The Davis Strait crust—a transform margin between two oceanic basins

Abstract

The Davis Strait is located between Canada and Greenland and connects
the Labrador Sea and the Baffin Bay basins. Both basins formed in
Cretaceous to Eocene time and were connected by a transform fault system
in the Davis Strait. Whether the crust in the central Davis Strait is
oceanic or continental has been disputed. This information is needed to
understand the evolution of this transform margin during the separation
of the North American plate and Greenland. We here present a 315-km-long
east-west-oriented profile that crosses the Davis Strait and two major
transform fault systems-the Ungava Fault Complex and the Hudson Fracture
Zone. By forward modelling of data from 12 ocean bottom seismographs, we
develop a P-wave velocity model. We compare this model with a density
model from ship-borne gravity data. Seismic reflection and magnetic
anomaly data support and complement the interpretation. Most of the
crust is covered by basalt flows that indicate extensive volcanism in
the Davis Strait. While the upper crust is uniform, the middle and lower
crust are characterized by higher P-wave velocities and densities at the
location of the Ungava Fault Complex. Here, P-wave velocities of the
middle crust are 6.6 km s-1 and of the lower crust are 7.1 km
s-1 compared to 6.3 and 6.8 km s-1 outside this
area; densities are 2850 and 3050 kg m-3 compared to 2800 and
2900 kg m-3. We here interpret a 45-km-long section as
stretched and intruded crust or as new igneous crust that correlates
with oceanic crust in the southern Davis Strait. A high-velocity lower
crust (6.9-7.3 km s-1) indicates a high content of mafic
material. This mantle-derived material gradually intruded the lower
crust of the adjacent continental crust and can be related to the
Iceland mantle plume. With plate kinematic modelling, we can demonstrate
the importance of two transform fault systems in the Davis Strait: the
Ungava Fault Complex with transpression and the Hudson Fracture Zone
with pure strike-slip motion. We show that with recent poles of
rotation, most of the relative motion between the North American plate
and Greenland took place along the Hudson Fracture Zone.

Full text

Geophysical Journal International
Geophys. J. Int. (2013) 193, 78–97 doi: 10.1093/gji/ggs126
GJI Geodynamics and tectonics
The Davis Strait crust—a transform margin between two
oceanic basins
Sonja K. Suckro,1Karsten Gohl,1Thomas Funck,2Ingo Heyde,3
Bernd Schreckenberger,3Joanna Gerlings2,4 and Volkmar Damm3
1Alfred Wegener Institute for Polar and Marine Research (AWI), Am Alten Hafen 26, 27568 Bremerhaven, Germany. E-mail: Sonja.Suckro@awi.de
2Geological Survey of Denmark and Greenland (GEUS), Øster Voldgade 10, DK-1350 Copenhagen K, Denmark
3Federal Institute for Geosciences and Natural Resources (BGR), Stilleweg 2, 30655 Hanover, Germany
4Dalhousie University, Department of Earth Sciences, 1459 Oxford Street, Halifax, N.S., B3H 4R2, Canada
Accepted 2012 December 21. Received 2012 December 11; in original form 2012 August 8
SUMMARY
The Davis Strait is located between Canada and Greenland and connects the Labrador Sea and
the Baffin Bay basins. Both basins formed in Cretaceous to Eocene time and were connected
by a transform fault system in the Davis Strait. Whether the crust in the central Davis Strait
is oceanic or continental has been disputed. This information is needed to understand the
evolution of this transform margin during the separation of the North American plate and
Greenland. We here present a 315-km-long east–west-oriented profile that crosses the Davis
Strait and two major transform fault systems—the Ungava Fault Complex and the Hudson
Fracture Zone. By forward modelling of data from 12 ocean bottom seismographs, we develop
aP-wave velocity model. We compare this model with a density model from ship-borne gravity
data. Seismic reflection and magnetic anomaly data support and complement the interpretation.
Most of the crust is covered by basalt flows that indicate extensive volcanism in the Davis
Strait. While the upper crust is uniform, the middle and lower crust are characterized by higher
P-wave velocities and densities at the location of the Ungava Fault Complex. Here, P-wave
velocities of the middle crust are 6.6 km s−1and of the lower crust are 7.1 km s−1compared
to 6.3 and 6.8 km s−1outside this area; densities are 2850 and 3050 kg m−3compared to
2800 and 2900 kg m−3. We here interpret a 45-km-long section as stretched and intruded
crust or as new igneous crust that correlates with oceanic crust in the southern Davis Strait.
A high-velocity lower crust (6.9–7.3 km s−1) indicates a high content of mafic material. This
mantle-derived material gradually intruded the lower crust of the adjacent continental crust
and can be related to the Iceland mantle plume. With plate kinematic modelling, we can
demonstrate the importance of two transform fault systems in the Davis Strait: the Ungava
Fault Complex with transpression and the Hudson Fracture Zone with pure strike-slip motion.
We show that with recent poles of rotation, most of the relative motion between the North
American plate and Greenland took place along the Hudson Fracture Zone.
Key words: Plate motions; Transform faults; Continental margins: divergent; Crustal struc-
ture; Arctic region.
1 INTRODUCTION
The Davis Strait is located between Canada and Greenland and con-
nects the Baffin Bay in the north with the Labrador Sea in the south
(Fig. 1). The strait is a bathymetric high with water depths <700 m,
while the water depth in the Baffin Bay and the Labrador Sea ex-
ceeds 2000 m. Prominent tectonic features of the Davis Strait are
the Ungava Fault Complex and the Davis Strait High. A line of pos-
itive southwest–northeastward striking free-air gravity anomalies
marks the location of the Ungava Fault Complex, a major transform
fault (Funck et al. 2007; Gregersen & Skaarup 2007; Gerlings et al.
2009). In the centre of the strait, the Davis Strait High is character-
ized by outcropping basement between 66 and 67◦N (Dalhoff et al.
2006).
The Davis Strait area has experienced Paleogene volcanism. Out-
crops of volcanic sequences are located on Disko Island and the
adjacent Nuussuaq Peninsula (Storey et al. 1998; Pedersen et al.
2006). On the Canadian margin, volcanics are mapped at Cape
Dyer (Clarke & Upton 1971) and offshore in seismic reflection data
(Skaarup et al. 2006). Volcanics are drilled offshore at several wells
as indicated in Fig. 1(a).
The Davis Strait crust has long been a subject of debate. Sonobuoy
readings reveal a 22-km-thick crust, which is interpreted as a
thick pile of oceanic crust by Keen & Barrett (1972). Chalmers &
78 C
The Authors 2013. Published by Oxford University Press on behalf of The Royal Astronomical Society.
at Alfred Wegener Institut fuer Polar- und Meeresforschung Bibliothek on April 10, 2013http://gji.oxfordjournals.org/Downloaded from

Davis Strait crust 79
Figure 1. (a) Bathymetric map of the Davis Strait area (GEBCO_08 Grid, Version 20090202, http://www.gebco.net) with place names and locations of
wide-angle seismic data. Abbreviations are: (NP) Nuussuaq Peninsula, (DI) Disko Island, (CD) Cape Dyer. Line AWI-2008500, -600, -700 were acquired
during the MSM09/3 cruise of RV Merian in 2008 (Gohl et al. (2009); Funck et al. (2012); Suckro et al. (2012)); black dots and short black lines are locations
of sonobuoys and profiles of expandable sonobuoys from Keen & Barrett (1972); NUGGET-1 (Funck et al. 2007), NUGGET-2 (Gerlings et al. 2009), and
GR89-WA (Gohl & Smithson 1993) are seismic refraction lines; diamonds mark well locations: (d1) Hellefisk-1, (d2) Ikermiut, (d3) Kangamiut-1, (d4)
Nukik-2, (d5) Nukik-1, (d6) Qulleq-1, (d7) Gjoa G-37, (d8) Ralegh N-18, (d9) Hekja O-71; red diamonds: volcanics are drilled; black diamonds: Precambrian
rocks are drilled; white diamonds: neither is drilled; all well information are from the Natural Resources Canada, originator: Phil Moir. (b) Free-air gravity
anomalies derived from satellite altimetry of the offshore area (Sandwell & Smith 2009), version 18.1; grey shaded areas mark the extend of oceanic crust
on seismic refraction lines (Funck et al. 2007; Gerlings et al. 2009; Funck et al. 2012; Suckro et al. 2012); positive gravity anomalies that mark the Ungava
Fault Complex (UFC) are circled, as is the Davis Strait High (DSH) and the Nuuk Basin (NB); location of the Hudson Fracture Zone (HFZ) after Chalmers
& Pulvertaft (2001). (c) Closeup of the coinciding seismic refraction line AWI-20080700 with OBS locations marked by red dots and line BGR08-301 with
seismic reflection, gravity and magnetic anomaly data.
Pulvertaft (2001) interpret the crust as continental, while Srivastava
et al. (1982) argue that the Davis Strait High is a continental block
and the adjacent basins are underlain by oceanic crust. A seismic
refraction line in southern Davis Strait showed that continental crust
is separated by a 140-km-wide zone of oceanic crust (NUGGET-1,
Funck et al. 2007, Fig. 1b).
To determine the nature of the crust in the central Davis Strait,
a 226-km-long seismic refraction profile was recorded during the
at Alfred Wegener Institut fuer Polar- und Meeresforschung Bibliothek on April 10, 2013http://gji.oxfordjournals.org/Downloaded from

80 S. K. Suckro et al.
cruise MSM09/3 of RV Maria S. Merian in 2008 (Gohl et al. 2009).
Additionally, multi-channel seismic reflection (MCS), ship-borne
gravity and magnetic field data were collected on the same line
with an additional 90 km extend to the east. We here present the
results of P-wave velocity and gravity forward modelling together
with magnetic field and MCS data. The results are used in a plate
kinematic model to determine the role of the Ungava Fault Complex
in the evolution of the Davis Strait.
2 TECTONIC BACKGROUND OF THE
OPENING OF THE LABRADOR SEA
ANDTHEBAFFINBAY
The tectonic evolution of the Davis Strait is linked to the evolu-
tion of the Baffin Bay and the Labrador Sea. These have formed
in the Cretaceous to Eocene during the separation of Greenland
from the North American craton (e.g. Chalmers & Pulvertaft 2001;
Tessensohn & Piepjohn 2000). The time of initial rifting of North
America and Greenland is dated to earliest Cretaceous by Larsen
et al. (1999) from dyke intrusions in southern West Greenland. On
the Nuussuaq Peninsula, tectonic instability with three phases of
uplift occurred in the Maastrichtian (Chalmers et al. 1999). The age
of the oldest oceanic crust in the Labrador Sea is disputed. Roest &
Srivastava (1989) date it to magnetic chron 33 (80 Ma after Grad-
stein et al. 2004, which is used throughout this paper for dating),
while Chalmers & Laursen (1995) use chron 27N (62 Ma). Recent
seismic refraction and gravity data have now confirmed Paleocene
and Eocene oceanic crust in southern Baffin Bay (Funck et al. 2012;
Suckro et al. 2012).
A first volcanic pulse at 60.7–59.4 Ma is identified from volcanics
on Disko Island by Storey et al. (1998) and correlated with the
arrival of the Greenland–Iceland mantle plume in the Davis Strait
area. (Funck et al. 2007) attribute a thick high-velocity lower crust in
their P-wave velocity model of the NUGGET-1 line to the southward
flow of plume material.
During magnetic chron 24R (55 Ma), the relative motion of
Greenland to the North American craton changed from east to north-
east, as indicated by magnetic spreading anomalies in the Labrador
Sea (Roest & Srivastava 1989; Oakey 2005). This caused new frac-
tures and the breaking of Paleocene oceanic crust in the south-
ern Baffin Bay and the evolution of new spreading centres in the
Eoecene (Chalmers & Pulvertaft 2001; Oakey 2005; Suckro et al.
2012). The opening of the Norwegian–Greenland Sea is dated to
chron 24 (Talwani & Eldholm 1977; Olesen et al. 2007), therefore,
Greenland moved as an independent plate from this time until the
end of relative motion between Greenland and the North American
craton (Tessensohn & Piepjohn 2000). According to Storey et al.
(1998), the reorientation of spreading caused a second volcanic
pulse at 54.8–53.6 Ma in the Disko Island area.
Spreading ceased in the Labrador Sea at chron 13 (33 Ma) ac-
cording to Srivastava (1978), while separation of Greenland and
Eurasia and the opening of the Northeast Atlantic are still ongoing.
Since then sedimentation and subsidence are the dominant geologic
processes in the Baffin Bay and the Labrador Sea (Chalmers &
Pulvertaft 2001).
The Ungava Fault Complex consists of several northeast–
southwest striking faults that are oriented along positive gravity
anomalies in the Davis Strait (Fig. 1b, Sørensen 2006). The Ungava
Fault Complex marks the northwestern border of oceanic crust in
the Labrador Sea (Chalmers & Pulvertaft 2001). It is interpreted as
a transform system, linking seafloor spreading in the Labrador Sea
with spreading in the Baffin Bay (Rice & Shade 1982; Chalmers
& Pulvertaft 2001). Skaarup et al. (2006) interpret the Ungava
Fault Complex in the Davis Strait as the continent–ocean bound-
ary of the Greenland plate. East of the Ungava Fault Complex runs
the north–south striking Hudson Fracture Zone, which meets the
Ungava Fault Complex in the Davis Strait (Chalmers & Pulvertaft
2001). The Hudson Fracture Zone was first identified from magnetic
anomaly data by Srivastava (1978).
3DATAACQUISITION
Seismic and potential field data of this study were acquired during
the research cruise MSM09/3 of RV Maria S. Merian in 2008 (Gohl
et al. 2009). The profiles presented here were set up to determine
the crustal thickness and structure across the Davis Strait and the
Ungava Fault Complex (Fig. 1).
We collected seismic refraction data along the 226-km-long pro-
file AWI-20080700 with 12 ocean bottom seismometers (OBS)
(Fig. 1c). Technical details are listed in Table 1. On line BGR08-
301, we recorded MCS and potential field data. BGR08-301 coin-
cides with line AWI-20080700 and extends 90 km further eastwards
(Fig. 1c). Setup parameters of the MCS measurement are summa-
rized in Table 2.
Gravity data were recorded with a KSS31M sea gravimeter
(Bodensee Gravitymeter Geosystem GmbH) at 1 Hz sampling rate.
To reference the ship-borne gravity data, we carried out connection
measurements on land with a LaCoste&Romberg gravity metre at
the beginning and end of the cruise (Gohl et al. 2009). Magnetic
field data were recorded with an Overhauser SeaSPY marine mag-
netometer system towed approximately 600 m behind the vessel.
4SEISMICDATA
4.1 Seismic reflection data
The MCS data are common depth point (CDP) sorted to 6.25 m
and processed with ProMAXTM with the processing steps listed in
Table 3. We were able to remove the first seafloor multiple by a
surface-related multiple estimation procedure. The trade-off of this
Tab l e 1 . Setup parameters of the seismic refraction survey.
OBS type 3-component Mark seismometers,
4.5 Hz natural frequency, 1 hydrophone
OBS spacing nominally 18 km
Seismic source array of 16 G.GunsTM and2Bolt
TM guns
Volume G.GunTM array 50.8 L, 3100 in3
Operation pressure 145 bar
Vo l u m e 2 B o l t TM guns 64 L, 3906 in3
Operation pressure 120 bar
Total source volume 114.8 L, 7006 in3
Shot interval 60 s
Tab l e 2. Setup parameters of the seismic reflection
survey.
Streamer length 3450 m
Number of channels 276
Sampling rate 2 ms
Recording length 14 s
Seismic source array of 16 G.GunsTM
Operation pressure 100–135 bar
Total source volume 50.8 L, 3100 in3
Shot interval 18 s
at Alfred Wegener Institut fuer Polar- und Meeresforschung Bibliothek on April 10, 2013http://gji.oxfordjournals.org/Downloaded from

Davis Strait crust 81
Tab l e 3 . Processing steps applied to the MCS data of
line BGR08-301 in ProMAXTM.
Resampling: 4 ms
Apply geometry: common mid point binning of 6.25 m
Bandpass filter: (4 -) 8–80 (- 160) Hz
Velocity analysis
Surface related multiple estimation
Velocity analysis
Predictive deconvolution
Normal move out correction
Stack
Poststack Kirchhoff migration
procedure is that primary signals are also partly absorbed (white
band between 1 and 2 s from a model distance of 90–290 km,
Fig. 2). Multiples that are not suppressed by this procedure are
multiples of the basement at distances of 0–70 and 95–135 km.
At these locations, the acoustic basement is close to the seafloor
(less than 0.1 s two-way traveltime) and the remaining basement
multiples can easily be confused with seafloor multiples. But their
shape varies from the seafloor morphology, especially at 40 and
115 km model distance (Fig. 2).
We interpret the acoustic basement from the seismic reflection
data in order to use it in the P-wave velocity and the density mod-
els. From distances of 70–95 and 165–325 km, the basement is the
lowest continuous reflector and marks the base of stratified sedi-
mentary sequences. From a distance of 135–165 km, we use the top
of a series of high-amplitude reflectors below a more transparent
sediment succession. The base of these high-amplitude reflectors
cannot be defined from the seismic reflection data, but in combina-
tion with the P-wave velocity and density model, an interpretation
is discussed later. The deformation of sediments in this section will
be discussed later and is therefore highlighted in closeup B of Fig. 2.
As mentioned before, distances of 0–70 and 95–135 km are only
covered by very little sediment. Here, the basement morphology
is best determined from the basement multiples. Dipping reflector
Tab l e 4 . Statistical values of the P-wave velocity model calculated by ray-
invr and dmplstsqr (Zelt & Smith 1992). nis the number of observations;
pick uncertainties are averaged for all observations; RMS is the misfit be-
tween calculated and observed traveltime; the normalized χ2is a measure of
how well-calculated traveltimes are within the range of the pick uncertainty.
Layer nPick uncertainty (ms) RMS (ms) Normalized χ2
Psa −Psd 424 67 47 0.527
PsaP−Psd P211 89 70 0.915
Pse 644 82 48 0.401
PseP288 90 99 1.040
Pbas 84 92 46 0.278
PLVZ P113 70 50 0.296
Pc1707 100 68 0.692
Pc1P188 115 100 1.006
Pc22647 123 170 2.261
Pc2P429 166 153 0.764
Pc31217 189 279 2.064
PmP1286 158 351 4.682
Pn221 200 109 0.300
Total 8459 131 177 1.965
sequences from a distance of 55–67 km are also better visible in the
multiple (closeup A in Fig. 2).
4.2 P-wave velocity model
We relocalized the OBS positions with the arrival of the direct wave.
All refracted and reflected signals were picked with the software zp
(by B. Zelt, www.soest.hawaii.edu/users/bzelt/index.html), using a
bandpass filter of 4–15 Hz applied for the near offset signals (30 km
distance from the station) and 4–10 Hz for greater offsets. Picking
errors were assigned manually to each phase, taking into account
the signal to noise ratio. In Table 4, the assigned pick uncertainties
are summarized for each phase. Refracted phases are named as Player
and reflected phases PlayerP, except for the reflection at the Moho,
PmP, and the refraction in the upper mantle, Pn.
Figure 2. Final processing of MCS data along line BGR08-301; basement is marked in red; depth scale is approximated by average P-wave velocities of
sediments along the profile. Closeup A shows dipping reflectors in the basement multiple. Closeup B shows folded sediments.
at Alfred Wegener Institut fuer Polar- und Meeresforschung Bibliothek on April 10, 2013http://gji.oxfordjournals.org/Downloaded from

82 S. K. Suckro et al.
Figure 3. (a) P-wave velocity model with layer names. Interpretation of the layers are: sb, sc, sd are sediments; se are basalts intercalated with sediments; bas is
a basalt unit; lvz abbreviates low velocity zone and represents buried sediments; c1 is the upper crust, c2 the middle crust and c3 the lower crust. White triangles
indicate OBS locations; rotated numbers are OBS numbers; numbers on contour lines are P-wave velocities in km s−1; thick lines mark layer boundaries that
are constrained by reflected phases; white shaded areas are not passed by rays. (b) Gridded diagonal values of the resolution matrix of the P-wave velocity
model. Layers are annotated; white triangles indicate OBS locations; rotated numbers are OBS numbers.
By forward modelling with the software rayinvr (Zelt & Smith
1992), we obtained the P-wave velocity model in Fig. 3. Ray
coverage of the single layers is displayed in Fig. 4; modelling
of all stations is given in the Appendix in Figs A1 and A2;
examples of modelling for OBS 2, 8 and 11 are displayed in
Figs 5–7. The modelled layers are described in the following
paragraphs. The accuracy of the model depends on the data cov-
erage and quality; typical uncertainties of the P-wave velocity
are ±0.1 km s−1.
Wat e r : For the seawater, we used an average velocity of
1.47 km s−1, which was calculated from a conductivity temper-
ature density (CTD) measurement during the cruise (Gohl et al.
2009). We took the depth of the seafloor from bathymetry data of
the on-board multi-beam echo-sounder.
at Alfred Wegener Institut fuer Polar- und Meeresforschung Bibliothek on April 10, 2013http://gji.oxfordjournals.org/Downloaded from

Davis Strait crust 83
Figure 4. Ray coverage of the different layers in the P-wave velocity model (Fig. 3). Refracted phases are displayed in white, reflected in black.
sa, sb, sc, sd: Sediment layers with P-wave velocities ranging
from 1.5 to 3.5 km s−1are determined from the OBS data (Fig. 3).
The complex structure of the basement is incorporated from the
high-resolution MCS data (Fig. 2).
From a model distance of 68–100 km, a sediment basin with
P-wave velocities from 1.8 to 2.9 km s−1is modelled from phases
of OBS 4 and 5 (Figs 3 and 4). The sediment infill of the basin
at a model distance of 135–165 km consists of two units. A 1-
km-thick unit with P-wave velocities of 1.5–2.4 km s−1overlies a
0.5-km-thick unit with an average P-wave velocity of 3.3 km s−1
(Fig. 3). The low velocity of the upper unit is extrapolated from
the sediment package of the eastern basin. The lower sediment unit
is confirmed by Psd phases of OBS 8 (Fig. 4). The sediments east
of a model distance of 165 km, in the Nuuk Basin, are of similar
character. A 2-km-thick sediment sequence with P-wave velocities
of 1.5–2.6 km s−1overlies a 1-km-thick unit with an average P-wave
velocity of 3.3 km s−1(Fig. 3). Psb,Psc and Psd phases from OBS 9
to 12 confirm these sequences (Fig. 4).
se: We later interpret this layer, with P-wave velocities between
4.1 and 5.1 km s−1, partly as basalts and therefore name it here
separately from the other sediment layers.
At a model distance of 0–68 km, this layer is modelled with
P-wave velocities of 4.4–5.5 km s−1according to Pse phases of
OBS 1, 2 and 4 (Figs 3, 4 and A1). From a model, distance of
35–50 km lies a body of higher velocities (5.4 km s−1). From a
model distance of 68–95 km, the layer se is modelled with 2 km
thickness and is confirmed by Pse phases of OBS 5 (Figs 3 and A1).
From a model distance of 95–135 km, the layer se is only 0.5 km
thick and modelled with a P-wave velocity of 4.8 km s−1west of
117 km (Fig. 3). East of 117 km, a P-wave velocity of 4.1 km s−1is
at Alfred Wegener Institut fuer Polar- und Meeresforschung Bibliothek on April 10, 2013http://gji.oxfordjournals.org/Downloaded from

84 S. K. Suckro et al.
Figure 5. (a) Seismic section of OBS 2, displayed with a reduction velocity of 6 km s−1. (b) The same seismic section with picks in red; the pick length
corresponds to the assigned pick uncertainty; calculated traveltimes are displayed in black with thick black lines corresponding to the picks. (c) P-wave velocity
model with ray paths. Model layers are annotated; black rays indicate reflected phases, white rays refracted phases; thick lines correspond to the picks in the
central panel.
modelled. This velocity difference is needed to account for different
Pse phases from OBS 6 and 7. From a model distance of 135–
165 km, the layer se is 2 km thick and modelled with 4.9–5.2 km s−1,
according to Pse phases of OBS 8 (Figs 3 and 4). The thickness is
confirmed by OBS 9 (Fig. A2). East of a model distance of 165 km,
in the Nuuk Basin, the velocity structure is determined only by a
Pse phase of OBS 11 (Fig. 7), which indicates a P-wave velocity
of 4.0 km s−1.
at Alfred Wegener Institut fuer Polar- und Meeresforschung Bibliothek on April 10, 2013http://gji.oxfordjournals.org/Downloaded from

Davis Strait crust 85
Figure 6. (a) Cutout of the seismic section of OBS 8, displayed with a
reduction velocity of 6 km s−1. (b) The same seismic section with picks in
red. (c) The same seismic section with picks in red and calculated traveltimes
in black. (d) P-wave velocity model with ray paths; black rays indicate
reflected phases, white rays refracted phases; model layers are annotated.
bas: We modelled a separate body of higher P-wave velocities
than the surrounding layer se from a refracted phase Pbas of OBS 2
(Fig. 5). The average P-wave velocity is 5.4 km s−1and the thickness
is 1.5 km.
lvz: Low-velocity zones (LVZs) are modelled at a model distance
of 0–50 km and of 135–170 km. Phases in OBS 1 and 2 indicate a
LVZ at a model distance of 0–50 km by fading Pse and Pbas phases
and by a delay of crustal phases (Figs 5 and A1). We chose a velocity
of 4.9 km s−1for the LVZ, as this is the average P-wave velocity of
the surrounding layer se. The LVZ from a model distance of 135–
170 km was introduced due to delayed phases in OBS 8, as shown
in Fig. 6. The delay of 0.14 s is modelled with a 0.6-km-thick layer
of P-wave velocity of 4.9 km s−1.TheP-wave velocity of the LVZ
has to be smaller than 5.2 km s−1, which is the velocity at the base
of the overlying layer. We have chosen 4.9 km s−1, which is the P-
wave velocity at the top of the overlying layer se. It can therefore be
interpreted as part of this layer, which is later interpreted as basalts
intercalated with sediments.
c1:P-wave velocities of the first crustal layer range from
5.2 km s−1at the top to 5.8 km s−1at the base. The average P-
wave velocity is 5.5 km s−1, which is well confirmed by Pc1phases
throughout the model except for the western end of the model
(Figs 3 and 4). The thickness varies between 0.5 and 3.5 km along
the profile.
From a model distance of 0–68 km, the upper crust (c1) is 2.5–
3.0 km thick, while it thins from 3 to 0.5 km eastwards beneath the
sediment basin from a model distance of 68–95 km (Fig. 3). From
a model distance of 100–210 km, the thickness is more uniform
with 1.5–2.0 km. East of a model distance of 210 km, a thickening
to 3 km is modelled due to Pc1Pphases in OBS 11 and 12 (Figs 4
and A2). The top of the upper crust (c1) is modelled from the
basement interpretation of the MCS data from a model distance of
135–226 km (Figs 2 and 3).
c2: The second crustal layer is modelled with P-wave velocities
of 5.9–6.7 km s−1, except for a model distance of 40–95 km where
it is characterized by higher P-wave velocities of 6.5–6.8 km s−1
(Fig. 3). Here, the middle crust (c2) is only 2.5–4 km thick, while
it reaches 7.5–12 km thickness in the adjacent model parts. Except
for the model boundaries, the velocity structure is well confirmed
by Pc2phases (Fig. 4). The velocity structure west of a profile
distance of 40 km is only confirmed at the top of the layer by OBS
4 (Fig. A1). The velocity at the bottom can thus be chosen in a wide
range. An extrapolation of high velocities, such as in the thin lower
crust section from 40–95 km, did not lead to the required delay
of later crustal phases. We thus adopt a lower velocity structure,
similar to the model distances east of 95 km for the western profile
termination. Also, from model distances of 210–226 km, we use low
P-wave velocities of 5.9–6.3 km s−1instead of 5.9–6.7 km s−1to
model the delay of later arrivals. Fig. A2 shows that the PmPphase in
OBS 12 has travelled through crust with considerably lower P-wave
velocities than the PmPphase in OBS 10 and 11.
c3: The third crustal layer has P-wave velocities between 6.5 and
7.4 km s−1. Similar to the middle crust (c2), the lower crust (c3) is
characterized by higher P-wave velocities in the centre of the model
than at the sides (Fig. 3).
At a model distance of 50–160 km, P-wave velocity ranges from
6.8 to 7.4 km s−1. At 190–226 km, the average velocity is consider-
ably lower with only 6.7 km s−1. This velocity reduction is necessary
to account for the PmPphase in OBS 12. Fig. A2 shows that even
slower velocities are necessary for modelling of OBS 12, but this
would then change the fit of Pc3,PmPand Pnphases in OBS 10 and
11 and we thus did not further lower the P-wave velocities. Similar
to the modelling of the PmPphase of OBS 12, there is a misfit in
the modelling of the PmPphase of OBS 1. Another possibility of
modelling OBS 1 is with a deeper Moho at the eastern termination
of the profile. Because this leads to a misfit with the gravity model
and with data from OBS 4, we did not chose this option. At both
profile terminations, we chose the model that fits best to the data of
OBS with good ray coverage and to the gravity model. The lower
crust is well resolved from a model distance of 65–190 km by Pc3
and PmPphases (Figs 3 and 4). From 0 to 65 km, modelling only
depends on PmPphases (Fig. 4) and P-wave velocities are thus not
well constrained. The depth of the Moho varies between 21 and
24.5 km and is confirmed by various PmPphases (Fig. 4).
Mantle:AP-wave velocity of 7.8 km s−1is modelled at the top
of the mantle from a Pnphase of OBS 11 (Figs 7 and A2).
Table 4 summarizes statistical values as a measure of quality
for the model’s fit to the picked traveltimes. The root mean square
traveltime (RMS) error is calculated by rayinvr from the misfit of
calculated and picked traveltime. The normalized χ2is a measure
at Alfred Wegener Institut fuer Polar- und Meeresforschung Bibliothek on April 10, 2013http://gji.oxfordjournals.org/Downloaded from

86 S. K. Suckro et al.
Figure 7. (a) Seismic section of OBS 11, displayed with a reduction velocity of 8 km s−1. (b) The same seismic section with picks in red; the pick length
corresponds to the assigned pick uncertainty; calculated traveltimes are displayed in black with thick black lines corresponding to the picks. (c) P-wave velocity
model with ray paths. Model layers are annotated; black rays indicate reflected phases, white rays refracted phases; thick lines correspond to the picks in the
central panel.
of how well the calculated traveltimes are within the range of the
assigned pick uncertainties and should ideally be 1. The normalized
χ2of our model is 1.965, which is almost twice the ideal value. But
a comparison with the P-wave velocity models of Mackenzie et al.
(2005) (χ2of 2.563) and Voss & Jokat (2007) (χ2of 2.804 and
of 3.049) shows that χ2values greater than 2 are not uncommon.
The RMS error of our model is 177 ms, which is higher than the
values of the before mentioned publications, which range from 137
at Alfred Wegener Institut fuer Polar- und Meeresforschung Bibliothek on April 10, 2013http://gji.oxfordjournals.org/Downloaded from

Davis Strait crust 87
Tab l e 5 . Corrections applied to the gravity data.
−time shift due to overcritical damping of the sensor
−conversion from instrument reading units to mGal
−tie to world gravity net IGSN 71 with connection
measurements
−correction for the E¨
otv¨
os effect with navigation data
−correction for instrument drift during the cruise
−subtraction of normal gravity (GRS80)
to 164 ms. Especially phases from the lower crust contribute to the
high RMS error. We think that the high RMS error is mainly due
to the low signal to noise ratio of the OBS data. The model depicts
a complex crust, which various vertical features where scattering
of the deep phases can lower the signal amplitudes. Fig. 3 shows
the diagonal values of the resolution matrix as a colour grid. The
resolution is a measure of how well a velocity value is constrained
by all rays passing though it. The layers of the model are over all
well resolved, except for the profile terminations.
5 GRAVITY AND MAGNETIC
ANOMALY DATA
For free-air gravity anomalies, standard processing steps as listed
in Table 5 were applied to the gravity data. We obtained a density
model by forward modelling with the software GM-SYS (Geosoft,
Inc.). For the starting model (Fig. 8c), we used a simplified ge-
ometry of the P-wave velocity model. Line AWI-20080700 of the
P-wave velocity model only extends up to a model distance of
226 km, while gravity data were recorded on line BGR08-301 up to
a model distance of 315 km. Density values were derived from av-
erage P-wave velocities according to (Barton 1986). For simplicity,
we combined the upper three sediment layers with P-wave veloc-
ities of 1.7–2.9 km s−1to one density body of 2200 kg/m3(s1).
The two underlying layers of 3.1–5.6 km s−1are combined to one
layer of 2450 kg m−3density (s2). The basalt flow from a model
distance of 35–50 km is added to the first crustal layer. We used the
basement interpretation of the MCS data along the whole density
model (Figs 2 and 8).
Calculated free-air gravity values of the starting model are gen-
erally too high along the western part of the profile and too low
at the eastern part (Fig. 8b). We therefore divided the mantle at a
model distance of 170 km into a body of 3200 and 3300 kg m−3.
Where this density change was not sufficient, we adjusted the layer
boundaries. From a model distance of 117–135 km, we replaced the
second sediment layer (2450 kg m−3) by the first (2200 kg m−3),
to meet smaller free-air gravity values in this region. This density
change is also indicated by a lateral change in P-wave velocities
(4.8–4.1 km s−1) along line AWI-20080700. Further, we adjusted
the crustal layers east of a model distance of 225 km. This area is
not covered by the P-wave velocity model, so only the depth of the
basement is constrained by the MCS data. To fit the high free-air
gravity values east of 270 km, we modelled a shallowing of the
middle and lower crust.
The average difference between the calculated gravity of the
final model (Fig. 8d) and the observed free-air gravity values is
7.2 mGal, in contrast to 40.5 mGal for the starting model. The
greatest mismatches between modelled and observed gravity occur
at model distances of 0–65 km and 110–150 km. These regions are
in the vicinity of strong positive anomalies off the profile (Fig. 1b)
and we therefore interpret these as the influence of 3-D effects.
To obtain residual magnetic anomaly values, the appropriate
IGRF reference field values (IGRF-10) were removed from the
measured magnetic total intensity. It was necessary to add 100 nT
to the anomaly curve to meet the mean level of two published
magnetic maps (Verhoef et al. 1996; Maus et al. 2009). The mag-
netic anomalies (Fig. 8a) vary between positive and negative values
of −1146 nT (at a model distance of 32 km) and 1015 nT (47 km). In
general, magnetic anomalies have small amplitudes and long wave-
lengths at the locations of sedimentary basins (at model distances
of 68–100 km and east of 135 km) and high amplitudes with small
wavelengths where the basement is near the surface.
6 PLATE KINEMATICS
We use GPlates (www.gplates.org) to visualize the tectonic evolu-
tion of the Davis Strait area. For the relative motion of the Greenland
plate to the North American craton, we use the set of rotation poles
by Oakey (2005). This recent reconstruction complements the pre-
vious reconstruction from Roest & Srivastava (1989). The most
relevant time steps in the tectonic evolution of the Davis Strait, as
illustrated in Fig. 9, are:
90 Ma: Greenland separates from Canada in an eastwards di-
rection; rifting is active, but seafloor spreading has not started in
the Labrador Sea (Roest & Srivastava 1989; Chalmers & Laursen
1995).
57 Ma: Greenland and Canada are at a maximum east-west dis-
tance in the Davis Strait; the motion of Greenland changes from an
eastward to a northeastward direction (Srivastava 1978); seafloor
spreading is active in the Labrador Sea (Srivastava 1978; Chalmers
& Laursen 1995).
33 Ma: Seafloor spreading ceases in the Labrador Sea (Srivastava
1978); Greenland and Canada are placed at their modern configu-
ration.
Between 57 and 33 Ma, Greenland moved northwards by 310 km
relative to the North American craton. This resulted in a narrowing
of the central Davis Strait. If we use the location of the Hudson Frac-
ture Zone as shown in Chalmers & Pulvertaft (2001) for the plate
boundary, pure strike-slip motion occurs along this fault (Fig. 9e). If
we use the location of the Ungava Fault Complex instead, a crustal
overlap of 70 km width must be compensated. The area of this over-
lap coincides with the positive free-air gravity anomalies that are
associated with the Ungava Fault Complex. This is the area where
transpressional forces were compensated.
7 DISCUSSION
7.1 Basalts and sediments
Below the sediment packages sc and sd, we modelled a layer se
(Fig. 3). This layer with P-wave velocities of 4.1–5.1 km s−1is
similar to a layer with P-wave velocities of 4.3–5.3 km s−1, observed
on NUGGET-1 (Funck et al. 2007). This layer was drilled at the
Hekja O-71 and the Gjoa G-37 wells (Fig. 1a) and consists of basalts
intercalated with sediments (Klose et al. 1982). Due to the similarity
of the P-wave velocity character and the proximity to NUGGET-1,
we follow this interpretation for line AWI-20080700/BGR08-301.
At a model distance of 55–68 km, dipping reflectors in the MCS
data confirm this interpretation (basalt flows in closeup A in Fig. 2).
High amplitudes and frequencies of the magnetic anomaly data also
support the interpretation of volcanics near the surface (Fig. 8a).
at Alfred Wegener Institut fuer Polar- und Meeresforschung Bibliothek on April 10, 2013http://gji.oxfordjournals.org/Downloaded from

88 S. K. Suckro et al.
Figure 8. (a) Magnetic anomaly data along line BGR08-301. (b) Free-air gravity data along line BGR08-301. Observed gravity in black, calculated gravity of
the start model in blue (c), of the final model in red (d). (c) Start model of the density modelling; layer boundaries are taken from the P-wave velocity model
and average P-wave velocities are transferred to densities according to Barton (1986). Numbers inside the model indicate densities in kg m−3. (d) Final density
model.
The only indication of the separately modelled body bas with
P-wave velocities of 5.4 km s−1in the MCS data is an undulation of
the basement at a modal distance of 38 km (Fig. 2). It is confirmed
by the density model, where it is modelled with the same density
as the upper crust (model distance 35–50 km). Due to this high
density, we interpret this feature as a separate basalt unit, which
is not intercalated with sediments. Model distances 0–50 km are
underlain by a LVZ, which we interpret as sediments that were
covered by the basalt unit (Fig. 10).
From a model distance of 95–130 km, the layer se is much thinner
than modelled to the west (0.5 km instead of 2 km). High ampli-
tudes and frequencies of the magnetic anomaly data indicate that
volcanics are near the surface (Fig. 8a). From the available data, it is
not clear whether this sequence was deposited on this basement high
with only 0.5 km thickness, or if it was deposited before an uplift of
the basement with 2 km thickness like in the west. In the later case,
1.5 km of it were eroded due to uplift and exposure at the seafloor.
P-wave velocities of 4.8 km s−1from a model distance of 95–117 km
at Alfred Wegener Institut fuer Polar- und Meeresforschung Bibliothek on April 10, 2013http://gji.oxfordjournals.org/Downloaded from

Davis Strait crust 89
Figure 9. Tectonic evolution of the Davis Strait with poles of rotation from Oakey (2005). (a) Configuration at 90 Ma. (b) and (c) The maximum east–west
separation of the North American craton and Greenland is reached at 57 Ma. The area of additional crust relative to 90 Ma (stretched continental and oceanic
crust) is shaded in yellow; the location of the Ungava Fault Complex and the Hudson Fracture Zone are marked. (Lower row) Relative motion between Greenland
and Canada terminates at 33 Ma; the plates are at their present day configuration. (d) Case 1: The Ungava Fault Complex is used as a plate boundary; due to
the northward motion of Greenland, an overlap of crust needs to be compensated (shaded in orange). (e) Case 2: The Hudson Fracture Zone is used as a plate
boundary; only strike-slip motion is active without thickening or thinning of the crust. (f) Free-air gravity anomalies (Sandwell & Smith 2009), version 18.1,
with the outline of overlapping crust, the Ungava Fault Complex in blue, the Hudson Fracture Zone in red, and oceanic crust in the Labrador Sea as outlined
by Chalmers & Pulvertaft (2001) in white; line AWI-20080700/BGR08-301 in the Davis Strait as thick white line.
support this interpretation as do dipping reflectors in the MCS data
(Figs 2 and 11). A graben structure of the interpreted basement
separates this section from lower P-wave velocities (4.1 km s−1)
and densities (2450–2200 kg m−3) from a model distance of
117–130 km (Fig. 11). As P-wave velocities of basalts can range
between 3.5 and 6.5 km s−1due to varying composition and deposi-
tion (Christie et al. 2006), we here also interpret layer se as basalts
intercalated with sediments.
From a model distance of 130–165 km, high-amplitude reflec-
tions of low frequency line up in the MCS data (Fig. 2 with
closeup B). The reflection pattern is similar to drilled volcanics
in the vicinity of the Gjoa G-37 well (fig. 9 in Klose et al. 1982).
The P-wave velocity of 5.0 km s−1is also within the range for
basalts (Christie et al. 2006). This section is underlain by a LVZ,
which represents old sediments that were covered by the basalt
flows.
at Alfred Wegener Institut fuer Polar- und Meeresforschung Bibliothek on April 10, 2013http://gji.oxfordjournals.org/Downloaded from

90 S. K. Suckro et al.
Figure 10. Geological structure of line AWI-20080700/BGR08-301 compiled from the MCS data (Fig. 2), the P-wave velocity and the density models
(Figs 3 and 8).
Figure 11. Line drawing of the MCS data of line BGR08-301 overlain with the time-converted P-wave velocity model from Fig. 3.
East of a model distance of 165 km, in the Nuuk Basin (Fig. 1b),
P-wave velocities of layer se are only 4.6 km s−1(Fig. 3). This is the
only part of the profile, where we interpreted the lower boundary
of this layer as basement instead of the upper boundary. The top
of layer se causes a high-amplitude continuous reflection in the
MCS data from a model distance of 165–190 km (Fig. 2). This is
similar to reflections of the top of basalts from a model distance of
140–165 km. From 165 to 230 km, the upper boundary of layer se
is characterized by diffuse reflections, which can indicate a broken
surface (Fig. 2). Although P-wave velocities of layer se are lower
in the Nuuk Basin than along the rest of the model, we here also
interpret basalt flows due to the high-amplitude reflections in the
MCS data (Fig. 10).
7.2 Crustal structure
The P-wave velocity and density model consist of a three layered
crusts: the upper, middle, and lower crust. While the P-wave ve-
locity and density structure of the upper crust is uniform along the
profile, the middle crust is characterized by higher P-wave veloc-
ities and densities from a model distance of 50–95 km, like the
lower crust between 40 and 170 km. A lateral change was also mod-
elled in the mantle with smaller densities west of a model distance
of 170 km.
7.2.1 Stretched and highly intruded/igneous crust, model
distance: 50–95 km
The higher P-wave velocities and densities of the middle and lower
crust at a model distance of 40–100 km show an increased con-
tent of mafic material. This can be in the form of mafic intrusions
in a stretched and fractured continental crust, or in the form of
newly formed oceanic crust. The following paragraphs discuss both
options.
The average thickness of normal oceanic crust is 7.1 ±0.8 km and
of plume affected oceanic crust is 10.3 ±1.7 km (White et al. 1992).
This is only half of the crustal thickness of our model. From the
top of layer c1 to the base of layer c3 we measure 20 km thickness.
Oceanic crust of a similar thickness is reported at oceanic plateaus
as parts of large igneous provinces. Gohl & Uenzelmann-Neben
(2001) report that a 17-km-thick high-velocity lower crust (P-wave
at Alfred Wegener Institut fuer Polar- und Meeresforschung Bibliothek on April 10, 2013http://gji.oxfordjournals.org/Downloaded from

Davis Strait crust 91
velocities of 7.0–7.5 km s−1) overlains by a 3-km-thick layer of
P-wave velocities of 6.5–6.8 km s−1at the Agulhas Plateau. This
crustal structure is similar to the model of line AWI-20080700 with
P-wave velocities of 6.9–7.3 km s−1in a 15-km-thick lower crust
and 6.3–6.9 km s−1in a 3.5-km-thick middle crust. Therefore, an
interpretation of new igneous crust from the P-wave velocities is
possible.
Other locations of thick oceanic crust are the volcanic conti-
nental margins of East Greenland (Holbrook et al. 2001; Hopper
et al. 2003) (more than 30 km thickness to 18.3 km thickness
depending on the distance to the Iceland hotspot track) and the
Vøring Plateau (Mjelde et al. 2005) (23.5–9 km thickness). Like
the Davis Strait area, both locations were influenced by the Ice-
land mantle plume, with production of thick basalt flows during the
breakup process (Storey et al. 1998; Holbrook et al. 2001; Hopper
et al. 2003; Mjelde et al. 2005). Basalt flows are also present along
AWI-20080700/BGR08-301 with varying thickness. The basalts
from a model distance of 0–68 km are part of the seaward dipping
reflectors at the Baffin Island margin reported by Skaarup et al.
(2006).
A difference to the East Greenland margin and the Vøring Plateau
is the moderate P-wave velocities in the middle and lower crust.
Along AWI-20080700, the middle crust is 3.5 km thick with an
average P-wave velocity of 6.6 km s−1and the lower crust is 14 km
thick with an average P-wave velocity of 7.1 km s−1. Hopper et al.
(2003) model a crust with 6.6 km s−1at the top and 7.5 km s−1
at the base. P-wave velocity models of the East Greenland margin
shown in Holbrook et al. (2001) exceed 7.5 km s−1in the lower
crust. Mjelde et al. (2005) model a layer of 6.8, 7.1 and 7.3 km s−1.
It is therefore likely that the crust along AWI-20080700 does not
consist completely of new igneous material, but of highly intruded
continental crust. According to Rudnick & Fountain (1995), the
middle crust of rifted margins is 7.5 ±5.6 km thick with an average
P-wave velocity of 6.4 ±0.3 km s−1; the lower crust is 8.6 ±5.1 km
thick with a P-wave velocity of 7.0 ±0.3 km s−1. Although rifted
margins vary greatly, these global averages fit well to the layers of
our model (see above). This interpretation requires that the basalt
flows along the model are not products of a breakup, but that they are
related to volcanism along fractures of the Ungava Fault Complex.
Other methods that are used to identify oceanic crust are mag-
netic spreading anomalies and the basement morphology. Because
of the small scales (45 km of crust), no magnetic seafloor spread-
ing anomalies can be expected. The basement morphology is
only visible below the sedimentary basin from a model distance
of 68–95 km in the MCS data. But it cannot distinguished be-
tween a basalt covered continental crust and newly formed oceanic
crust.
As we cannot rule out either interpretation, we refer to the
crust between a model distance of 50–95 km as stretched and in-
truded/igneous crust in the following (Fig. 10).
We compare the crustal model along line AWI-20080700/
BGR08-301 to that of NUGGET-1 (Funck et al. 2007). Along both
profiles, the continental crust of Baffin Island and Greenland is
separated by thin crust with a high content of mafic material. On
NUGGET-1, Funck et al. (2007) modelled a 140-km-long section
of oceanic layers 2 (5.4–6.2 km s−1) and 3 (6.7–7.0 km s−1) un-
derlain by a thick magmatic underplating of P-wave velocities of
7.4 km s−1. On NUGGET-1 and AWI-20080700/BGR08-301, this
crust is divided into a western and an eastern section. On line AWI-
20080700/BGR08-301 at a model distance of 68 km, the upper
crust thins by 1.5 km and rises. The western part, from a model
distance of 50–68 km, is covered by a thick succession of basalts
intercalated with sediments. The eastern part, from 68–100 km, is
also covered by basalts and by a sedimentary basin. On NUGGET-1,
a graben structure filled with basalts divides the western and eastern
section. We interpret the sharp boundary between the eastern and
western segment of intruded/igneous crust as a transform fault of
the Ungava Fault Complex.
Funck et al. (2007) propose that the western part of the
oceanic crust is related to the volcanic type margin of Baffin Is-
land and Labrador. We expand this interpretation to line AWI-
20080700/BGR08-301, as we also imaged basalt flows at the west-
ern end of our profile in the models, the MCS and the magnetic
anomaly data (Figs 2, 3 and 8). These volcanics, southeast of
Cape Dyer, are partly exposed at the seafloor and are mapped by
Skaarup et al. (2006) from seismic reflection lines and potential field
data.
Funck et al. (2007) further describe the evolution of oceanic
crust at the eastern segment as an upwelling of magma in areas of
transtensional movement along the Ungava Fault Complex. From
the plate kinematic reconstruction (Fig. 9), we know that in the
period between 57 and 33 Ma, strike-slip motion and compression
were active in the Davis Strait. The stretched crust must there-
fore have evolved prior to 57 Ma when the strait was opening.
The intruded/igneous crust along line AWI-20080700/BGR08-301
and the oceanic crust along NUGGET-1 (Funck et al. 2007) are
both in line with gravity anomalies of the Ungava Fault Com-
plex. We therefore propose that stretched and intruded crust/oceanic
crust is present between both lines along the Ungava Fault Com-
plex. The location of the Ungava Fault Complex therefore marks
the plate boundary between Baffin Island and Greenland prior to
57 Ma.
7.2.2 High-velocity lower crust
P-wave velocities of the lower crust higher than 7.0 km s−1are often
interpreted as magmatic underplating (Furlong & Fountain 1986;
Marillier & Reid 1990). Underplating has also been reported on the
nearby lines GR89-WA (Gohl & Smithson 1993) and NUGGET-
1 and -2 (Funck et al. 2007; Gerlings et al. 2009) in Fig. 1. P-
wave velocities of these magmatic underplatings are higher than
the velocities we have modelled on line AWI-20080700 (in the
range of 7.4–7.7 km s−1instead of 6.9–7.4 km s−1). As there is no
boundary detected between lower crust and an underplated body, we
interpret a gradual increase of mafic material from the sides to the
centre of the model. The thickening of the lower crust from a model
distance of 30–100 km shows that mafic material was added to
the lower crust. This is similar to the interpretation of a magmatic
underplating along other profiles (GR89-WA (Gohl & Smithson
1993), NUGGET-1 (Funck et al. 2007) and NUGGET-2 (Gerlings
et al. 2009)).
Lower mantle densities in the free-air gravity model indicate that
the high-velocity lower crust is underlain by a hotter mantle than the
eastern part of line AWI-20080700/BGR08-301. The high content
of mafic material in the centre of the models can be the result of
decompressional mantle melts during extension of the lithosphere
(McKenzie & Bickle 1988) and/or due to the influence of a mantle
plume (White & McKenzie 1989).
Funck et al. (2007) relate the magmatic underplating along
NUGGET line 1 to the Greenland–Iceland mantle plume. Volcanics
of Disco Island are dated to 61 Ma and have been related to the
Iceland plume (Storey et al. 1998). Funck et al. (2007) suggest that
according to the hypothesis of Sleep (1997), buoyant plume mate-
at Alfred Wegener Institut fuer Polar- und Meeresforschung Bibliothek on April 10, 2013http://gji.oxfordjournals.org/Downloaded from

92 S. K. Suckro et al.
rial flowed southwards along thin lithosphere in the central Davis
Strait. Although we cannot confirm the origin of the mafic material
along line AWI-20080700/BGR08-301, it supports the hypothesis
of Funck et al. (2007) that material of the Iceland plume was chan-
nelled southwards along thinned lithosphere in the Davis Strait.
7.2.3 Continental crust, model distance:
0–50 and 95–315 km
We interpret the crust, west of a model distance of 40 km and east of
a model distance of 100 km, as rifted continental crust according to
the P-wave velocity compilation from Rudnick & Fountain (1995)
and the thickness of up to 19 km.
The section from a model distance of 95–135 km is the Davis
Strait High, which crops out farther north. Although the Davis
Strait area was a rifting system prior to 57 Ma (see Section
6), the Davis Strait High is elevated to seafloor level instead
of having subsided. As Chalmers & Pulvertaft (2001) have pro-
posed, this indicates that compressional forces within the Ungava
Fault Complex caused an uplift of continental crust. We suggest
that the presence of buoyant plume material has supported this
uplift.
Steps in the basement morphology indicate faults at a model
distance of 68, 95, 135 and 165 km (Fig. 2). From the P-wave
velocity and density model, we introduced an additional fault at
the western border of continental to intruded/igneous crust at a
model distance of 50 km. The faults from a model distance of
50–135 km are within the transform fault system of the Ungava
Fault Complex (Sørensen 2006) and we therefore interpret them as
transform faults with a normal component (Fig. 10). The fault at
165 km lies at the location of the Hudson Fracture Zone (Chalmers
& Pulvertaft 2001), which is also a transform fault with a normal
component. The continental crust is broken into several segments
that have been uplifted relative to one another and were transported
along transform faults of the Ungava Fault Complex and the Hudson
Fracture Zone.
Model distance 0–50 km: Stretched continental crust of 6–16 km
thickness, divided into upper, middle and lower crust, covered by
basalts intercalated with sediments.
Model distance 50–95 km: Stretched and intruded crust or
new igneous crust with a high-velocity mafic lower crust, covered
by basalts intercalated with sediments and partly by a sediment
basin.
Model distance 95–226 km: Stretched continental crust of 12–
19 km thickness, with a high-velocity lower crust merging into less
intruded lower crust from west to east, covered by sediments and
partly by basalt flows.
7.3 Ungava Fault Complex and Hudson
Fracture Zone
Transform faults of the Ungava Fault Complex are recently derived
by Sørensen (2006) from Bouguer gravity data. Our new models
and data offer new constraints on the location of these faults. We
use regional magnetic anomaly and satellite-derived gravity data to
extend the faults perpendicular to our profile (Fig. 12).
The fault at a model distance of 95 km separates intruded/igneous
crust from the Davis Strait High and matches exactly the location
that Sørensen (2006) proposes (Fig. 12). On our line, the eastern
border of the Davis Strait High lies 14 km east of the location from
Sørensen (2006). We also propose a more north–south striking trend
from the gravity data. The fault that bounds the crust of the Nuuk
Basin to the west (at a model distance of 170 km) is not mapped by
Sørensen (2006). It lies on the Hudson Fracture Zone, which is a
north–south striking fault (Srivastava 1978; Chalmers & Pulvertaft
2001), that is not clearly imaged by the regional potential field data.
While the eastern boundary of the intruded/igneous crust coincides
well with the existing fault map, the western boundary needs to be
shifted eastwards by 40 km. The north–south extent of this fault is
well indicated by a polarity change in the magnetic anomaly data
(our interpretation in Fig. 12b). Furthermore, the fault within the
intruded/igneous crust is well marked by a polarity change. On our
profile, this fault had to be shifted 14 km eastwards relative to the
Sørensen (2006) interpretation.
To determine the role of the Ungava Fault Complex and the
Hudson Fracture Zone in the time between 57–33 Ma, we develop
two-plate tectonic end-member models:
In the first case, we use the Ungava Fault Complex as a plate
boundary and neglect the Hudson Fracture Zone: Although trans-
form forces dominate the Ungava Fault Complex, compressional
forces also occur and must compensate overlapping crust of 70 km
width (Fig. 9d). Evidence for compression is the varying thickness
of the crust along our line. The middle crust of the Davis Strait
High is, for example, 2.5 km thicker than that of the adjacent east-
ern crust (at a model distance of 140–170 km). This can be due
to compression. However, these units may have been transported
to their present position along the Greenland margin via transform
faults of the Ungava Fault Complex, and thus the crustal thickness
does not need to be equal. If a deformation in the scale of 70 km
has occurred, this should also affect the pre-Eocene sediments that
directly overly the basement. Deformed sediments are present at a
model distance of 140–170 km (east the Davis Strait High, closeup
B in Fig. 2). Balancing the bulged sediments only leads to a lat-
eral extension of 0.5 km, which is far from the expected value of
70 km. On the Davis Strait High, there is no sediment cover detected,
which could verify deformations. We conclude that our models im-
age transform faults of the Ungava Fault Complex dividing the
crust, but compression can only have occurred in a scale of a few
kilometres.
In the second case, we use the Hudson Fracture Zone as a plate
boundary and neglect the Ungava Fault Complex: No compressional
forces occur in the Davis Strait area, only strike-slip motion along
the Hudson Fracture Zone connects the opening of the Labrador
Sea and the Baffin Bay (Fig. 9e). Although this model fits the
plate kinematics and the sediment record, some motion must have
occurred along the Ungava Fault Complex which is clearly imaged
by the data we here present and by the regional potential field data.
Given that the poles of rotation from (Oakey 2005) are correct,
the Ungava Fault Complex and the Hudson Fracture Zone must have
been active in the Davis Strait area. We propose that a change took
place from transpression along the Ungava Fault Complex to strike-
slip motion along the Hudson Fracture zone. Prior to 57 Ma, Davis
Strait was opening and highly stretched and intruded continen-
tal crust (line AWI-20080700/BGR08-301) or oceanic-type crust
(NUGGET-1, Funck et al. (2007)) evolved within the location of
the Ungava Fault Complex, which marks the plate boundary at that
time. When the Greenland motion relative to the North American
craton changed to a more northward direction at 57 Ma, transpres-
sion along the Ungava Fault Complex was active as a result of its
relative weak lithosphere. As the northward motion of Greenland
continued, the stress was no longer compensated by the deformation
of crust within the Ungava Fault Complex, but the Hudson Fracture
Zone evolved with pure strike-slip motion. Although the Hudson
at Alfred Wegener Institut fuer Polar- und Meeresforschung Bibliothek on April 10, 2013http://gji.oxfordjournals.org/Downloaded from

Davis Strait crust 93
Figure 12. (a) Locations of seismic refraction lines (AWI-20080700, GR89-WA (Gohl & Smithson 1993), NUGGET-1 (Funck et al. 2007), NUGGET-2
(Gerlings et al. 2009)). On the profiles, fractures and interpretations are marked. (b) Magnetic anomaly data (EMAG2 V2, Maus et al. (2009)) overlain with
the same data as in the upper left panel. Locations of faults of the Ungava Fault Complex after Sørensen (2006), the location of the Hudson Fracture Zone
after Chalmers & Pulvertaft (2001) and our interpretation are marked. (c) Satellite-derived free-air gravity anomalies ((Sandwell & Smith 2009), version 18.1)
overlain with the same information as in the upper right panel. (d) Bouguer gravity anomalies reduced to sea level (DNSC08 free-air gravity data (Andersen
et al. 2008) and Smith & Sandwell (1997) topography, version 13.1, used with code from Fullea et al. 2008) overlain with the same information as in the upper
right panel.
Fracture Zone is not well imaged by the regional gravity data and
has thus often been neglected in the literature, it likely compensated
most of the relative motion between the North American craton and
Greenland. As the crust along the Hudson Fracture Zone was not
deformed with respect to its thickness, it is not indicated by the
regional gravity data.
8 CONCLUSIONS
To determine the nature of the central Davis Strait crust, we devel-
oped a P-wave velocity and a density model, and interpret these
with additional seismic reflection and magnetic anomaly data (Figs.
2, 3 and 8). The profile is dominated by continental crust that is
at Alfred Wegener Institut fuer Polar- und Meeresforschung Bibliothek on April 10, 2013http://gji.oxfordjournals.org/Downloaded from

94 S. K. Suckro et al.
separated by a 45-km-long section of stretched and intruded/new
igneous crust (Fig. 10). It is similar in the P-wave velocity and
density structure to oceanic crust along NUGGET-1 in the northern
Labrador Sea, Fig. 1 (Funck et al. 2007). On both profiles, this
section is divided into an eastern and a western segment by a trans-
form fault of the Ungava Fault Complex. We suggest that oceanic
crust/stretched and intruded crust is also present between both
lines and follows the gravity anomalies that mark the Ungava Fault
Complex (Figs 12c and d). Beneath the intruded/igneous crust lies
a thick high-velocity lower crust (Fig. 10) that can be related to the
Iceland plume which influenced the Davis Strait region in the Pa-
leocene (Lawver & M¨
uller 1994; Storey et al. 1998). We infer that
buoyant plume material was channelled southwards along thinned
lithosphere in the Davis Strait and formed a zone of magmatic un-
derplating in the northern Labrador Sea. Resulting volcanic activity
along the Baffin Island margin is also indicated by basalts flows
along our profile (Fig. 2).
The Davis Strait is dominated by the transform fault system of the
Ungava Fault Complex and the Hudson Fracture Zone. We analysed
the role of both fault systems for the Davis Strait area with plate
kinematic modelling (Fig. 9). While the Davis Strait was opening
prior to 57 Ma, stretched and intruded crust evolved along the lo-
cation of the Ungava Fault Complex, which was the plate boundary
at that time. When the Greenland motion changed to a more north-
ward component, transpressional motion had to be compensated
and the Ungava Fault Complex evolved. Crust was deformed and
transported along transform faults. At some point, compressional
deformation of the crust caused more stress than could be compen-
sated and the Hudson Fracture Zone with pure strike-slip motion
evolved. As this transform fault is not accompanied by crustal thick-
ening or thinning, it is not well represented by the regional potential
field data and has thus not been recognized to the same extent as
the Ungava Fault Complex has. As we only find evidence of mi-
nor compression along our profile, most of the motion between the
North American plate and Greenland at 57–33 Ma must have taken
place along the Hudson Fracture Zone.
ACKNOWLEDGMENTS
We thank the master and crew of RV Merian for their support dur-
ing the cruise. For processing of the MCS data, we thank Ewald
L¨
uschen. Tabea Altenbernd, Martin Block and Sonja Breuer con-
tributed in several discussions to the interpretation of the MCS data.
For providing the OBS to TF via an EU grant in 2008 (contract
RITA-CT-2004505311), we acknowledge Ernst Fl¨
uh from Geomar.
We thank the German Research Council DFG for funding the cruise
MSM09/3. The data analysis and study was financed by institutional
funds of AWI and BGR. We also thank two anonymous reviewers
for improving the manuscript.
REFERENCES
Andersen, O., Knudsen, P., Berry, P.& Kenyon, S., 2008. The DNSC08 ocean
wide altimetry derived gravity anomaly field, talk, EGU-2008, Vienna,
Austria.
Barton, P., 1986. The relationship between seismic velocity and density
in continental crust - a useful constraint?, Geophys. J. R. astr. Soc., 87,
195–208.
Chalmers, J. & Laursen, K., 1995. Labrador Sea: the extent of continental
and oceanic crust and the timing of the onset of seafloor spreading, Mar.
Pet. Geol., 12, 205–217.
Chalmers, J. & Pulvertaft, T., 2001. Development of continental margins
of the Labrador Sea: a review, in Non-Volcanic Rifting of Continental
Margins: A Comparison of Evidence from Land and Sea, Vol. 187, pp.
77–105, eds Wilson, R., Whitmarsh, R., Taylor, B. & Froitzheim, N.,
Geol. Soc. London.
Chalmers, J., Pulvertaft, T., Marcussen, C. & Pedersen, A., 1999. New insight
into the structure of the Nuussuaq Basin, central West Greenland, Mar.
Pet. Geol., 16, 197–224.
Christie, P., Gollifer, I. & Cowper, D., 2006. Borehole seismic studies of
volcanic succession from the Lopra-1/1a borehole in the Faroe Islands,
northern North Atlantic, Geol. Surv. Denmark Greenland Bull., 9, 23–
40.
Clarke, D. & Upton, B., 1971. Tertiary basalts of Baffin Island: field relations
and tectonic setting, Can. J. Earth Sci., 8(248), 248–258.
Dalhoff, F. et al., 2006. Continental crust in the Davis Strait: new evidence
from seabed sampling, Geol. Surv. Denmark Greenland Bull., 10, 33–
36.
Fullea, J., Fern´
andez, M. & Zeyen, H., 2008. FA2BOUG-A FORTRAN 90
code to compute Bouguer gravity anomalies from gridded free-air anoma-
lies: application to the Atlantic-Mediterranean transition zone, Comput.
Geosci., 34(12), 1665–1681.
Funck, T., Gohl, K., Damm, V. & Heyde, I., 2012. Tectonic evolution of
southern Baffin Bay and Davis Strait: results from a seismic refraction
transect between Canada and Greenland, J. geophys. Res., 117(B04107),
doi:10.1029/2011JB009110.
Funck, T., Jackson, H., Louden, K. & Klingelh¨
ofer, F., 2007. Seismic study of
the transform-rifted margin in Davis Strait betweenBaffin Island (Canada)
and Greenland: what happens when a plume meets a transform, J. geophys.
Res., 112(B04402), doi:10.1029/2006JB004308.
Furlong, K. & Fountain, D., 1986. Continental crustal underplating: thermal
considerations and seismic-petrologic consequences, J. geophys. Res.,
91(B8), 8285–8294.
Gerlings, J., Funck, T., Jackson, H., Louden, K. & Klingelh ¨
ofer, F., 2009.
Seismic evidence for plume-derived volcanism during formation of the
continental margin in southern Davis Strait and northern Labrador Sea,
Geophys. J. Int., 176, 980–994.
Gohl, K., Schreckenberger, B. & Funck, T., 2009. The expedition of the
research vessel “Maria S. Merian” to the Davis Strait and Baffin Bay in
2008 (MSM09/3), in Berichte zur Polar- und Meeresforschung (Reports
on Polar and Marine Research), Vol. 587, p. 104, eds Gohl, K., Schreck-
enberger, B. & Funck, T., Alfred Wegener Institute for Polar and Marine
Research, Bremerhaven, ISSN: 1866-3192.
Gohl, K. & Smithson, S., 1993. Structure of Archean crust and passive
margin of southwest Greenland from seismic wide-angle data, J. geophys.
Res., 98(B4), 6623–6638.
Gohl, K. & Uenzelmann-Neben, G., 2001. The crustal role of the Agul-
has Plateau, southwest Indian Ocean: evidence from seismic profiling,
Geophys. J. Int., 144(3), 632–646.
Gradstein, F., Ogg, J. & Smith, A., 2004. A Geological Time Scale 2004,
Cambridge University Press.
Gregersen, U. & Skaarup, N., 2007. A mid-Cretaceous prograding sedimen-
tary complex in the Sisimiut Basin, offshore West Greenland - stratigraphy
and hydrocarbon potential, Mar. Pet. Geol., 24, 15–28.
Holbrook, W. et al., 2001. Mantle thermal structure and active upwelling
during continental breakup in the North Atlantic, Earth planet. Sci. Lett.,
190, 251–266.
Hopper, J., Dahl-Jensen, T., Holbrook, W., Larsen, H., Lizarralde, D.,
Korenaga, J., Kent, G. & Kelemen, P., 2003. Structure of the SE
Greenland margin from seismic reflection and refraction data: implica-
tions for nascent spreading center subsidence and asymmetric crustal
accretion during North Atlantic opening, J. geophys. Res., 108(B5),
doi:10.1029/2002JB001996.
Keen, C. & Barrett, D., 1972. Seismic refraction studies in Baffin Bay:
an example of a developing ocean basin, Geophys. J. R. astr. Soc., 30,
253–271.
Klose, G., Malterre, E., McMillan, N. & Zinkan, C., 1982. Petroleum
exploration offshore southern Baffin Island, Northern Labrador Sea,
Canada, in Arctic Geology and Geophysics: Proceedings of the Third
at Alfred Wegener Institut fuer Polar- und Meeresforschung Bibliothek on April 10, 2013http://gji.oxfordjournals.org/Downloaded from

Davis Strait crust 95
International Symposium on Arctic Geology, pp. 233–244, eds Embry,
A. & Balkwill, H., Canadian Society of Petroleum Geologists, Calgary,
Alberta.
Larsen, L., Rex, D., Watt, W. & Guise, P., 1999. 40Ar-39 Ar dating of alkali
basaltic dykes along the southwest coast of Greenland: Cretaceous and
Tertiary igneous activity along the eastern margin of the Labrador Sea,
Geol. Greenland Surv. Bull., 184, 19–29.
Lawver,L.&M
¨
uller, R., 1994. Iceland hotspot track, Geology, 22, 311–
314.
Mackenzie, G., Thybo, H. & Maguire, P., 2005. Crustal velocity structure
across the Main Ethiopian Rift: results from two-dimensional wide-angle
seismic modelling, Geophys. J. Int., 162(3), 994–1006.
Marillier, F. & Reid, I., 1990. Crustal underplating beneath the carboniferous
Magdalen basin (Eastern Canada): evidence from seismic reflection and
refraction, in The Potential of Deep Seismic Profiling for Hydrocarbon
Exploration, pp. 209–218, eds Pinet, B. & Bois, C., Editions Technip,
Paris.
Maus, S. et al., 2009. EMAG2: A 2-arc min resolution Earth mag-
netic anomaly grid compiled from satellite, airborne, and marine
magnetic measurements, Geochem. Geophys. Geosyst., 10(Q08005),
doi:10.1029/2009GC002471.
McKenzie, D. & Bickle, J., 1988. The volume and composition of melt
generated by extension of the lithosphere, J. Petrol., 29(3), 625–679.
Mjelde, R., Raum, T., Myhren, B., Shimamura, H., Murai, Y., Takanami,
T., Karpuz, R. & Næss, U., 2005. Continent-ocean transition on
the Vøring Plateau, NE Atlantic, derived from densely sampled
ocean bottom seismometer data, J. geophys. Res., 110(B05101),
doi:10.1029/2004JB003026.
Oakey, G., 2005. Cenozoic evolution and lithosphere dynamics of the Baffin
Bay-Nares Strait region of Arctic Canada and Greenland, PhD thesis,
Vrije Universiteit, Amsterdam.
Olesen, O. et al., 2007. An improved tectonic model for the Eocene opening
of the Norwegian-Greenland Sea: use of modern magnetic data, Mar.
Pet.Geol., 24, 56–66.
Pedersen, A., Larsen, L., Pedersen, G. & Dueholm, K., 2006. Five slices
through the Nuussuaq Basin, West Greenland, Geol. Surv. Denmark Bull.,
10, 53–56.
Rice, P. & Shade, B., 1982. Reflection seismic interpretation and seafloor
spreading history of Baffin Bay, in Arctic Geology and Geophysics: Pro-
ceedings of the Third International Symposium on Arctic Geology, pp.
245–265, eds Embry, A. & Balkwill, H., Canadian Society of Petroleum
Geologists, Calgary, Alberta.
Roest, W. & Srivastava, S., 1989. Sea-floor spreading in the Labrador Sea:
a new reconstruction, Geology, 17, 1000–1003.
Rudnick, R. & Fountain, D., 1995. Nature and composition of the con-
tinental crust: a lower crustal perspective, Rev. Geophys., 33(3), 267–
309.
Sandwell, D. & Smith, W., 2009. Global marine gravity from retracked
Geosat and ERS-1 altimetry: ridge segmentation versus spreading rate, J.
geophys. Res., 114(B01411), doi:10.1029/2008JB006008.
Skaarup, N., Jackson, H. & Oakey, G., 2006. Margin segmentation of Baffin
Bay/Davis Strait, eastern Canada based on seismic reflection and potential
field data, Mar. Pet. Geol., 23, 127–144.
Sleep, N., 1997. Lateral flow and ponding of starting plume material,
J. geophys. Res., 102(B5), 10 001–10 012.
Smith, W. & Sandwell, D., 1997. Global seafloor topography from satellite
altimetry and ship depth soundings, Science, 277, 1957–1962.
Sørensen, A., 2006. Stratigraphy, structure and petroleum potential of the
Lady Franklin and Maniitsoq Basins, offshore southern West Greenland,
Pet. Geosci., 12, 221–234.
Srivastava, S., 1978. Evolution of the Labrador Sea and its bearing on the
early evolution of the North Atlantic, Geophys. J. R. astr. Soc., 52, 313–
357.
Srivastava, S., MacLean, B., Macnab, R. & Jackson, H., 1982. Davis Strait:
structure and evolution as obtained from a systematic geophysical survey,
in Arctic Geology and Geophysics: Proceedings of the Third International
Symposium on Arctic Geology, pp. 267–278, eds Embry, A. & Balkwill,
H., Canadian Society of Petroleum Geologists, Calgary, Alberta.
Storey, M., Duncan, R., Larsen, A. & Larsen, H., 1998. 40Ar/39 Ar
geochronoly of the West Greenland Tertiary volcanic province, Earth
planet. Sci. Lett., 160, 569–586.
Suckro, S. et al., 2012. The crustal structure of southern Baffin Bay: im-
plications from a seismic refraction experiment, Geophys. J. Int., 190(1),
37–58.
Talwani, M. & Eldholm, O., 1977. Evolution of the Norwegian-Greenland
Sea, Geol. Soc. Am. Bull., 88, 969–999.
Tessensohn, F. & Piepjohn, K., 2000. Eocene compressive deformation
in Arctic Canada, North Greenland and Svalbard and its plate tectonic
causes, Polarforschung, 68, 121–124.
Verhoef, J., Roest, W., Macnab, R. & Arkani-Hamed, J., 1996. Magnetic
anomalies of the Arctic and North Atlantic oceans and adjacent land
areas, in Open File Report 3125A, Geological Survey of Canada, Calgary,
Canada.
Voss, M. & Jokat, W., 2007. Continent-ocean transition and voluminous
magmatic underplating derived from p-wave velocity modelling of the
east Greenland continental margin, Geophys. J. Int., 170(2), 580–604.
White, R. & McKenzie, D., 1989. Magmatism at rift zones: the generation of
volcanic continental margins and flood basalts, J. geophys. Res., 94(B6),
7685–7729.
White, R., McKenzie, D. & O’Nions, R.K., 1992. Oceanic crustal thickness
from seismic measurements and rare Earth element inversion, J. geophys.
Res., 97(B13), 19 683–19 715.
Zelt, C. & Smith, R., 1992. Seismic traveltime inversion for 2-D crustal
velocity structure, Geophys. J. Int., 108, 16–34.
at Alfred Wegener Institut fuer Polar- und Meeresforschung Bibliothek on April 10, 2013http://gji.oxfordjournals.org/Downloaded from

96 S. K. Suckro et al.
APPENDIX A: RAYTRACING IN THE P-WAVE VELOCITY MODEL FOR ALL OBS
Figure A1. Raytracing in the P-wave velocity model for OBS 1–6. (Top panels) Picked phases in red with vertical bar length according to the assigned pick
uncertainty; calculated traveltimes as thin black lines; phase names are annotated; a reduction velocity of 6 km s−1is used. (Lower panels) Raypaths of the
corresponding phases in the P-wave velocity model. For clarity, only every 10th ray is plotted.
at Alfred Wegener Institut fuer Polar- und Meeresforschung Bibliothek on April 10, 2013http://gji.oxfordjournals.org/Downloaded from

Davis Strait crust 97
Figure A2. Raytracing in the P-wave velocity model for OBS 7–12. (Top panels) Picked phases in red with vertical bar length according to the assigned pick
uncertainty; calculated traveltimes as thin black lines; phase names are annotated; a reduction velocity of 6 km s−1is used. (Lower panels) Raypaths of the
corresponding phases in the P-wave velocity model. For clarity, only every 10th ray is plotted.
at Alfred Wegener Institut fuer Polar- und Meeresforschung Bibliothek on April 10, 2013http://gji.oxfordjournals.org/Downloaded from