Tectonophysics 580 (2012) 88–99
Contents lists available at SciVerse ScienceDirect
Tectonophysics
j ourna l homepage: www.e lsev ie r .com/ locate / tectoA new interpretation of the sedimentary cover in the western Siljan Ring area, central
Sweden, based on seismic data
Christopher Juhlin a,⁎, Erik Sturkell b, Jan Ove R. Ebbestad c, Oliver Lehnert d,
Anette E.S. Högström e, Guido Meinhold f
a Department of Earth Sciences–Geophysics, Uppsala University, Villavägen 16, SE‐752 36 Uppsala, Sweden
b Department of Earth Sciences, University of Gothenburg, PO Box 460, SE-405 30 Göteborg, Sweden
c Museum of Evolution, Uppsala University, Norbyvägen 16, SE‐752 36 Uppsala, Sweden
d GeoZentrum Nordbayern, Universität Erlangen, Schloßgarten 5, D-91054 Erlangen, Germany
e Tromsø University Museum, Natural Sciences, N-9037 Tromsø, Norway
f Sedimentology and Environmental Geology, Geoscience Centre, University of Göttingen, Goldschmidtstrasse 3, D-37077 Göttingen, Germany⁎ Corresponding author. Tel.: +46 18 4712392; fax: +
E-mail address: Christopher.Juhlin@geo.uu.se (C. Juh
0040-1951 © 2012 Elsevier B.V.
http://dx.doi.org/10.1016/j.tecto.2012.08.040
Open access under CC BYa b s t r a c ta r t i c l e i n f oArticle history:
Received 22 May 2012
Received in revised form 14 August 2012
Accepted 27 August 2012
Available online 1 September 2012
Keywords:
Impact crater
Silurian
Ordovician
Reflection seismic
Refraction seismic
BoreholeTwo new reflection seismic profiles over the Paleozoic successions of the western part of the Siljan Ring impact
structure show a contrasting seismic signature. The more southerly c. 10 km long Mora profile reveals a highly
disturbed structure, with only a few kilometers of relatively horizontally layered structures observed. However,
interpretations of refracted arrivals in the data, that can be correlated to reflections, indicate the Silurian clastic
rocks to be about 200 m thick in the central part of the profile.Weak reflections from about 600 mdepth suggest
a 400 m thick Ordovician limestone sequence to be present. Cores from the area show a mainly shale lithology
for the Silurian and only a thin sequence of Ordovician strata, suggesting a rapid thickening of the Ordovician
towards the north. On the more northern c. 12 km Orsa profile clear reflections from the Paleozoic successions
are seen along the entire profile, except on the southernmost few kilometers. Based on interpretations of
refracted arrivals, the Silurian succession appears to be considerably thinner here, and possibly absent at some
locations. The Ordovician is also interpreted to be thinner in this area, with a maximum thickness of about
200–300 m along most of the profile. A deeper reflection from about 2 km within the crystalline basement
may represent a dolerite sill. The lack of clear basement reflections on the Mora profile can be attributed to
near-surface conditions and the acquisition geometry. The seismic data and recent coring in the area suggest
the presence of a deeper paleo-basin towards the southwest with significantly more shales being deposited
and the Paleozoic successions being severely disturbed. The shallow coring and seismic data will help form the
basis for locating future boreholes for deeper drilling to study impact processes and the Paleozoic evolution of
central Sweden.
© 2012 Elsevier B.V. Open access under CC BY-NC-ND license.1. Introduction
During the summer of 2011 two reflection seismic profiles were
acquired in the western part of the Siljan impact structure (Fig. 1). The
work was conducted under the framework of the project ‘Concentric
Impact Structures in the Paleozoic’ (CISP), which aims to study impact
mechanics, tectonics and the paleobiology of the Siljan and Lockne
impact structures. CISP is financed 2010–2012 by the Swedish Research
Council (VR) as one of two currently VR funded projects that are part of
the Swedish Deep Drilling Program (SDDP; Lorenz, 2010). To date, the
overall objective for the CISP studies has been to establish reliable geo-
logical models for the planned deeper drilling in the Siljan and Lockne46 18 501 110.
lin).
-NC-ND license.impact structures under the SDDP umbrella; for an outline of detailed
scientific objectives see Högström et al. (2010).
Acquisition of new reflection seismic data has been the first step in
defining the drilling locations in the Siljan impact crater. Previously
acquired seismic data in the area were obtained in conjunction with
the search for abiogenic methane in the 1980s and early 1990s (Bodén
and Erikson, 1988). A total of nine lines were acquired, mostly in the
eastern and northern parts of the Siljan Ring area, but one of them,
Line 9, extended from the central part of the crater nearly to the town
of Mora (Fig. 1). These profiles show the upper crustal Transcandinavian
Igneous Belt (TIB) granitic rock in the area to be fairly transparent, but
with strong high amplitude sub-horizontal reflections at distinct levels
in the upper 10 km. Deep drilling within the crater rim revealed these
high amplitude reflections to be generated by relatively thin (5–50 m)
dolerite sills (Juhlin, 1990; Papasikas and Juhlin, 1997). A deeper reflec-
tive zone is present at about 12–15 km throughout the area (Juhojuntti
Precambrian bedrock
0 10 km
TownsLate Silurian–Devonian
sediments (Orsa Sandstone)
Ordovician sediments
Igrene drill sitesKullsberg Mounds
Silurian sediments
Legend
Transpression ridges
Siljan
SWEDEN
600km
N
Solberga 1
Boda Mounds
Stumsnäs 1
Mora 1
N
Lake
Ore
Lake
Skattungen
Mora
Orsa
Lake
Orsa
Rättvik
Lake Siljan
Key sections
Amtjärn
Fjäcka
Kårgärde
Or
sa 
pro
file
Mora profile
Seismic profiles
Styggforsen
Skattungbyn
Fig. 1. Location map of the Siljan Ring, showing distribution of sedimentary rocks, key geological sections, the drill sites for cores taken by Igrene AB, the position of the new seismic
profiles, and the transpression ridges of Kenkmann and von Dalwigk (2000a,b).
Modified from Ebbestad and Högström (2007).
89C. Juhlin et al. / Tectonophysics 580 (2012) 88–99and Juhlin, 1998) thatmayhave adifferent origin. North of the SiljanRing,
where Svecofennian Domain rocks are present at the surface, relatively
strong lower crustal reflectivity is observed with the Moho interpreted
to be at about 45 km (Juhojuntti and Juhlin, 1998).
In this paper we present results from two new seismic profiles that
mainly pass over the Paleozoic sedimentary successions of the western
part of the Siljan Ring. One profile, about 10 km long and striking in
the W–E direction, was located just north of the town of Mora and
extends onto the crystalline basement at its eastern end (Fig. 1). The
other, about 12 km long and striking SW–NE, was located north of the
town of Orsa and traversed only Paleozoic rocks (Fig. 1). Objectives of
the profiles were to map the thickness of the sedimentary rocks and
their internal structure, and if possible, identify structures in the crystal-
line basement below. Results from the profiles will be used as a starting
point for defining future deep drilling locations in the area.2. Geological setting
The Siljan structure was formed by a bolide that impacted the
terrestrial landscape of central Sweden in the late Devonian (latest
Frasnian; 377 Ma, Reimold et al., 2004, 2005). The structure is con-
centric with a central rise composed of Proterozoic magmatic rocks
or younger Proterozoic meta-sediments and meta-volcanic rocks
(western side) across a 33 km wide area in the east–west direction.
The obvious circular structure outside the central rise (here called
the ring) is outlined mainly by a ring of lakes and consists largely of
marine Ordovician and Silurian sediments and Silurian to Devonian
terrestrial sandstones, extending the east–west diameter to about
45 km (Fig. 1). It is northern Europe's largest crater, and may have
played a casual role in the Fransnian/Famennian mass extinction
(Reimold et al., 2004).
90 C. Juhlin et al. / Tectonophysics 580 (2012) 88–99The transient crater that formed at impact may have been anywhere
between 22 and 42 km (Grieve, 1988; Holm et al., 2011; Juhlin and
Pedersen, 1987, 1993; Kenkmann and von Dalwigk, 2000a). The pre-
served central risemay represent the peak ring (Grieve, 1988), although
Henkel and Aaro (2005) suggested that the peak ring was 20 km wide.
The final diameter of the crater may have been anywhere between 52
and 90 km (Grieve, 1988; Henkel and Aaro, 2005; Holm et al., 2011;
Kenkmann and von Dalwigk, 2000a,b) (Fig. 2).
The ring structure itself has been denoted as an annular ring graben,
a mega-block zone, or a mega-breccia (Collini, 1988; Henkel and Aaro,
2005; Juhlin and Pedersen, 1987).Within themega-blocks the Paleozoic
successions are partly well preserved, albeit poorly exposed, and the
stratigraphy is often continuous. It is clear that the mega-blocks now
preserved in the ring could not have been within the transient crater,
but rather represent radial inflow of the collapsed rim of the transient
crater in concert with mass flow from the central area (Henkel and
Aaro, 2005). A large transient crater would imply significant horizontal
displacements of the mega-blocks (6–7 km; Kenkmann and von
Dalwigk, 2000b), while a smaller crater would imply primarily vertical
movements of blocks through localized sagging (listric faulting), which
is what the reflection seismic data show in this paper, as well as in
Juhlin and Pedersen (1987, 1993).Lake
Ore
Lake
Orsa
Lake Siljan
Rättvik
0 10 km
Stumsnäs 1
Lake
re
Lake Siljan
52 km A
Stumsnäs 1
Lake
Ore
Lake
Orsa
Lake Siljan
75 km C
Fig. 2. Crater models developed for the Siljan impact structure. The entire circles denote the
final crater. The center of the crater is based on estimates by Juhlin and Pedersen (1987). (
Estimates by Grieve (1988) and Juhlin and Pedersen (1993) based on seismic studies. (b) A
(2000a) who delimited the size from distribution of fractures in the crystalline bedrock in
showing the estimated peak ring. Based on geomorphological data by Henkel and Aaro (
size of 90 km. Based on the 32–38 km estimate of the transient crater and the maximum fi
features.Even though the reflection seismic data indicate a relatively simple
structure, at the scale of the resolution of the data, there are geological
indications of more complex structures in the area. Outcrops in the
Siljan Ring are few, but the heavily tectonized nature of the sedimentary
successions is often evident, generally with sharply inclined or nearly
overturned packages of crystalline basement and/or sediments. One of
the more important sedimentary features in the area is the large Late
Ordovician Boda Limestonemudmounds, which are quarried commer-
cially. Several of these are steeply inclined and all are heavily fractured
(mega-brecciated; Henkel and Aaro, 2005) as a result of immediate
post-impact disturbances.
Recent investigations of new drill cores from the Siljan Ring have dra-
matically changed the established stratigraphic view of the successions
within the ring (Lehnert et al., 2012). Instead of the classical Ordovician
carbonate succession separated by a stratigraphic gap from the Silurian
shale dominated sequence, it is now clear that several facies belts exist.
During the Middle–Late Ordovician there was a carbonate platform suc-
cession towards the present day east, separated by a backbulge basin
from the peripheral to the present day west in the Mora area. In Mora
an erosional unconformity is observed, showing a considerable hiatus
spanning the early Darriwilian to the upper Llandovery (preliminary
dating; Lehnert et al., 2012). About 224 m of Silurian strata overlayLake
Ore
Lake
Orsa
Lake Siljan
Rättvik
Stumsnäs 1
Mora 1
Lake
re
Lake Siljan
65 km B
Stumsnäs 1
Lake
Ore
Lake
Orsa
Lake Siljan
90 km D
diameter of the transient crater, while the stippled circle denotes the diameter of the
a) A final crater measuring 52 km and pre-erosion transient crater measuring 26 km.
42 km transient crater and a final size of 65 km. Based on Kenkmann and von Dalwigk
terpreted as impact‐related. (c) A final crater of at least 75 km, with the inner ring he
2005). (d) A transient crater with a pre-erosion mean diameter of 35 km and a final
nal diameter suggested by Holm et al. (2011), from distribution of shock metamorphic
Table 1
Data acquisition.
Mora Orsa
Spread type Split
Number of channels 320 (160–160)
Near offset 0 m
Geophone spacing 10 m
Geophone type 28 Hz single
Source spacing 20 m
Source type VIBSIST 3000
Hit interval between hammer blows 100–200 ms
Sweeps per source point 3–4
Nominal fold 80
Recording instrument SERCEL 428UL
Sample rate 1 ms
Field low cut Out
Field high cut 400 Hz
Record length 30 s
Profile length 10 km 12 km
Source points 391 542
Dates acquired 3/6–9/6, 2011 10/6–15/6, 2011
91C. Juhlin et al. / Tectonophysics 580 (2012) 88–99here a thin Lower–Middle Ordovician carbonate succession (about 21 m)
in theMora 001 core (Fig. 1), drilled and owned by Igrene AB. There is no
evident contact to the Orsa Sandstone or equivalent strata as is consistent
with the partly coeval Nederberga Formation in the Skattungbyn area in
the northern part of the ring structure (Fig. 1). The age of the uppermost
preserved strata exposed at Mora is yet to be determined.
Earlier acquired reflection seismic data show approximately 350 m of
Paleozoic sediments preserved above the crystalline basement in the
eastern part of the ring (Collini, 1988), and Thorslund (in Rondot, 1975)
estimated a total of 400–500 m of these sediments to be present.
Where exposed, the sedimentary successions startwith the Tremadocian,
although Cambrian sediments may also be present in the area (Ebbestad
and Högström, 2007). There are no continuous sections in the Siljan area,
so the stratigraphy has been pieced together froma few key sections such
as Kårgärde near the town of Orsa, Fjäcka on the northeast side of the
ring, Amtjärn near the town of Rättvik, Styggforsen near Boda Church
and the new Igrene drill cores (see Fig. 1). Exposures of the Ordovician–
Silurian boundary are seen in a few quarries where the large Boda Lime-
stone mud mounds are exposed. The development of strata in between
the larger mounds and the Silurian succession has not been well under-
stood, and the transition to the terrestrial Silurian/Devonian Orsa Sand-
stone is only seen in a few drill cores (Petalas, 1985). As a consequence,
the total thickness of the Paleozoic succession in the Siljan Ring is not
known and, with respect to distinct facies belts, may have been variable
depending on the different paleo-environments. Regardless, the Siljan
area has become the standard reference area for a number of Ordovician
and Silurian type formations, units and/or biozones in Sweden (see
Bergström, 2007).
Rough calculations of sediment thickness usingmeasurements quot-
ed for each unit in Ebbestad and Högström (2007) and Lehnert et al. (in
press) suggest that the classic Ordovician succession is at the minimum
115 m thick and more than 133 m thick, albeit here excluding the
~35 m thick Kullsberg mounds and the ~100 m thick Boda mounds
which formed during Late Ordovician times on the outer platform.
Adding the thickness of the Boda Limestone (e.g. thickness of all beds
from the Obolus beds to the Fjäcka Shale plus Boda Limestone), gives
an estimate of between 185 and 215 m for the Ordovician sediments.
In the Mora 001 core, the Silurian succession is at least 224 m thick
(Lehnert et al., 2012), although it does not reflect the depositional thick-
ness over the entire Siljan area. The thickest succession of Orsa Sand-
stone known is at least 58.9 m (Petalas, 1985). Adding up these highly
uncertain estimates gives a minimum thickness of the Paleozoic sedi-
ments in the order of 460 to 500 m.
3. Data acquisition
Seismic data (Table 1) were acquired along two profiles in the west-
ern part of the Siljan Ring area in the summer of 2011 (Fig. 1). Since ob-
jectives were to both image the Paleozoic sedimentary rocks and the
deeper crystalline basement (down to 3 km) along longer profiles a
compromise between station spacing and profile length was necessary.
In addition, it was expected that signal penetration could be an issue at
some locations due to large amounts of sand in the near-surface. There-
fore, a relatively powerful sourcewith a source spacing of 20 m and a re-
ceiver spacing of 10 m was employed. If the Paleozoic rocks extend
down to 500 m then this source and receiver spacing corresponds to
an effective fold of about 25 at this depth with a normal moveout
(NMO) stretch of about 50%. At 250 m depth the effective fold is half
of this. This lower fold at shallower depth implies that degraded images
in those parts of the final stacked sectionwhere the source couplingwas
poor may be expected.
Datawere first acquired along theMora profile during a 6 day period
in the beginning of June 2011 using a tractor mountedmechanical ham-
mer (see Juhlin et al. 2010 for details on the source). Acquisition started
on the eastern end of the profile on granitic bedrock and proceeded
westwards. It was not possible to lay cables across the outlet to LakeOrsa (stations 1016–1035 in Fig. 3) so the acquisition geometry had to
be adjusted in this area. A cabled fixed spreadwas used east of the outlet
with 36 wireless units west of the outlet while the source was active on
the eastern part.When all possible source points had been activated east
of the outlet the cabled spread was moved to the west of the outlet and
the 36 wireless units to the east of it. This strategy allowed us to acquire
data to fill the gap below the outlet without having to repeat source
points and avoided us having to transport the source back and forth
through the town of Mora. In addition to the logistics with the outlet,
there were many locations along the Mora profile where the source
could not be activated due to the near-surface being too soft or that
the profile ran too close to residential areas. These locationsweremostly
on the eastern half of theprofile (Fig. 3). Surface deposits east of the Lake
Orsa outlet consist mainly of moraine while west of the outlet the sur-
face deposits are mainly glacial and fluvial sands.
After completing the Mora profile the spreadwas moved to the Orsa
profile (Fig. 4) and acquisition continued using the same mechanical
source. Logistics were less problematic along the Orsa profile, but
there were some locations in the southern part of the profile where it
was not possible to activate the source (Fig. 4). South of station 350 sur-
face deposits are mainly fluvial sands while to the north of it glacial
deposits dominate. In addition to the cabled spread that was rolled
along during the acquisition, twelve 3-component wireless units were
placed out along the profile at a spacing of 1 km. These units recorded
nearly all of the source points during the acquisition, but presentation
of data from these units is not the focus of this paper. However, these
data may be useful in the future for building velocity models based on
refracted waves.
4. Data processing
Datawere decoded using a shift and stackmethod based on Park et al.
(1996). Data quality varied significantly along the profiles depending
upon the near surface conditions. The thickness of the uppermost loose
deposits varies considerably in the area, as does the groundwater level. In-
terestingly, the depth to the groundwater level appeared to have less
influence on the data quality than may have been expected. First break
quality is nearly the same regardless of how much dry sand is present
(Fig. 5). The thickness of the groundwater bearing loose sediments
appears to be a more important factor influencing data quality. Where
these loose sediments are thick a very ‘ringy’ signal after the first arrival
is present. This signal masks weaker reflections and multiple reflections
are generated that arrive within a cone that is limited by a velocity of
about 1700 m/s. Within this cone it is difficult, if not impossible, to trace
reflections with the receiver spacing used. Any reflections arriving at
late times will be masked by this cone. When this groundwater-bearing
470 480
6760
6770
W-E (km)
N
-S
 (k
m)
Silurian clastics
Ordovician carbonates
Neoproterozoic dolerites
Precambrian felsic intrusives
Precambrian felsic extrusives
Precambrian sedimentary rocks
Fig. 3. Mora profile location. Blue line indicates where the geophones were placed. When filled with red then the source was activated at those locations along the profile. Yellow
line shows the CDP processing line that the data were stacked along. Geological background map provided by the Geological Survey of Sweden (www.sgu.se).
92 C. Juhlin et al. / Tectonophysics 580 (2012) 88–99layer of loose sediments is thinner or absent this cone is less prevalent
(Fig. 5a, left side of spread) and reflections are more easily observed in
raw source gathers. The thickness of this problematic layer varies consid-
erably along theMora profile, while along theOrsa profile this layer is less
of a problem and appears to be presentmainly in the southern part of the
profilewherefluvial sediments are present at the surface (Fig. 6a). Further
north along the Orsa profile the data quality is generally good with clear
first arrivals across the entire spread (Fig. 6b). Note how the apparent ve-
locity increases at the further offsets in Figs. 5a and 6b, indicating higher
velocity rock at depth.
The variable data quality along theprofiles, especially theMoraprofile,
alongwith variable velocities, both vertically and horizontally, was a pro-
cessing challenge. It is difficult to estimate the velocity in the dry loose de-
posits since the direct wave through them is only recorded on the closest
geophone or two, but based on these arrivals a velocity of 600 m/s is rea-
sonable for these deposits. When water saturated, these loose deposits
have a velocity on the order of 1700 m/s. The first main refractor that is
observed has a velocity of about 3000–3200 m/s, corresponding to the
uppermost Paleozoic rocks. A deeper refractor is sometimes observed
with a velocity on the order of 4000–4200 m/s that probably represents
a deeper succession of Paleozoic rocks. East of the Lake Orsa outlet on
the Mora profile, refractor velocities of about 5000 m/s are observed.
Bothprofiles followed roads and forest tracks, resulting in crooked line
data being acquired. One option for handling such data is to define a sla-
lom stacking line that follows the general trend of the mid-points. How-
ever, this strategy can result in false structures being introduced into
the final stacks where the profile changes direction rapidly (e.g. Wu etal., 1995). Instead, straight stacking lines were used with Common Data
Point (CDP) bins defined perpendicular to the stacking line with the
true acquisition geometry still being honored. That is, it was not assumed
that the data had been acquired along straight lines. Since the stacking (or
CDP) lines are straight they can be shifted at will in the direction normal
to them and the same result will be obtained in the processing. Therefore,
the location of the stacking lines in Figs. 3 and 4 is only schematic, and
they may be shifted as explained above without affecting the processing
results.
Mostly standard processing steps were applied to the data after CDP
binning. Due to that most of the reflections are observed best at wide
angles a large normal NMO stretch was allowed. This large stretch re-
duces the apparent frequency content of the data in the stacked sec-
tions. However, if the maximum stretch was decreased to a more
normal value of 40% then some reflections were completely muted in
the source gathers and no additional reflections appeared in the stacks.
Increasing the allowable stretch to above 70% did not only increase am-
plitudes in the stacked sections, but also decreased the frequency con-
tent even more. A stretch value of 70% in the NMO processing step
therefore seemed to be a reasonable compromise. Due to the often
‘ringy’ nature of the data in the Mora profile and the more prevalent
noise cone, both a top mute and a bottom mute were applied before
stacking. Top muting was to ensure that first arrivals were not stacked
in as reflected energy. Bottom muting was applied to reduce the influ-
ence of the noise cone in the uppermost 300 ms of the stacked section.
A complete list of processing steps for the two profiles is provided in
Table 2.
Silurian clastics
Ordovician carbonates
Neoproterozoic dolerites
Precambrian felsic intrusives
Precambrian felsic extrusives
Precambrian sedimentary rocks
W-E (km)
N
-S
 (k
m)
480
6780
Fig. 4. Orsa profile location. Blue line indicates where the geophones were placed. When filled with red then the source was activated at those locations along the profile. Yellow line
shows the CDP processing line that the data were stacked along. Geological background map provided by the Geological Survey of Sweden (www.sgu.se).
93C. Juhlin et al. / Tectonophysics 580 (2012) 88–995. Results
Stacked and migrated sections are shown for the Mora and Orsa
profiles in Figs. 7 and 8, respectively.
The Mora profile is characterized by reflective packages of variable
dip with the only longer clear sub-horizontal reflection being present
in the west between CDPs 350 and 700 at an approximate time of
200 ms (c. 250 m). The rest of the profile either shows a transparent
seismic response, reflections with either a predominantly easterly or
westerly dip, or shorter packages of sub-horizontal reflections (Fig. 7).
There are no clear reflections from below 500 ms. Note the generally
higher frequency content in the data east of CDP 1500 where the bed-
rock consists of Ordovician carbonate rocks and Precambrian crystalline
rocks. Dipping events in the upper 400 ms, with opposite dips, between
CDPs 500 and 700 may be from real steeply dipping structures since
they have apparent velocities of about 3500 m/s. However, they may
also result from direct or surfacewaves that have not been removed en-
tirely during the processing. The more moderately, mainly east dipping
events, between CDPs 800 and 1000 probably represent a sub-surface
structure that has been significantly disturbed.
The Orsa profile shows a more consistent image with mainly sub-
horizontal reflections present in the uppermost 400 ms throughout
the profile (Fig. 8). Only in the south are transparent zones present, be-
tween CDPs 100 to 300 and 550 to 650. Dipping reflections are most ap-
parent between CDPs 300 and 500 in the south and in the northernmostpart of the profile around CDP 2200. A deeper sub-horizontal reflective
zone is present in the central parts of the profile between 800 and
900 ms. There are also indications of dipping reflections above this
sub-horizontal reflectivity. Evidence for faulting is present throughout
most of the section. Faults with minor throws are obvious between
CDPs 1250 and 1700 in the upper 200 ms. Further north, between CDPs
1750 and 2000, a graben-like structure is present in the upper 400 ms
containing faults with throws on the order of 100 ms (200 m).
6. Seismic modeling
Based on the refractions observed in the source gathers and the pres-
ence of wide-angle reflections there are clear velocity contrasts in the
sub-surface along the profiles that can produce reflections. However,
the stacked and migrated images show a complex and sometimes spo-
radic reflectivity pattern that is difficult to interpret based on geometry
alone. It is not even clear if the crystalline Precambrian basement has
been imaged along the profiles when only inspecting the stacked sec-
tions. To aid in the interpretation, CDP super gathers of the raw data
were generated to allow velocity as a function of depth (time) to be
estimated along the profiles (Fig. 9). Data from five adjacent CDPs
were binned at an offset interval of 10 m and then stacked within com-
mon offset bins so that only one trace is presented for each offset in the
super gathers. Onmost super gathers there are clear breaks in the slope
of the first arrivals with the slope decreasing with offset, indicating the
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
Ti
m
e 
(s)
700 750 800 850 900 950 1000 1050
Station
SP842
300 350 400 450 500 550 600
Station
SP460
Fig. 5. (left panel) Source gather from SP 842 on the Mora profile (Fig. 3) showing a relatively small delay in the zero crossing of the main first break energy which indicates that the
slow dry near-surface loose deposits are relatively thin at this source location. (right panel) Source gather from SP 460 on the Mora profile (Fig. 3) showing a longer delay in the zero
crossing, a ringy signal after the first arrival and a cone multiple noise. The latter two characteristics of the data are probably due to a relatively thick groundwater layer at this
location which energy gets trapped in.
94 C. Juhlin et al. / Tectonophysics 580 (2012) 88–99arrival of refractions from deeper layers. The arrival times were then
modeled for every 50th CDP assuming a horizontally layered sub-
surface and increasing velocity with depth to generate velocity-depth
models along the profiles. The assumption of a horizontally layered
sub-surface is true along much of the profiles, but is obviously violated
on some parts. Dipping layers and rapid lateral variations in velocity can
significantly perturb the arrival times compared to the 1D assumption.0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
Ti
m
e 
(s)
50 100 150 200
Station
SP30
400
Fig. 6. (left panel) Source gather from SP 30 on the Orsa profile (Fig. 4) showing a moderat
arrivals and some signs of a noise cone due to trapped energy in the groundwater layer. (ri
sponse from the more northern part of the profile. Here, the ringyness is less prevalent andHowever, the velocity-depth models do show general consistency
along the profiles and aid in the interpretation of the reflection seismic
images.
As a further check on the modeling, synthetic seismograms were
generated based on the obtained velocity-depth models using an
elastic modeling program (Juhlin, 1995). Poisson's ratio was assumed
to be on the order of 0.33 and the density increasing with increasing450 500 550 600 650 700
Station
SP548
e delay in the zero crossing of the main first break energy, a ‘ringy’ signal after the first
ght panel) Source gather from SP 460 on the Orsa profile (Fig. 4) showing a typical re-
the multiple noise cone is generally absent.
Table 2
Data processing.
Step Parameters
Mora Orsa
1 Read decoded VIBSIST data
2 Bulk static shift to zero time
3 Apply geometry
4 Pick first breaks
5 Spherical divergence correction
6 Trace editing
7 Trace balance: 0–3000 ms
8 Spectral equalization:
0–600 ms: 50–80–200–240 Hz
700–1500 ms: 40–70–180–240 Hz
9 Time variant bandpass filter:
0–400 ms: 50–80–240–360 Hz
450–600 ms: 45–70–210–300 Hz
700–1000 ms: 40–60–180–270 Hz
1100–3000 ms: 35–50–150–225 Hz
10 Refraction statics: datum 160 m,
replacement velocity
3000 m/s
Refraction statics: datum 180 m,
replacement velocity 3000 m/s
11 Residual statics
12 Median filter: 11 traces, 3
samples, 5300 m/s, subtract
Median filter: 11 traces, 3 samples, 1600,
2000, 2400 m/s, subtract
13 AGC: 50 ms
14 Mute: top and bottom No mute
15 Residual statics
16 Velocity analysis
17 NMO correction: 70% stretch mute
18 Trace balance
19 FX Decon: 19 trace window
20 No dip filter Dip filter 1.5 ms/trace cutoff
21 Trace balance
22 Stolt migration: 200–3000, 1000–4000 ms/s
95C. Juhlin et al. / Tectonophysics 580 (2012) 88–99P-wave velocity for the modeling. Fig. 9a shows a CDP super gather at
CDP 600 on the Mora profile (Fig. 7), a location where the structure
appears to be relatively sub-horizontal. The first arrivals calculated0 2 4
Distanc
0
1
Ti
m
e 
(s)
Ti
m
e 
(s)
200 400 600 800 100
CD
W
0 2 4
Distan
0
1
200 400 600 800 10
CD
W
a
b
Fig. 7. Mora profile showing the (a) stfrom the velocity-depth model (Table 3) match well with the observed
first arrivals (Fig. 9a). Note that the low velocity unsaturated loose de-
posits (layer 1 in Table 3) appear to be relatively thick here (17 m),
but that signals are still clear out to 1500 m offsets in the real data.
Most of the general characteristics of the modeled data agree well
with observations in the real data. In particular, the strong multiples,
both reflected and refracted, and the noise cone due to the water satu-
rated loose sediments (layer 2 in Table 3) are similar in the two sections
(Fig. 9a). One discrepancy is the reflected wave from the base of layer 3
that arrives earlier in themodeled data than in the real data. The reason
for this is not clear, but could be related to anisotropy, with propagation
velocities faster in the horizontal direction than in the vertical direction.
Further west, at CDP 300 (Fig. 9b), the first arrivals are less clear at far
offsets in spite of low velocity layer 1 being thinner (Table 3), implying
that other factors are affecting the data quality than the uppermost dry
loose sediments. First arrivals are modeled well by a 5 layer model, but
there is no clear reflection from the base of layer 4 (the equivalent to
layer 3 at CDP 600). This may be due to the interface being disturbed
at this location or generally poorer data quality in this area. Reflections
at this location are also absent on the stacked section (Fig. 7). Note that
the noise cone is not modeled as well here as at CDP 600. Further east,
at CDP 1300 (Fig. 9c), first arrivals from the equivalent of layer 3 at
CDP 600 and layer 4 at CDP 300 are absent. Instead first arrivals from a
layer with a slope corresponding to 1550 m/s are present to offsets of al-
most 500 m. A refracted wave with an apparent velocity of 5100 m/s
then arrives first out to about 800 m. At greater offsets it is more difficult
to identify the first arrival, but the apparent velocity increases up to
about 7500 m/s, a very high value. Significant multiple energy is also
present in the synthetic data as is a noise cone that corresponds well in
time with the real data. The reflection from the base of layer 2 arrives
within this noise. Therefore, it is unlikely that this reflection would be
imaged on the real data, although signs of it are present.
The velocity-depth models that were determined along both pro-
files have been plotted onto the stacked migrated sections to aid in6 8
e (km)
0 1200 1400 1600 1800
P
E
6 8
ce (km)
00 1200 1400 1600 1800
P
E
acked and (b) migrated sections.
0 2 4 6 8 10
Distance (km)
0
a
b
1
Ti
m
e 
(s)
Ti
m
e 
(s)
200 400 600 800 1000 1200 1400 1600 1800 2000 2200
CDP
SW NE
0 2 4 6 8 10
Distance (km)
0
1
200 400 600 800 1000 1200 1400 1600 1800 2000 2200
CDP
SW NE
Fig. 8. Orsa profile showing the (a) stacked and (b) migrated sections.
96 C. Juhlin et al. / Tectonophysics 580 (2012) 88–99the interpretation (Figs. 10 and 11). Prior to plotting, the depths were
converted to time and a static shift corresponding to the refraction static
correction at each CDP location was applied. This allows a rough com-
parison to be made between the refraction models and the reflection
images. The refraction models and the interpretation of the reflections
images are discussed in the next section.
7. Discussion
The reflection seismic images do not show a simple layered structure
along the profiles, indicating that complex geological conditions are
present in some parts. Interpretation of the profiles based solely on the
reflection images is difficult. However, themodeled refractors aid signif-
icantly in the interpretation, especially if lithologies can be assigned to
the velocities. Unfortunately, no velocity logs are available through the
sedimentary successions from any of the boreholes. Therefore, we
have assumed the following in our interpretation: velocities less than
2100 m/s correspond to glacial sediments, velocities in the range of
2500–3200 m/s correspond to Silurian clastic rocks, velocities in the
range of 3800–4200 m/s correspond to Ordovician limestones, veloci-
ties in the range of 4500–5500 m/s correspond to highly fractured crys-
talline rock or Precambrian meta-sedimentary/volcanic rocks and
velocities greater than 5500 m/s correspond to mainly intact crystalline
bedrock.
In the Mora section (Fig. 10), the refractor velocity corresponding to
the interpreted top of the Silurian is relatively consistent along the pro-
file at about 3000–3200 m/s. This refractor is only present in the south-
ern and northern parts of the Orsa profile (Fig. 11), indicating the
Silurian to be thinner here, or perhaps absent in the central part. The
interpreted refraction velocity at the top of the Ordovician is generallyFig. 9. CDP super gathers with corresponding elastic modeling at CDP locations (a) 600, (b)
calculated from the velocity-depth models in Table 3 while the green line represents the refl
300 and the base of layer 2 for CDP 1300. Data area plotted with at reduced times with thehigher and more variable in the Orsa profile than in the Mora profile.
In addition, the velocities below the base of the interpreted Ordovician
on the Orsa profile are high at about 6000 m/s, indicating intact crystal-
line basement is present. On the Mora profile, first arrivals with such a
high apparent velocity are not observed, except at CDP 1300, suggesting
that the crystalline basement is deeper and has not been mapped. At
around CDP 1300 (Fig. 9c) higher apparent velocities are observed, but
this is very local. The 5100 m/s velocities at CDP 1300 are probably too
low to represent intact crystalline basement and the 7500 m/s velocity
may be due to geometrical effects or to a high velocity mafic intrusion
at depth. The deeper reflections corresponding to this high velocity
boundary are also a sign of a strong impedance contrast at depth (Fig. 10).
The only clear reflection from within the basement is observed on
the Orsa profile, a gently dipping event at about 900 ms (Fig. 11).
Given that sub-horizontal reflections further east at similar depths are
generated by thin dolerite sills (e.g. Juhlin, 1990), it is likely that this
reflection also is generated from a dolerite sill at about 2 km depth.
However, it is also possible that it represents a detachment surface on
which mega-blocks have slid. Therefore, it is of interest to drill through
the interface generating this reflection.
Recent studies of new drill cores in the Siljan Ring (Lehnert et al.,
2012) have shown that the classical distribution in the Siljan area of an
Ordovician carbonate platform followed by Silurian shales and mud-
stone and the Silurian/Devonian terrestrial Orsa Sandstone has to be
re-evaluated and ancient facies belts have to be reconstructed in detail
for different time slices. The development of the Lower–Middle Ordovi-
cian carbonate succession, from the contact to the Precambrian base-
ment to the upper part of an overlying limestone (Holen Formation) is
approximately uniform across the Siljan district at about 20 m. In the
Mora area this limestone exhibits exposure plus erosion, interpreted as300 and (c) CDP 1300 along the Mora profile. Blue lines represent the first break times
ection time corresponding to the base of layer 3 for CDP 600, the base of layer 4 for CDP
reduction velocity given by RVEL.
0.0
a
b
c
0.1
0.2
0.3
Ti
m
e 
(s)
0.0
0.1
0.2
0.3
Ti
m
e 
(s)
0.0
0.1
0.2
0.3
Ti
m
e 
(s)
0 200 400 600 800 1000 1200 1400
Offset (m)
0 200 400 600 800 1000 1200 1400
Offset (m)
0 200 400 600 800 1000 1200 1400
Offset (m)
0 200 400 600 800 1000 1200 1400
Offset (m)
0 200 400 600 800 1000 1200 1400
Offset (m)
0 200 400 600 800 1000 1200 1400
Offset (m)
CDP=300RVEL=3000.0
CDP=600RVEL=3000.0
CDP=1300RVEL=3000.0
97C. Juhlin et al. / Tectonophysics 580 (2012) 88–99
Table 3
Modeling parameters used for Fig. 9. HS refers to a half-space.
Layer CDP 300 CDP 600 CDP 1300
Base depth (m) Vp (m/s) Base depth (m) Vp (m/s) Base depth (m) Vp (m/s)
1 8 600 17 600 5 600
2 27 1600 50 1700 179 1550
3 120 2520 254 3200 402 5100
4 217 3200 HS 4050 HS 7500
5 HS 4100
Fig. 10. Interpreted Mora profile. Times to refractors are plotted as short red lines at those CDP locations where a reasonable time-offset curve for the first arrivals could be picked.
The velocity of the refracting interface is shown below each red line. Shown also are interpretations of the boundaries of the different geological units based on the reflection seismic
images and the refraction interpretation. Yellow—base of the water saturated loose sediments, green—base of the Silurian, blue—base of the Ordovician, black—dolerite intrusion,
gray—fault.
98 C. Juhlin et al. / Tectonophysics 580 (2012) 88–99due to the formation of a peripheral bulge in the Caledonian foreland
basin. Later, the limestonewas covered by amore than 200 m thick Silu-
rian clastic succession. This Silurian shale basin in the Mora area (prob-
ably thinning or partially eroded towards the east) may partly explain
why the basement is apparently deeper along the Mora seismic profile
than along the Orsa profile.
The only location along the Mora profile that the reflection seismic
data indicate a relatively undisturbed layering of the sedimentary
successions is between CDPs 350 and 700. Here the interpreted top
of the Ordovician corresponds to a clear seismic reflection, indicating
little internal deformation, and the refraction modeling indicates the
Silurian to be about 200 m thick below CDP 600 (Table 3). If the
deeper reflection at 0.4 s (Fig. 10) corresponds to the top of the base-
ment then the unit below is about 400 m thick, giving a total thickness
of the Paleozoic successions of about 600 m here. Whether the lower-
most unit consists entirely of Ordovician limestones or whether there
may be Cambrian units present at depth cannot be determined from
the seismic data. On the rest of theMora profile there are only short seg-
ments where reflections directly correlate with the results from the
refraction interpretation, indicating that the sedimentary successions
are heavily disturbed. The depth to the crystalline basement is relatively
well determined along the Orsa profile, generally being on the order of
200–300 m, except south of CDP 500 and in the heavily faulted northern
part of the profile (CDPs 1750 to 2000) that was mentioned earlier. This
is significantly less than along theMora profile, suggesting that the facies
may be different north of Lake Orsa compared to that found in the Mora
area.
Intense thrusting during the emplacement of megablocks is observed
in the yet to be described Stumsnäs 1 (Fig. 1) core of Igrene AB (repeated
thrusting of granitic basement over fault-bounded Ordovician limestone
units) and the Nittsjö trench (Fig. 1) (thrust slices of Lower/Middle Ordo-
vicianmarls in structural contactwith Proterozoicmeta-volcanoclastics).
In addition, the occurrence of a Silurian clastic succession (NederbergaFormation), a facies typical of the Silurian shale basin in the western
part of the impact structure, in the north-central part of the ring structure
at Skattungbyn reflects also some large-scale overthrusting and not sim-
ple sliding of blocks from the rim of the crater towards the central uplift.
Similar thrusting may be expected to be present in the Mora area and
may explain the lack of coherency of reflections along most of the Mora
profile.
8. Conclusions
The reflection seismic data presented here shed new light on the
Paleozoic successions in the western Siljan Ring. They show a complex
and disturbed reflection pattern, especially along the Mora profile. Re-
flections are more coherent on the northern Orsa profile. Interpretation
of the reflection images was greatly aided by combining information
from refraction modeling, allowing depths to the top of Ordovician
limestone to be estimated along both profiles.
New core data from a borehole drilled in the town of Mora shows a
thick Silurian shale sequence (more than 200 m) above the Ordovician
limestone in this part of the Siljan Ring. Where reflections are clear
from the interpreted base of the Silurian on the Mora seismic profile
(CDPs 350 to 700) we estimate the Silurian to be about 200 m thick, in-
dicating that this shale sequence extends at least this far to the north.
Depth to the crystalline bedrock is generally greater along theMora pro-
file than theOrsa profile, with the Paleozoic successions interpreted to be
as thick as 600 m in thewestern part of theMora profile. Both the reflec-
tion images and the refraction interpretation indicate the bedrock to be
highly faulted. In the northern part of the Orsa profile the data indicate
normal faultswith throws of up to 200 m. Significant faulting of the bed-
rock is also interpreted in the eastern part of the sedimentary successions
on the Mora profile.
In order to study the distribution of significant sedimentary changes
(limestones vs. shales vs. sandstone units) and to reconstruct the
Fig. 11. Interpreted Orsa profile. Times to refractors are plotted as short red lines at those CDP locations where a reasonable time-offset curve for the first arrivals could be picked.
The velocity of the refracting interface is shown below each red line. Shown also are interpretations of the boundaries of the different geological units based on the reflection seismic
images and the refraction interpretation. Yellow—base of the water saturated loose sediments, green—base of the Silurian, blue—base of the Ordovician, black—dolerite intrusion,
gray—fault, pink area—possible reflective zone in the Precambrian basement.
99C. Juhlin et al. / Tectonophysics 580 (2012) 88–99extension of coeval facies belts, the position of megablocks and thrust
slices, higher resolution shallow seismic studies together with new
drill cores are required. One new core must include the Silurian
siliciclastic succession in the Orsa area, where lithostratigraphic control
is needed for seismic interpretation of the data reported here. Future
higher resolution shallow seismic studies will help to document the
contacts between various Silurian sandstone and shale units as well as
between Silurian siliciclastics and underlying Ordovician shelf carbon-
ates resting on Precambrian basement rocks. Both additional seismic
data and shallow drilling are required before the location of a deeper
borehole can be finalized.
Acknowledgments
The Swedish Research Council (VR) funded the reflection seismic data
acquisition (project number 2009‐4492) and is gratefully acknowledged.
GLOBE Claritas™ under license from the Institute of Geological and
Nuclear Sciences Limited, Lower Hutt, New Zealand was used to process
the seismic data.
References
Bergström, S.M., 2007. The Ordovician conodont biostratigraphy in the Siljan region, south-
central Sweden: a brief review of an international reference standard. In: Ebbestad,
J.O.R., Wickström, L.M., Högström, A.E.S. (Eds.), WOGOGOB 2007. 9th Meeting of the
Working Group on Ordovician Geology of Baltoscandia. Field Guide and Abstracts:
Sveriges Geologiska Undersökning Rapporter och Meddelanden, 128, pp. 26–41. 63–78.
Bodén, A., Erikson, K.G., 1988. Deep Drilling in Crystalline Bedrock; Vol. 1: The Deep
Gas Drilling in the Siljan Impact Structure, Sweden and Astroblemes. Springer
Verlag, Berlin . 364 pp.
Collini, B., 1988. Geological Setting of the Siljan Ring Structure Deep Drilling in Crystal-
line Bedrock, Vol. 1: The Deep Gas Drilling in the Siljan Impact Structure, Sweden
and Astroblemes. Springer Verlag, Berlin, pp. 349–354.
Ebbestad, J.O.R., Högström, A.E.S., 2007. Ordovician of the Siljan district, Sweden. In:
Ebbestad, J.O.R., Wickström, L.M., Högström, A.E.S. (Eds.), WOGOGOB 2007. 9th
meeting of the Working Group on Ordovician Geology of Baltoscandia. Field
guide and Abstracts: SGU Rapporter och meddelanden, 128, pp. 7–26.
Grieve, R.A.F., 1988. The formation of large impact structures and constraints on the nature of
Siljan. In: Boden, A., Eriksson, K.G. (Eds.), DeepDrilling in Crystalline Bedrock. : TheDeep
Gas Drilling in the Siljan Impact Structure, Sweden and Astroblemes, Proceedings of the
International Symposium, vol. 1. Springer Verlag, Berlin, pp. 328–348.
Henkel, H., Aaro, S., 2005. Geophysical investigations of the Siljan impact structure—a
short review. In: Koeberl, Henkel, H. (Eds.), Impact Tectonics. Springer Verlag, Berlin,
pp. 247–283.
Högström, A.E.S., Sturkell, E., Ebbestad, J.O.R., Lindström, M., Ormö, J., 2010. Concentric
impact structures in the Palaeozoic of Sweden—the Lockne and Siljan craters. GFF
132, 65–70.Holm, S., Alwmark, C., Alvarez,W., Schmitz, B., 2011. Shock barometry of the Siljan impact
structure, Sweden. Meteoritics and Planetary Science 46, 1888–1909.
Juhlin, C., 1990. Interpretation of the reflections in the Siljan Ring area based on results
from the Gravberg-I borehole. Tectonophysics 173, 345–360.
Juhlin, C., 1995. Finite difference elastic wave propagation in 2-D heterogeneous trans-
versely isotropic media. Geophysical Prospecting 43, 843–858.
Juhlin, C., Pedersen, L.B., 1987. Reflection seismic investigations of the Siljan impact
structure, Sweden. Journal of Geophysical Research 92, 14113–14122.
Juhlin, C., Pedersen, L.B., 1993. Further constraints on the formation of the Siljan impact crater
from seismic reflection studies. Geologiska Föreningens i Stockholm Förhandlingar 115,
151–158.
Juhlin, C., Dehghannejad, M., Lund, B., Malehmir, A., Pratt, G., 2010. Reflection seismic
imaging of the end-glacial Pärvie Fault system, northern Sweden. Journal of Ap-
plied Geophysics 70, 307–316.
Juhojuntti, N., Juhlin, C., 1998. Seismic lower crustal reflectivity and signal penetration
in the Siljan Ring area, Central Sweden. Tectonophysics 288, 17–30.
Kenkmann, T., von Dalwigk, I., 2000a. Centro‐symmetrical material flow during crater
modification: structural implications. Lunar and Planetary Science Conference
XXXI, Abstract 1041.
Kenkmann, T., vonDalwigk, I., 2000b. Radial transpression ridges: a new structural feature
of complex impact craters. Meteoritics and Planetary Science 35, 1189–1202.
Lehnert, O., Meinhold, G., Bergström, S.M., Calner, M., Ebbestad, J.O.R., Egenhoff, S., 597
Frisk, Å.M., Hannah, J.L., Högström, A.E.S., Huff, W.D., Juhlin, C., Maletz, J., Stein, H.J.,
Sturkell, E., Vandenbroucke, T.R.A., 2012. New Ordovician-Silurian drill cores from
the Siljan impact structure in central Sweden - an integral part of the Swedish
Deep Drilling Program. GFF 134, 87–98. http://dx.doi.org/10.1080/11035897.
2012.692707.
Lorenz, H., 2010. The Swedish Deep Drilling Program: for science and society. GFF 132,
25–27.
Papasikas, N., Juhlin, C., 1997. Interpretation of reflections from the central part of the
Siljan Ring impact structure based on results from the Stenberg-1 borehole.
Tectonophysics 269, 237–245.
Park, C.B., Miller, R.D., Steeples, D.W., Black, R.A., 1996. Swept impact seismic technique
(SIST). Geophysics 61, 1789–1803.
Petalas, C., 1985. Sedimentary petrology of the Orsa sandstone, central Sweden.
UUDMP research report, 40, pp. 1–138.
Reimold,W.U., Kelley, S.P., Sherlock, S., Henkel, H., Koeberl, C., 2004. Laser argon dating of
melt breccias from the Siljan impact structure, Sweden—implications for possible
relations to late Devonian Extinction events. Lunar and Planetary Science Conference,
XXXV, Abstract 1480.
Reimold, U., Kelley, S.P., Sherlock, S., Henkel, H., Koeberl, C., 2005. Laser argon dating of
melt breccias from the Siljan impact structure, Sweden: implications for a possible
relationship to Late Devonian extinction events. Meteoritics and Planetary Science
40, 591–607.
Rondot, J., 1975. Comparison entre les astroblemes de Siljan, Suede et de Charlevoix,
Quebec, vol. 6. Bulletin of the Geological Institute of the University of Uppsala,
N.S., pp. 85–92.
Wu, J., Milkereit, B., Boerner, D.E., 1995. Seismic imaging of the enigmatic Sudbury
Structure. Journal of Geophysical Research 100 (B3), 4117–4130.