3 km, which
indicates that the amount of serpentinization as a function of
depth is more or less similar throughout the transition zone.
However, on line 1 in Newfoundland Basin, basement
above the 7.6 km/s serpentinite layer has a more variable
thickness between 1 and 3.5 km. At the location of the
thinnest crust (km 290, Figure 10), the crust appears to be
more intensely faulted than in other areas [Hopper et al.,
2003]. Because faulting normally facilitates serpentinization
by providing pathways for water, we would expect a higher
degree of serpentinization and, hence, low Pwave velocities
reaching greater depths at this location. However, the
opposite is observed. Another difference between the
two transition zones is the reflection character above
the 7.6 km/s serpentinite layer. In the southern IAP, there
is an upper unreflective layer on top of a reflective layer
[Pickup et al., 1996; Dean et al., 2000], while the MCS
record on line 1 [Hopper et al., 2003] shows high reflec-
tivity at the top and reduced reflectivity below. Pickup et al.
[1996] do not see reflectivity in the upper 1–1.5 s, while all
the reflectivity on our line 1 is in the upper second [Hopper
et al., 2003].
[48] The third alternative to explain the transition zone is
that it is magma-starved oceanic crust accreted during slow
or ultraslow spreading. Velocities within the two layers of
the transitional crust between km 240 and 270 fit those of
oceanic layers 2 and 3. In oceanic crust, layer 2 has an
average thickness of 2.11 ± 0.55 km and velocities between
2.5 and 6.6 km/s, while layer 3 has velocities of 6.69 ±
0.26 km/s [White et al., 1992]. Variations in oceanic crustal
thickness are mainly related to thickness variations within
layer 3, while the thickness of layer 2 remains fairly
constant [Mutter and Mutter, 1993]. A refraction seismic
study of the ultraslow spreading Mohns Ridge (16 mm/yr
full spreading rate) reveals a total crustal thickness of 4 km
with a close-to-normal thickness layer 2 (1.4 – 1.7 km),
while oceanic layer 3 appears to be thinned compared to
normal oceanic crust [Klingelho¨ fer et al.,2000].The
transitional crust on line 1 can be interpreted in these terms.
Layer 2 is observed everywhere with a thickness of 1 to
2 km, while thinning of the crust coincides with thinning
or even complete absence of layer 3. Thin oceanic crust
(2 to 4 km) underlain by serpentinized mantle is reported off
western Iberia [Pinheiro et al., 1992; Whitmarsh et al.,
1993, 1996]. Presently, the slowest spreading ridge in the
world is the Gakkel Ridge (<10 mm/yr full spreading rate);
based on gravity studies there, Coakley and Cochran [1998]
estimate the crustal thickness at 1 to 4 km, which fits with
our observations on line 1. Results from a refraction seismic
experiment across the Gakkel Ridge show the presence of
thin (1.9 to 3.3 km) oceanic crust with absence of layer 3 on
most records [Jokat et al., 2003]. Velocities in the upper-
most mantle are between 7.5 and 7.8 km/s and are attributed
to partly serpentinized mantle. This structure is very similar
to our transitional crust at Flemish Cap.
[49] In summary, the velocity structure in the transition
zone is compatible with thin oceanic crust created by
ultraslow seafloor spreading with limited and/or episodic
magma supply operating in parallel with tectonic extension.
Extensional faults and fractures within the crust provided
pathways for seawater to reach and partially serpentinize the
underlying uppermost mantle. On the basis of this interpre-
tation, the ocean-continent boundary (OCB) is located at
km 242, at the location of OBS 19 (Figure 10). This location
is marked on the potential field maps in Figure 13. On the
free-air gravity map (Figure 13b), the OCB is located just
seaward of the prominent gravity low (30 mGal) along the
continental rise and slope. Farther seaward, the gravity
signature is rather smooth with values of 30 mGal (see
also Figure 12). The OCB on line 1 appears to be abrupt
rather than transitional as at the Iberia/Galicia margin where
there is a zone of exhumed continental mantle between the
continental crust and igneous oceanic crust [Whitmarsh et
al., 1996; Dean et al., 2000].
[50] On the magnetic anomaly map, the OCB plots at a
position where the pattern changes from a long-wavelength
magnetic low (200 nT) in the northwest to a more
irregular pattern with shorter wavelengths and anomalies
in the order of ±100 nT to the southeast. The irregular
pattern continues seaward of line 1, and Srivastava et al.
[2000] interpret this region to be created by slow seafloor
spreading based on magnetic evidence. In the area crossed
by line 1, they identified magnetic anomalies M0 and M3.
Anomaly M0 lies seaward of our interpreted OCB, but M3
is located well within the zone that is clearly thinned
continental crust. Hence M0 appears to be the oldest
magnetic anomaly along the northernmost segment of the
Newfoundland margin that can be explained by seafloor
spreading.
[51] The OCB position is consistent with interpreted OCBs
elsewhere along the margin. A refraction seismic line
(line 7, Figure 2) on southern Grand Banks [Reid, 1994]
shows thin, probably basaltic crust (1 to 4 km thick with
velocities of 4.5 km/s) on the eastern segment of the line,
underlain by serpentinized mantle. The interpreted OCB is
located just seaward of the prominent slope anomaly in
gravity (Figure 13b), similar to line 1. Northeast of Flemish
Cap, the conjugate to Goban Spur (Figure 1), Keen et al.
[1987] locate the OCB on reflection seismic line 85-3 in a
position also similar to that on line 1 (Figure 13). The slope
anomaly in gravity is less pronounced on line 85-3, but
oceanic crust is interpreted to start 15 km seaward of the
trough. The magnetic pattern along the oceanic part of line
85-3 also resembles that on line 1.
[52]Todd and Reid [1989] obtained seven short refraction
lines just south of line 1 (Figure 13), and they interpret the
OCB in a position very similar to line 1. Their lines T9-T11
lie within the gravity slope anomaly (Figure 13b) and show
velocities typical of continental crust. Farther seaward (line
T18), a 1.5-km-thick layer with velocities of 4.5 km/s is
observed, which Todd and Reid [1989] interpret as oceanic
crust. The underlying layer with velocities of 7.3 km/s was
originally thought to be oceanic layer 3. However, Reid
[1994] reinterpreted this layer to be possible serpentinized
mantle, which fits well with our interpretation of line 1.
[53] The crust at the seaward end of line 1 is interpreted to
be oceanic crust of close to normal thickness despite the
reduced resolution seaward of km 300 as pointed out in
the error analysis. The available data indicate a division of
the crust in two layers with velocities in the order of 4.5 and
FUNCK ET AL.: SEISMIC REFRACTION STUDY AT FLEMISH CAP MARGIN EPM 10 -15

7 km/s, respectively, and crustal thickening is required by
several OBS. These velocities, even with higher uncertain-
ties, would be difficult to explain with continental crust or
serpentinized mantle. Magnetic interpretations of Srivastava
et al. [2000] support an oceanic character as well.
5.3. Mantle Serpentinization
[54] The velocity model (Figure 10) shows a layer with
velocities of 7.6 –8.0 km/s in an 80-km-wide zone under-
lying the thinned continental crust and the thin transitional
crust with oceanic character. This layer is interpreted as
partially serpentinized mantle and it is analogous to similar
layers at other nonvolcanic rifted margins, for example off
Iberia [Whitmarsh et al., 1996; Chian et al., 1999; Dean et
al., 2000], Greenland and Labrador [Chian and Louden,
1994; Chian et al., 1995a, 1995b], and the southern Grand
Banks [Reid, 1994]. The velocity range observed on line 1
suggests a serpentinite fraction of <10% in the mantle
[Miller and Christensen, 1997].
[55]Pe´rez-Gussinye´andReston[2001] correlate the
onset of serpentinization beneath continental crust with
embrittlement of the entire crust during progressive exten-
sion. Once all of the overlying crust is in the brittle regime,
seawater can reach the mantle through connected networks
of faults and fractures to initiate serpentinization of the
mantle. The entire crust becomes brittle at a stretching
factor b
b
between 3 and 5, depending on the strain rate
[Pe´rez-Gussinye´ and Reston, 2001]. On line 1, apparently
serpentinized mantle first appears beneath crust 8to4km
thick. With an original crustal thickness of 30.4 km, this
corresponds to b
b
= 3.8 to 7.6. These values compare well
with a factor of 5.3 obtained for the Iberia Abyssal Plain
[Pe´rez-Gussinye´etal., 2001].
[56] The high-velocity layer (7.6–8.0 km/s) beneath the
thin oceanic crust is interpreted in a similar way. The thin
oceanic crust is characterized by numerous faults, many of
which can be traced down to the Moho [Hopper et al., 2003],
and it is likely seawater penetrated to, and serpentinized, the
upper mantle. Velocities of 7.6 km/s within the serpentinized
mantle are similar to 7.5 km/s mantle velocities found
beneath thin oceanic crust at the ultraslow spreading Mohns
Ridge [Klingelho¨fer et al., 2000] or the Gakkel Ridge [Jokat
et al., 2003].
5.4. Comparison With Conjugate Margin
[57] Line 1 at Flemish Cap is conjugate to R/WAR seismic
line 6 [Whitmarsh et al., 1996] at the western margin of
Galicia Bank. Velocity models show strong similarities
between the lines. At Galicia Bank, a 60-km-wide zone with
serpentinized mantle underlies both thin continental crust
and, farther seaward, thin oceanic crust (Figure 14). A
peridotite ridge separates oceanic from continental crust.
At Galicia Bank, the serpentinized mantle has velocities
between 7.2 and 7.6 km/s in the deeper parts and as low as
3.5 km/s in the peridotite ridge exposed at the seafloor.
Velocities at the top of oceanic layer 2 vary between 4.6 and
4.8 km/s. Within layer 3, velocities are near 7.0 km/s. All
these velocities are similar to our line 1.
[58] The thickness of the oceanic crust overlying the
serpentinized mantle varies between 2.5 and 3.5 km, which
is very similar to the Flemish Cap margin. Whitmarsh et al.
[1996] attribute the limited melt generation and thus the thin
oceanic crust to cooling of slowly upwelling mantle by
conductive heat loss during a long period of stretching of
the continental lithosphere prior to breakup. In addition, the
spreading rate probably was low; Srivastava et al. [2000]
assumed a full rate of 14 mm/yr.
[59] There are some limitations when comparing the
continental sections on either side of the rift. On the Galicia
side, the existing profiles continue onto relatively unex-
tended continental crust. At Flemish Cap, the extension of
SCREECH line 1 would run into Flemish Pass and farther
into Orphan Basin (Figure 1). However, the continental
crust in Orphan Basin [Chian et al., 2001] was extended
during a later phase of rifting at 110 Ma [Srivastava et al.,
1988]. This prevents a full reconstruction of the prerift
structure of our line 1. Nevertheless, with these limitations
in mind, we see some differences between the two margins
(Figure 14). Most notably, the crustal thinning at Galicia
extends over a wider zone. Beneath Galicia Bank maximum
crustal thickness is on the order of 20 – 22 km [Whitmarsh et
al., 1996; Gonza´lez et al., 1999] and in Galicia Interior
Basin, between Galicia Bank and Iberia, crust thins to as
little as 12 km [Gonza´ lez et al., 1999]. Full thickness
continental crust in northwest Iberia is 32 km [Cordoba
et al., 1987]. This value is comparable to the maximum
Moho depth of 30 km at Flemish Cap. All the continental
crust seaward of the Iberia mainland has been extended.
Initial rifting occurred in Triassic-Liassic time, but it appears
to have concentrated in Galicia Interior Basin in the Early
Cretaceous (Valanginian) and then on the seaward margin of
Galicia Bank beginning in Hauterivian time, prior to break-
up [Murillas et al., 1990]. Similar rift phases [Tankard and
Welsink, 1987; Driscoll et al., 1995] formed basins in the
Grand Banks, including Flemish Pass Basin, but they appear
not to have extended the crust of Flemish Cap.
[60] According to Whitmarsh et al. [1996], the continental
crust at Galicia Bank is divided into only two layers with
velocities between 5.2 and 6.9 km/s, whereas the Flemish
Cap crust exhibits three layers with velocities ranging from
6.0 to 6.7 km/s. From these results, it would appear that the
two continental blocks have different internal structure or
composition. However, Gonza´lez et al. [1999] modeled
the Galicia crust with three layers of velocity 6.0 – 6.1 km/s,
6.4–6.5 km/s, and 6.8– 6.9 km/s, similar to our results at
Flemish Cap (Figure 10). The Whitmarsh et al. [1996] results
were constrained with just two OBS deployments along the
continental segment, so uncertainties in these data may
explain some of the differences with our Flemish Cap results.
[61] Figure 14a shows a reconstruction of the Flemish
Cap-Galicia Bank rift at the time of initial seafloor spread-
ing, i.e., matching the contact between the continental and
oceanic crust at Flemish Cap with the western limit of the
peridotite ridge off Galicia [Whitmarsh et al., 1996]. The
reconstruction shows that the final breakup position, with
respect to distribution of thinned continental crust, is
skewed to the Canadian side of the rift. The Flemish Cap
margin appears very narrow and abrupt whereas the conju-
gate is a broad zone of thinned continental crust. This is
similar to models of rifting that examine possible causes of
asymmetric margin pairs that show wide rift mode of
extension is likely to preferentially break along one edge,
leaving a narrow/wide conjugate margin pair [Dunbar and
Sawyer, 1989; Hopper and Buck, 1996]. As would be
EPM 10 -16 FUNCK ET AL.: SEISMIC REFRACTION STUDY AT FLEMISH CAP MARGIN

Figure 14. Reconstruction of the cross sections of the Flemish Cap-Galicia Bank margins at (a) the
initiation of seafloor spreading and (b) at the time of magnetic anomaly M0 (118 Ma) using the
identification of magnetic anomalies given by Srivastava et al. [2000]. The model for Galicia Bank is
taken from Whitmarsh et al. [1996]; the model for the Galicia Interior Basin and Galicia is from Gonza´lez
et al. [1999]. Horizontal scale is distance from the ridge/spreading axis (dashed line). Continental crust is
marked by pluses, oceanic layer 3 is marked by v’s, layer 2 is marked by inverted v’s, and serpentinized
mantle by gray shading. Abbreviations are Sed., sediments; Serp., serpentinized.
FUNCK ET AL.: SEISMIC REFRACTION STUDY AT FLEMISH CAP MARGIN EPM 10 -17

expected from the distribution of thinned continental crust,
Galicia Bank experienced more intense faulting than Flem-
ish Cap [Reston et al.,1996;Mauffret and Montadert,
1987]. Prominent and extensive detachment faults known
as ‘‘S’’ reflections developed under continental extension on
the Galicia margin [Reston et al., 1996]. At Flemish Cap,
there is no evidence for subhorizontal detachments in
continental crust [Hopper et al., 2003]. A final asymmetry
is observed in basement depth. Galicia crust is on the
order of 1 km deeper than Flemish Cap crust at the OCB
(Figure 14a) and this asymmetry persists into the oldest
ocean crust (Figure 14b).
[62] Figure 14b reconstructs the Flemish Cap to the
Galicia Bank margin at the time of magnetic anomaly
M0 (118 Ma) as identified by Srivastava et al. [2000].
The reconstruction indicates a 32-km-wide zone of oceanic
crust on the Flemish Cap side, very similar to the 31 km
found at Galicia, and it suggests a symmetric seafloor
spreading pattern following final breakup. Using a half-
spreading rate of 7 mm/yr as assumed by Srivastava et al.
[2000], the onset of seafloor spreading dates to 4.5 m.y.
prior to magnetic anomaly M0 (123 Ma, Barremian on the
timescale of Kent and Gradstein [1986]).
[63] While seafloor spreading appears to be symmetric
between the conjugate margins (Figure 14b), there are
differences in thickness of oceanic crust on the two sides.
At Galicia Bank, the zone with 2.5- to 3.5-km-thin oceanic
crust is only 17 km wide before the crust thickens seaward
to 7 km. In contrast, all the oceanic crust landward of
anomaly M0 at Flemish Cap is only 3 km thick. The
resulting 4 km misfit at M0 could relate either to problems
with identification of magnetic anomalies or to problems
with the velocity models. Some caution certainly is required
when using the magnetic interpretations of Srivastava et al.
[2000]. For instance, their location of magnetic anomaly M3
at Flemish Cap (Figure 13a) lies within crust that appears to
be continental according to our velocity model and MCS
interpretations. If both the Flemish Cap and Galicia velocity
models are correct, then a rather awkward model of
asymmetric melt accretion or a ridge jump is required to
explain the differences in crustal thickness. In the latter
case, very asymmetric seafloor spreading must have oc-
curred prior to anomaly M0.
[64] Alternatively, it is possible that the velocity models
are incorrect. The thickness of the oceanic crust at Flemish
Cap is well constrained northwest of OBS 26 (Figure 10).
We therefore revisit the experiment of Whitmarsh et al.
[1996] to examine whether their model would permit any
modifications. It is noteworthy that the two conjugate lines
would fit very well if the interpreted lower oceanic crust
(layer 3) at Galicia Bank was actually serpentinized mantle
(Figure 14b). First, looking at the gravity model presented
by Whitmarsh et al. [1996, Figure 9], one can see that the
observed gravity (18 mGal) shows no deviation at the
transition between the thin oceanic crust and normal-thick-
ness oceanic crust. However, the calculated gravity shows a
seaward drop of 12 mGal at the transition. A better fit
between observed and calculated gravity could be obtained
by continuing the serpentinized mantle farther seaward.
[65] If we examine the Galicia seismic data, it is impor-
tant to note that only three OBS were deployed along line 6,
and the segment of oceanic crust in question was not
constrained primarily by observations from line 6 but from
a cross line. This cross line likewise had only three OBS
deployments and the crossing point was close to the
southern end of the line where resolution was generally
reduced. On the cross line, velocities in layer 3 were
modeled with 7.1–7.2 km/s, which is rather high for
layer 3, and velocities in the underlying mantle were
modeled with 7.6 km/s, which is rather low for mantle.
The southernmost OBS showed almost no lateral amplitude
variation in first arrivals, while the synthetic seismogram
computed from the velocity model exhibited striking
variations, sometimes with almost no energy at all. The
amplitudes for this station could be fitted much better if
layer 3 had a higher velocity gradient and a smooth
transition into mantle velocities at the bottom. This model
would merge the separate refraction branches (7.2 and
7.6 km/s) into one branch without large lateral amplitude
variations, and it would be compatible with the conjugate
Flemish Cap line, i.e., thin oceanic crust underlain by
partially serpentinized mantle. Hence a reanalysis of this
section of the Galicia velocity model is justified before we
attempt firm conclusions about the symmetry or asymmetry
of pre-M0 oceanic crust.
6. Conclusions
[66] Our refraction seismic study across the southeastern
Flemish Cap continental margin shows a 30-km-thick sec-
tion of continental crust with velocities distributed in three
layers: 6.0 – 6.2 km/s, 6.3 – 6.4 km/s, and 6.6 – 6.7 km/s.
These velocities are similar to those of Appalachian crust
known from other studies. The continental crust thins
dramatically to only 1.2 km over a distance of 90 km,
where a fault marks the contact with thin oceanic crust
farther seaward. The first 55 km of oceanic crust that
accreted to the margin has a thickness of 3 km but it thins
locally to only 1 km. It is divided into two layers: layer 2
with velocities of 4.7 to 4.9 km/s and layer 3 with a
velocity of 6.9 km/s. Observed variations in oceanic crustal
thickness correlate closely with thickness variations in
layer 3, and there is a complete absence of layer 3 when
total crustal thickness decreases to less than 1.5 km. We
attribute the occurrence of this unusually thin oceanic crust
to ultraslow seafloor spreading. Assuming a full-spreading
rate of 14 mm/y [Srivastava et al., 2000], the 55-km-
wide zone of thin ocean crust represents about 8 m.y. of
seafloor spreading, beginning near 123 Ma (Barremian). At
the seaward end of our profile, the observed seismic phases
are modeled with a velocity structure similar to normal
thickness oceanic crust, although the resolution in this zone
is reduced.
[67] Both the zone of thin oceanic crust (55 km wide) and
the zone of continental crust stretched to a thickness of
less than 6 km (22 km wide) are underlain by partially
serpentinized mantle with velocities between 7.6 and
8.0 km/s. Theoretical calculations of Pe´rez-Gussinye´ and
Reston [2001] indicate that the entire continental crust
becomes brittle when stretched by more than a factor, b
b
,
and this can provide pathways for seawater to reach
the mantle and initiate serpentinization. At Flemish Cap,
the factor b
b
is 4–8. This is consistent with estimates of
b
b
=5.3byPe´rez-Gussinye´ and Reston [2001] for the Iberia
EPM 10 -18 FUNCK ET AL.: SEISMIC REFRACTION STUDY AT FLEMISH CAP MARGIN

margin. We observe no tectonic exposure of upper mantle
such as that in the peridotite ridge at the seaward margin of
Galicia Bank or in an extensive, 130-km-wide zone beneath
Iberia Abyssal Plain south of Galicia Bank [Pickup et al.,
1996].
[68] On the basis of magnetic anomalies identified by
Srivastava et al. [2000], about 30 km of oceanic crust was
accreted on both the Flemish Cap and Galicia margins
between the onset of seafloor spreading and magnetic
anomaly M0. Although the results of Whitmarsh et al.
[1996] indicate that the outer part of this crust had normal
thickness and velocity on the Galicia margin, an examina-
tion of their data suggests that this crust might actually be
thin and overlie serpentinized mantle, much as we observe
on the Flemish Cap margin. If this proves to be the case, the
oceanic crust would be much more symmetric around the
ridge axis at the time of anomaly M0. Neither side of
the spreading axis shows evidence for significant synrift
magmatism, as attested to by the unusually thin crust and an
absence of any high-velocity underplated crustal layers.
[69] Extension of the continental crust in this rift
occurred over a wide zone, bounded on both sides by
sharp decreases from normal crustal thicknesses of 30 km
to thicknesses of <2 km (Flemish margin) to 12 km
(Iberia margin, descending into the Galicia Interior Basin).
On the basis of presently observed crustal thicknesses,
extension factors varied widely through the rift, ranging
from b1.5 beneath Galicia Bank, to b2.3 in Galicia
Interior Basin, to b15–25 at the juncture of Flemish
Cap with the outer margin of Galicia Bank. These reflect a
complex interplay of multiple zones of rifting over several
rift episodes that began in late Triassic time and continued
in to the Early Cretaceous. Final breakup about Barremian
time focused extension near the foot of Flemish Cap,
asymmetrically isolating most rifted continental crust on
the Iberia margin.
[70]Acknowledgments. We thank the officers, crew, scientists, tech-
nicians, and students who helped to conduct the seismic experiment during
R/V Oceanus cruise 359-2 and R/V Maurice Ewing cruise 00-07. This work
was supported by National Science Foundation grant OCE 9819053, the
Danish Research Foundation (Danmarks Grundforskningsfond), and the
Natural Science and Engineering Research Council of Canada. B. Tucholke
also acknowledges support by the Henry Bryant Bigelow Chair in Ocean-
ography at Woods Hole Oceanographic Institution. We also thank two
anonymous reviewers and the Associate Editor for their helpful comments
on the manuscript. Prestack depth migration of the coincident reflection
seismic data was made possible by a grant to the GEOMAR Data
Processing Center (European Union contract HPRI-CT-1999-00037).
Woods Hole Oceanographic Institution contribution 11000.
References
Boillot, G., E. L. Winterer, A. W. Meyer, and Shipboard Scientific Party,
Proceedings of the Ocean Drilling Program, Initial Reports, vol. 103,
663 pp., Ocean Drill. Program, College Station, Tex., 1987.
Chian, D., and K. E. Louden, The continent-ocean crustal transition across
the southwest Greenland margin, J. Geophys. Res.,99, 9117 – 9135,
1994.
Chian, D., C. Keen, I. Reid, and K. E. Louden, Evolution of nonvolcanic
rifted margins: New results from the conjugate margins of the Labrador
Sea, Geology,23, 589 – 592, 1995a.
Chian, D., K. E. Louden, and I. Reid, Crustal structure of the Labrador Sea
conjugate margin and implications for the formation of nonvolcanic con-
tinental margins, J. Geophys. Res.,100, 24,239 – 24,253, 1995b.
Chian, D., K. E. Louden, T. A. Minshull, and R. B. Whitmarsh, Deep
structure of the ocean-continent transition in the southern Iberia Abyssal
Plain from seismic refraction profiles: Ocean Drilling Program (Legs 149
and 173) transect, J. Geophys. Res.,104, 7443 – 7462, 1999.
Chian, D., I. D. Reid, and H. R. Jackson, Crustal structure beneath Orphan
Basin and implications for nonvolcanic continental rifting, J. Geophys.
Res.,106, 10,923 – 10,940, 2001.
Coakley, B. J., and J. R. Cochran, Gravity evidence of very thin crust at
the Gakkel Ridge (Arctic Ocean), Earth Planet. Sci. Lett.,162, 81 – 95,
1998.
Cordoba, D., E. Banda, and J. Ansorge, The Hercynian crust in northwest-
ern Spain: A seismic survey, Tectonophysics,132, 321 – 333, 1987.
Dean, S. M., T. A. Minshull, R. B. Whitmarsh, and K. E. Louden, Deep
structure of the ocean-continent transition in the southern Iberia Abyssal
Plain from seismic refraction profiles: The IAM-9 transect at 4020
0
N,
J. Geophys. Res.,105, 5859 – 5885, 2000.
Discovery 215 Working Group, Deep structure in the vicinity of the ocean-
continent transition zone under the southern Iberia Abyssal Plain, Geol-
ogy,26, 743 – 746, 1998.
Driscoll, N. W., J. R. Hogg, N. Christie-Blick, and G. D. Karner, Exten-
sional tectonics in the Jeanne d’Arc Basin, offshore Newfoundland:
Implications for the timing of break-up between Grand Banks and Iberia,
in The Tectonics, Sedimentation and Palaeoceanography of the North
Atlantic Region, edited by R. A. Scrutton et al., Geol. Soc. Spec. Publ.,
90, 1 – 28, 1995.
Dunbar, J. A., and D. S. Sawyer, Patterns of continental extension along the
conjugate margins of the Central and North Atlantic Oceans and Labra-
dor Sea, Tectonics,8, 1059 – 1077, 1989.
Funck, T., K. E. Louden, and I. D. Reid, Crustal structure of the Grenville
Province in southeastern Labrador from refraction seismic data: Evidence
for a high-velocity lower crustal wedge, Can. J. Earth Sci.,38, 1463 –
1478, 2001.
Gonza´lez, A., D. Co´rdoba, and D. Vales, Seismic crustal structure of
Galicia continental margin, NW Iberian Peninsula, Geophys. Res. Lett.,
26, 1061 – 1064, 1999.
Grant, A. C., and K. D. McAlpine, The continental margin around New-
foundland, in Geology of Canada, vol. 2, Geology of the Continental
Margin of Eastern Canada, edited by M. J. Keen and G. J. Williams,
pp. 239 – 292, Geol. Surv. of Can., Ottawa, 1990.
Hall, J., F. Marillier, and S. Dehler, Geophysical studies of the structure of
the Appalachian orogen in the Atlantic borderlands of Canada, Can. J.
Earth Sci.,35, 1205 – 1221, 1998.
Holbrook, W. S., W. D. Mooney, and N. I. Christensen, The seismic veloc-
ity structure of the deep continental crust, in Continental Lower Crust,
edited by D. M. Fountain, R. Arculus, and R. Kay, pp. 1– 43, Elsevier
Sci., New York, 1992.
Hopper, J. R., and W. R. Buck, The effect of lower crustal flow on con-
tinental extension and passive margin formation, J. Geophys. Res.,101,
20,175 – 20,194, 1996.
Hopper, J. R., T. Funck, B. E. Tucholke, H. C. Larsen, W. S. Holbrook,
K. E. Louden, D. Shillington, and H. Lau, Continental breakup and the
onset of ultra-slow seafloor spreading off Flemish Cap on the Newfound-
land rifted margin, Geology, in press, 2003.
Jokat, W., O. Ritzmann, M. C. Schmidt-Aursch, S. Drachev, S. Gauger, and
J. Snow, Geophysical evidence for reduced melt production on the Arctic
ultraslow Gakkel mid-ocean ridge, Nature,423, 962 – 965, doi:10.1038/
nature01706, 2003.
Keen, C. E., and D. L. Barrett, Thinned and subsided continental crust on
the rifted margin of eastern Canada: Crustal structure, thermal evolution
and subsidence history, Geophys. J. R. Astron. Soc.,65, 443 – 4 65,
1981.
Keen, C. E., R. Boutilier, B. de Voogd, B. Mudford, and M. E. Enachescu,
Crustal geometry and extensional models for the Grand Banks, eastern
Canada: Constraints from deep seismic reflection data, in Sedimentary
Basins and Basin-Forming Mechanisms, edited by C. Beaumont and A. J.
Tankard, Mem. Can. Soc. Pet. Geol.,12, 101 – 115, 1987.
Keen, C. E., C. Peddy, B. de Voogd, and D. Matthews, Conjugate margins
of Canada and Europe: Results from deep reflection profiling, Geology,
17, 173 – 176, 1989.
Kent, D. V., and F. M. Gradstein, A Jurassic to recent chronology, in The
Geology of North America, vol. M, The Western North Atlantic Region,
edited by P. R. Vogt and B. E. Tucholke, pp. 45– 50, Geol. Soc. of Am.,
Boulder, Colo., 1986.
King, L. H., G. B. Fader, W. H. Poole, and R. K. Wanless, Geological
setting and age of the Flemish Cap granodiorite, east of the Grand Banks
of Newfoundland, Can. J. Earth Sci.,22, 1286 – 1298, 1985.
Klingelho¨ fer, F., L. Ge´ li, L. Matias, N. Steinsland, and J. Mohr, Crustal
structure of a super-slow spreading centre: A seismic refraction study of
Mohns Ridge, 72N, Geophys. J. Int.,141, 509 – 526, 2000.
Louden, K. E., and D. Chian, The deep structure of non-volcanic rifted
continental margins, Philos. Trans. R. Soc. London, Ser. A,357, 767–
804, 1999.
Louden, K. E., and H. Lau, Insights from scientific drilling on rifted con-
tinental margins, Geosci. Can.,28, 187 – 195, 2001.
FUNCK ET AL.: SEISMIC REFRACTION STUDY AT FLEMISH CAP MARGIN EPM 10 -19

Ludwig, W. J., J. E. Nafe, and C. L. Drake, Seismic refraction, in The Sea,
edited by A. E. Maxwell, pp. 53 – 84, Wiley-Intersci., Hoboken, N. J.,
1970.
Lutter, W. J., and R. L. Nowack, Inversion for crustal structure using
reflections from the PASSCAL Ouachita experiment, J. Geophys. Res.,
95, 4633 – 4646, 1990.
Marillier, F., et al., Lithoprobe East onshore-offshore seismic refraction
survey - constraints on interpretation of reflection data in the Newfound-
land Appalachians, Tectonophysics,232, 43 – 58, 1994.
Mauffret, A., and L. Montadert, Rift tectonics on the passive continental
margin off Galicia (Spain), Mar. Pet. Geol.,4, 49 – 70, 1987.
Miller, D. J., and N. I. Christensen, Seismic velocities of lower crustal and
upper mantle rocks from the slow-spreading Mid-Atlantic Ridge, south of
the Kane Transform zone (MARK), Proc. Ocean Drill. Program Sci.
Results,153, 437 – 454, 1997.
Minshull, T. A., S. M. Dean, R. S. White, and R. B. Whitmarsh, Anomalous
melt production after continental break-up in the southern Iberia Abyssal
Plain, in Non Volcanic Rifting of Continental Margins:A Comparison of
Evidence From Land and Sea, edited by R. C. L. Wilson et al., Geol. Soc.
Spec. Publ.,187, 537 – 550, 2001.
Murillas, J., D. Mougenot, G. Boillot, M. C. Comas, E. Banda,
and A. Mauffret, Structure and evolution of the Galicia Interior Basin
(Atlantic western Iberian continental margin), Tectonophysics,184, 297 –
319, 1990.
Mutter, C. Z., and J. C. Mutter, Variations in thickness of layer 3 dominate
oceanic crustal structure, Earth Planet. Sci. Lett.,117 , 295 – 317, 1993.
Pe´ rez-Gussinye´ , M., and T. J. Reston, Rheological evolution during exten-
sion at nonvolcanic rifted margins: Onset of serpentinization and devel-
opment of detachments leading to continental breakup, J. Geophys. Res.,
106, 3961 – 3975, 2001.
Pe´ rez-Gussinye´ , M., T. J. Reston, and J. Phipps Morgan, Serpentinization
and magmatism during extension at non-volcanic margins: The effect
of initial lithospheric structure, in Non Volcanic Rifting of Continental
Margins:A Comparison of Evidence From Land and Sea,editedby
R. C. L. Wilson et al., Geol. Soc. Spec. Publ.,187, 551 – 576, 2001.
Pe´ rez-Gussinye´ , M., C. R. Ranero, T. J. Reston, and D. Sawyer, Structure
and mechanisms of extension at the Galicia Interior Basin, west of Iberia,
J. Geophys. Res.,108(B5), 2245, doi:10.1029/2001JB000901, 2003.
Pickup, S. L. B., R. B. Whitmarsh, C. M. R. Fowler, and T. J. Reston,
Insight into the nature of the ocean-continent transition off west Iberia
from a deep multichannel seismic reflection profile, Geology,24, 1079 –
1082, 1996.
Pinheiro, L. M., R. B. Whitmarsh, and P. R. Miles, The ocean-continent
boundary off the western continental margin of Iberia-II. Crustal structure
in the Tagus Abyssal Plain, Geophys. J. Int.,109, 106 – 124, 1992.
Reid, I., Crustal structure of a nonvolcanic rifted margin east of Newfound-
land, J. Geophys. Res.,99, 15,161 – 15,180, 1994.
Reid, I. D., and C. E. Keen, High seismic velocities associated with reflec-
tions from within the lower oceanic crust near the continental margin of
eastern Canada, Earth Planet. Sci. Lett.,99, 118 – 126, 1990.
Reston, T. J., C. M. Krawczyk, and D. Klaeschen, The S reflector west of
Galicia (Spain): Evidence from prestack depth migration for detachment
faulting during continental breakup, J. Geophys. Res.,101, 8075 – 8091,
1996.
Sandwell, D. T., and W. H. F. Smith, Marine gravity anomaly from Geosat
and ERS 1 satellite altimetry, J. Geophys. Res.,102, 10,039 – 10,054, 1997.
Sawyer, D. S., R. B. Whitmarsh, A. Klaus, and Shipboard Scientific Party,
Proceedings of the Ocean Drilling Program, Initial Reports, vol. 149,
719 pp., Ocean Drill. Program, College Station, Tex., 1994.
Srivastava, S. P., J. Verhoef, and R. Macnab, Results from a detailed aero-
magnetic survey across the northeast Newfoundland margin, Part II:
Early opening of the North Atlantic between the British Isles and New-
foundland, Mar. Pet. Geol.,5, 324 – 337, 1988.
Srivastava, S. P., J.-C. Sibuet, S. Cande, W. R. Roest, and I. D. Reid,
Magnetic evidence for slow seafloor spreading during the formation of
the Newfoundland and Iberian margins, Earth Planet. Sci. Lett.,182,61–
76, 2000.
Talwani, M., J. L. Worzel, and M. Landisman, Rapid gravity computations
for two-dimensional bodies with application to the Mendocino submarine
fracture zone, J. Geophys. Res.,64, 49 – 59, 1959.
Tankard, A. J., and H. J. Welsink, Mesozoic extension and styles of basin
formation in Atlantic Canada, AAPG Bull.,71, 1210 – 1232, 1987.
Todd, B. J., and I. Reid, The continent-ocean boundary south of Flemish
Cap: Constraints from seismic refraction and gravity, Can. J. Earth Sci.,
26, 1392 – 1407, 1989.
Tucholke, B. E., J. A. Austin Jr., and E. Uchupi, Crustal structure and rift-
drift evolution of the Newfoundland Basin, in Extensional Tectonics and
Stratigraphy of the North Atlantic Margins, edited by A. J. Tankard and
H. R. Balkwill, AAPG Mem.,46, 247 – 263, 1989.
Verhoef, J., W. R. Roest, and R. Macnab, J. Arkani-Hamed, and Project
Team, Magnetic anomalies of the Arctic and North Atlantic Oceans and
adjacent land areas, Geol. Surv. Can. Open File Rep., 3125a, 1996.
White, R. S., D. McKenzie, and R. K. O’Nions, Oceanic crustal thickness
from seismic measurements and rare earth element inversions, J. Geo-
phys. Res.,97, 19,683 – 19,715, 1992.
Whitmarsh, R. B., L. M. Pinheiro, P. R. Miles, M. Recq, and J.-C. Sibuet,
Thin crust at the Western Iberia ocean-continent transition and ophiolites,
Tectonics,12, 1230 – 1239, 1993.
Whitmarsh, R. B., R. S. White, S. J. Horsefield, J.-C. Sibuet, M. Recq, and
V. Louvel, The ocean-continent boundary off the western continental
margin of Iberia: Crustal structure west of Galicia Bank, J. Geophys.
Res.,101, 28,291 – 28,314, 1996.
Whitmarsh, R. B., M.-O. Beslier, P. J. Wallace, and Shipboard Scientific
Party, Proceedings of the Ocean Drilling Program, Initial Reports,
vol. 173, 493 pp., Ocean Drill. Program, College Station, Tex., 2000.
Zelt, C. A., Modelling strategies and model assessment for wide-angle
seismic traveltime data, Geophys. J. Int.,139, 183 – 204, 1999.
Zelt, C. A., and D. A. Forsyth, Modeling wide-angle seismic data for crustal
structure: Southeastern Grenville Province, J. Geophys. Res.,99,
11,687– 11,704, 1994.
Zelt, C. A., and R. B. Smith, Seismic traveltime inversion for 2-D crustal
velocity structure, Geophys. J. Int.,108, 16 – 34, 1992.
Zelt, C. A., A. Sain, J. V. Naumenko, and D. S. Sawyer, Assessment of
crustal velocity models using seismic refraction and reflection tomography,
Geophys. J. Int.,153, 609 – 626, doi:10.1046/j.1365-246X.2003.01919,
2003.

T. Funck, J. R. Hopper, and H. C. Larsen, Danish Lithosphere Centre,
Øster Voldgade 10, L, 1350 Copenhagen K, Denmark. (tf@dlc.ku.dk)
K. E. Louden, Department of Oceanography, Dalhousie University,
Halifax, Nova Scotia, Canada B3H 4J1. (keith.louden@dal.ca)
B. E. Tucholke, Department of Geology and Geophysics, Woods
Hole Oceanographic Institution, Woods Hole, MA 02543-1541, USA.
(btucholke@whoi.edu)
W. S. Holbrook, Department of Geology and Geophysics, University of
Wyoming, Laramie, WY 82071-3006, USA. (steveh@uwyo.edu)
EPM 10 -20 FUNCK ET AL.: SEISMIC REFRACTION STUDY AT FLEMISH CAP MARGIN