Geophys. J. R . astr. SOC.

(1 984) 78,66 1-689

A seismic study of deep geological structure in the

Bristol Channel area, SW Britain

J. Mechie Geophysikalisches Institut, Universitat Karlsruhe, Hertzstrasse 16,

D-7500 Karlsruhe 21, Federal Republic o f Germany

Downloaded from https://academic.oup.com/gji/article/78/3/661/648611 by guest on 24 May 2022

M. Brooks Department of Geology, University college, Cardiff, PO BOX 78,
Cardiff CFl I X L , Wales

Received 1983 November 14; in original form 1983 July 1 1

Summary. Results from eight seismic refraction lines, 35-90 km long, in the
Bristol Channel area are presented. The data, mostly land recordings of marine
shots, have been interpreted mainly by ray-tracing and time-term modelling.
Upper layer velocities through Palaeozoic rocks usually fall within the range
4.8-5.2 km s-' . Below the Carboniferous Limestone with a normal velocity
of 5.1-5.2 kms-', the Old Red Sandstone with a velocity of 4.7-4.8 kms-'
acts as a low velocity layer, as d o parts of the underlying Lower Palaeozoic
succession. In the central South Wales/Bristol Channel area and the Mendips,
a 5.4-5.5 kms-' refractor is correlated with a horizon at or near the top of
the Lower Palaeozoic succession. Under the whole area, except for north
Devon, a 6.0-6.2 km s-l basal refractor has been located and is correlated with
Precambrian crystalline basement rocks. In general, this refractor deepens
southwards from a series of basement highs, which existed before the major
movements of the Variscan orogeny in South Wales, resulting in a southerly
thickening of the pre Upper Carboniferous supra-basement sequence. In north
Devon, a 6.2kms-' refractor at shallow depth, interpreted as a horizon in
the Devonian or Lower Palaeozoic succession, overlies a deep reflector that
may represent the Precambrian crystalline basement.

1 Introduction and geology

In the period 1973-78, a series of seismic refraction experiments was completed in the
Bristol Channel and surrounding land areas to obtain information on the geological structure
of the t o p 5-10km of the crust (Fig. 1). Brooks, Bayerly & Llewellyn (1977), Bayerly &
Brooks (1980) and Llewellyn (1981) have provided interpretations of refraction lines com-
pleted before 1976 (thin lines, Fig. 1). Interpretation of subsequent lines (thick lines, Fig. 1)
and a regional time-term analysis of a 6.0-6.2 kms-' basal refractor, using all available data,
was the purpose of a PhD project (Mechie 1980) and is summarized in this paper.
A geological map of the region is shown in Fig. 2. The land areas around the Bristol
Channel are dominated by Variscan structures including the South Wales and Pembrokeshire
Coalfields, the Mendip Hills, and the North Devon anticline. However, over large parts of the
662 J. Mechie and M.Brooks

S O U T H W A L E

Downloaded from https://academic.oup.com/gji/article/78/3/661/648611 by guest on 24 May 2022

Figure 1. Large-scale seismic refraction lines in the Bristol Channel area.

region, including the Bristol Channel, the Vale of Glamorgan, SE Wales, and Somerset the
Variscan terrains are concealed beneath a variable cover of Mesozoic rocks.
The South Wales and Pembrokeshire Coalfields represent a synclinal zone containing a
thick sequence of Upper Carboniferous rocks. In the South Wales Coalfield, about lOkm
east of refraction line B2, a maximum thickness of about 3 km of Upper Carboniferous
rocks is preserved. North of the coalfield progressively older rocks occur, including Old Red
Sandstone (ORS) and the thick Lower Palaeozoic sequence of the Welsh basin. In south-east
Carmarthenshire (Cope 1977) and the St David’s district of Pembrokeshire, Precambrian
igneous and sedimentary rocks are preserved in local inliers. Line L crossed one such inlier,
the Hayscastle anticline, and thus provided the opportunity to test for correlation between
the 6.0-6.2 k m s-l basal refractor and Precambrian rocks. In Pembrokeshire, the southern
edge of the coalfield is partly defined by the Johnson thrust. Along this fault, Precambrian
and Silurian rocks of the Johnston-Benton fault block (Brooks, Mechie & Llewellyn 1983),
which lines K and L crossed, have been thrust northwards over Upper Carboniferous rocks.
In southern Pembrokeshire, several E-W trending Variscan structures give rise to outcrops
of Lower Ordovician to Upper Carboniferous rocks (see Fig. 2).
South of the South Wales Coalfield, in Gower and the Vale of Glamorgan, the dominant
Palaeozoic rock-type exposed is Carboniferous Limestone. However, a series of en echelon
WNW-ESE trending Variscan anticlines bring ORS and Silurian rocks to the surface. In
south Gower, in the vicinity of lines B2 and C2, the Carboniferous Limestone is about 1 km
thick but on the north and east crops of the coalfield, which lines B2 and J respectively
crossed, the Carboniferous Limestone is very thin (30-200 m). Between the east crop and the
Silurian rocks exposed in the core of the Usk anticline, a wide outcrop of ORS occurs which
line J traversed.
In the Mendip Hills, where line I was located, Carboniferous Limestone, ORS, and Silurian
rocks including volcanics emerge from beneath the Mesozoic cover in a series of en echelon,
E-W trending Variscan periclines (Fig. 2). In north Devon and west Somerset, where line F2
was situated, a thick Devonian succession crops out in the major North Devon anticline
(Fig. 2).
Beneath the Bristol Channel thick Mesozoic sequences are preserved in major E-W
trending synclines. The most important of these is the Bristol Channel syncline which con-
 b
z
35'
ul
z
9
3
2B
2.
B
9
rr
3
3
Figure 2. Geology of the Bristol Channel area. %
!
r;
Key: 8
Onshore Offshore
0 Cretaceous Upper Carboniferous a Marine Devonian Tertiary 635 Permo-Trias -+anticline
Jurassic Lower Carboniferous Lower Palaeozoic Middle-late Jurassic Palaeozoic -TT fault (tick on down thrown side)
Permo-Trias Old Red Sandstone Precambrian Early Jurassic + syncline TT thrust
a\
JT Johnston thrust; BF Benton fault; BCFZ Bristol Channel fault zone; HA Hayscastle anticlines; PCS Pembrokeshire Coalfield syncline; CBA Cefn Bryn anticline;
CaCoA Cardiff -Cowbridge anticline; UA Usk anticline; BCS Bristol Channel syncline; NDA North Devon anticline; LG Lundy granite (Tertiary); other unreferenced
structures - see Mechie (1980).
664 J. Mechie and M. Brooks
tains about 2 km of Mesozoic rocks (Brooks & James 1975) of which the youngest are of
late Jurassic age (Lloyd ef d. 1973; Evans & Thompson 1979). The northern limb of the
syncline is complicated by the major E-W trending Bristol Channel fault zone (Fig. 2). In
the outer part of Barnstaple Bay, where the western shots of line F2 were located, a Tertiary
basin (Fletcher 1975) containing about 350m of sediments (Brooks & James 1975) is
located immediately east of the Sticklepath fault.
In the study area, the prediction of the deep geological structure using seismic refraction
datais complicated by the severity of the Variscan structures and the likely presence of major
unconformities at depth. Direct information on the deep geology is limited to two boreholes
drilled by Cambrian Exploration Ltd in the southern part of the South Wales Coalfield. These
reveal the presence of 1-1.6 km of ORS and 350 km (base not reached) of Lower Palaeozoic
rocks.

Downloaded from https://academic.oup.com/gji/article/78/3/661/648611 by guest on 24 May 2022

Previous seismic studies (Bayerly & Brooks 1980) suggested that the precarboniferous
supra-basement sequence thickens from under the central part of the South Wales Coalfield
to beneath Gower and the Vale of Glamorgan. In the Bristol Channel, the success of the
short marine refraction lines in delineating the structure of the Palaeozoic rocks beneath the
thick Mesozoic cover has been limited 'due t o the problem of overlapping velocity ranges for
the main Upper Palaeozoic rock-types' (Brooks & Al-Saadi 1977).

2 Collection, processing, and methods of interpretation of data

Lines B2, C2, F2, I, J , K and L (Fig. 1) were established during a research cruise of RRS
JohnMurray in 1976. Forty-one shotswere fired using Superflex 200 explosive cord (Nobels).
The shot instant was recorded by hydrophones towed in the vicinity of the explosive charges.
Each marine shot was recorded on land at 10 three-component stations, using MARS 66
(Berckhemer 1976) and Racal Geostore analogue tape recording systems with HS-10 or
Willmore Mk I11 seismometers and HBG or MSF time signal receivers. A total of 410
recordings was made during the experiment.
During 1977-78, 42 additional recordings were obtained from four marine shots, fired
from MV Edward Forbes, and 10 quarry blasts, in order to improve the regional time-term
analysis of the basal refractor.
Shot position fixing at sea was achieved by radar and Decca fixes and the maximum posi-
tionalerror isestimated t o be 2 200 m. On land, the positions of recording stations and quarry
blast locations were located with an accuracy off. 25 m or better.
Lines B2, C2 and F2 each comprise a long line of eight or nine shots fired into an in-line
recording array of ten stations, while lines 1 and J have essentially one shot-point fired into
an in-line recording array of around 25 stations (Fig. 1). In Pembrokeshire, two conventional
reversed lines, K and L, were established.
Data were digitized at 100 samples s-'. Figs 3, 6, 7 , 9 , 10, 12, 15 and 16 show, for 2.25 s
of the vertical components, normalized and filtered P-wave record sections, plotted with a
reduction velocity of 6 km s-'. In these record sections the seismograms have been relatively
shifted t o correct first arrival times for water depth, elevation, Mesozoic rocks and, on line
B2, Upper Carboniferous rocks. Of the corrections which were made, only those for Mesozoic
and Upper Carboniferous rocks on line B2 were significant. In the ray-tracing analysis, the
observed travel-times were only corrected for water depth and elevation. The Mesozoic and
Upper Carboniferous rocks were included in the ray-tracing models. The direct and refracted
phases Po, PA, P1 and P z , drawn in on the record sections, have been fitted to the arrivals as
T-X straight lines by least-squares. Thus, apparent velocities and intercept times, together
with standard error estimates, have been obtained for the phases. The Po and P A phases
represent direct or refracted waves through the upper layers. The PI phase represents a
 A seismic study in the Bristol Channel area 665
refracted wave from an intermediate layer with a true velocity of about 5.5 km s-'. The P2
phase represents the wave from the 6.0-6.2 km s-l basal refractor.
As a first step in interpreting the data, a planar-layer interpretation was derived for each
line (Johnson 1976). In the case of lines B2, C2 and F2, the interstation and intershot velo-
cities were used as the forward and reverse apparent velocities respectively. In most cases,
the planar-layer interpretations were taken as starting models for two-dimensional modelling
using the ray-tracing method for laterally inhomogeneous media with curved interfaces
(Cerven9, Langer & PEenEik 1974). The starting models were adjusted by manual iteration
until the observed travel times closely matched the calculated (ray-tracing) travel-times. The
average absolute differences between the observed and calculated travel-times are given in
Table 1. In addition, the time-term method (Willmore & Bancroft 1960; Berry & West 1966)
was used to obtain a three-dimensional picture of the 6.0-6.2 km s-l basal refractor using

Downloaded from https://academic.oup.com/gji/article/78/3/661/648611 by guest on 24 May 2022

arrivals from this refractor obtained from the 1976-78 experiments together with those
from previous experiments (Bayerly & Brooks 1980; Llewellyn 1981).
Table 1. Average absolute differences between observed travel times and those calculated by ray-tracing
for phases P, and P, from the various seismic refraction lines.

Phase Pl p,
Line Difference No. of Difference No. of
(S) observations (S) observations
L - 0.03 16
K - - 0.05 16
82 0.03 12 0.05 19
c2 0.05 13 0.07 26
J 0.04 6 0.04 5
I 0.03 9 0.02 6

3 Lines in Pembrokeshire - K, L
Two lines, K and L (Fig. l), both reversed with effectively one shot-point at each end and
18-20 recording stations about 2 km apart, were completed in Pembrokeshire. Both lines,
for which the record sections are shown in Fig. 3, were approximately 4 0 k m long and
crossed at a common recording station, K12-L8.
Line L trends NNW-SSE, almost perpendicular to the local geological strike, and crosses
a variety of Caledonian and Variscan structures. The northern part of the line (recording
stations L11-L20) mainly crosses ENE-WSW trending Caledonian structures, the chief of
which is the Hayscastle anticline exposing Precambrian igneous rocks in its core (Figs 2 and
4). The southern part of the line (recording stations L1-L10) crosses E-W trending Variscan
structures, including the synclinal structure of the Pembrokeshire Coalfield and, at the
southern edge of the coalfield, the Johnston thrust defining the northern margin of the
Johnson-Benton fault block. The southern boundary of the block is defined by the Benton
fault, which brings ORS against ORS, Silurian, or Precambrian rocks along a southerly
dipping fault plane (Figs 2 and 4) (Hancock, Dunne & Tringham 1983). Line K runs almost
parallel to the trend of local Variscan structures. West of station K9, it follows the line of
the Johnston-Benton fault block while east of station K9, it follows approximately the
Carboniferous Limestone outcrop on the southern limb of the Pembrokeshire coalfield
(Figs 2 and 5).
On record section L(S) (Fig. 3a), the direct phase Po has a velocity of 5.01 f 0.06 km s-'
which represents an average velocity through a varied sequence of Precambrian to Upper
Carboniferous rocks. However, along the southern part of line L there is a 13 km section of
666 J. Mechie and M. Brooks

u
w
m
b z
Y

Downloaded from https://academic.oup.com/gji/article/78/3/661/648611 by guest on 24 May 2022

5

Y
t
0
W
0
3
0
dw
E

D I S T A N C E IN K M
Figure 3. Normalized, band-pass filtered record sections for (a) L(S), (b) L(N), (c) K(W) and (d) K(E),
plotted with a reduction velocity of 6 km s-', and showing least-squares fits of the P-wave arrivals.

which 9.2km are represented by ORS and the velocity of 4.85 f 0.13 km s-l across this
13 km section must be close to the local ORS velocity. The onsets on the first two traces
from record section L(N) (Fig. 3b) poorly define the direct phase Po with a velocity of
4.86 km s-'. This is the only case in the study area where a direct phase through the varied
Upper Cambrian-Lower Ordovician sequence has been observed and thus the local velocity
represents the only direct evidence of the velocity of this major rock interval. On record
section K(W) (Fig. 3c), the direct phase Po has a velocity of 5.22 f 0.06 km s-l representing,
again, an average velocity through a varied sequence of Precambrian to Upper Carboniferous
rocks. Finally, the direct phase P o , on record section K(E) (Fig. 3d), has a velocity of
*
5.20 0.19 km s-l representing Carboniferous rocks.
The remainder of the onsets define arrivals of the phase P2 from the basal refractor, with
apparent velocities of 6.46 f 0.12, 5.96 k 0.07, 6.19 k 0.09, and 6.82 k 0.1 1kms-' on
record sections L(S), L(N), K(W) and K(E) respectively (Fig. 3a-d).
Two-layer dipping models for lines L and K gave true velocities for phase P2 of 6.19 f
0.10 and 6.48 f 0.18 km s-' respectively, while time-term solutions for the data of 'lines
L and K resulted in refractor velocities of 6.05-6.22 km s-l . In addition t o these results,
 A seismic study in the Bristol Channel area 667
true velocity values of 6.0-6.2kms-' for the basal refractor are almost everywhere
encountered throughout the South Wales/Bristol Channel region (see Bayerly & Brooks
1980; Brooks et al. 1983; sections 4 and 5), and thus the value of 6.48 km s-' obtained
from the planar-layer interpretation of line K appears anomalous. As the two apparent
velocities observed along line K are determined from different sections of the refractor,
however, it is possible that both are up-dip apparent velocities (see p. 669). Thus, in the
ray-tracing interpretation, in which a vertical velocity gradient is necessary in a layer in
order to make the ray curve, true velocities of 6.1 and 6.3 km s-l were used at the top
and bottom, respectively, of the basal layer (Figs 4 and 5).

NNW K12JT BF SSE

Downloaded from https://academic.oup.com/gji/article/78/3/661/648611 by guest on 24 May 2022

LI

JT BF
jl6km

tracing velocities in km s-'; faults --; other geological boundaries -;

a preUpper Carboniferous supra-basement rocks; Silurian;
=
Figure 4. Combined ray-tracing and geological model for line L. Key: ray-tracing interfaces -;

Precambrian volcanics;
ray-
Upper Carboniferous;
Pre-
cambrm crystalline basement rocks; JT Johnston thrust; BF Benton fault.

WNW BFJT JT LE ESE

ikrn
Figure 5. Combined ray-tracing and geological model for tine K. Key as for Fig. 4.

Anomalies in the arrival times of refracted rays, such as early arrival times at stations in
the vicinity of the Hayscastle anticline, seemed t o be better correlated with the structure of
the basal refractor than with variations in top layer velocity. Thus, in attempting to interpret
the anomalies in terms of refractor structure, the rock-types above the basal refractor on
both lines were modelled with the constant velocities (Figs 4 and 5) derived for the direct
phase Po (Fig. 3a-d). For the models shown in Figs 4 and 5, ignoring variations in velocity
of surface rock-types may lead to maximum errors of 0.03 s for rays travelling through the
Precambrian rocks of the Johnston-Benton fault block.
Time-term solutions for the data of lines L and K showed the shallowest depths to the
basal refractor, of 0-0.7 km, beneath station L16, immediately north of the outcropping
Precambrian in the Hayscastle anticline. Thus, the refractor depth beneath station L16 was
set equal to 0.3-0.4km in the ray-tracing interpretation (Figs 4 and 5), for which the
average absolute differences between calculated and observed travel-times are 0.03 and
0.05 s for lines L and K respectively (Table 1).
The ray-tracing interpretation shows that, where the Precambrian volcanics crop out in
the Hayscastle anticline, the basal refractor is encountered at depths of 1-2 km and that, at
least locally, it must represent Precambrian rocks. This 6.0-6.2 km s-l refractor, encountered
 668 J . Mechie and M. Brooks
throughout the South Wales/Bristol Channel region, is thought to represent a Precambrian
crystalline basement, partly intrusive igneous and partly metamorphic in character. In con-
sidering the possibility of a 1-2 km thickness of volcanics in the Hayscastle anticline, it
may be noted that 13 km to the west, in the St David's area, a thickness of 1.5 km (base
unseen) of Precambrian volcanics has been recorded (George 1970). North of the Hays-
castle anticline, the ray-tracing interpretation suggests that the basal refractor shallows and
that, consequently, the Precambrian volcanics decrease by about 1.7 km in thickness towards
station L16 (Fig. 4).
North and south o f station 16, the depths to the refractor increase to 1.5-2 km under the
northern end of the line and to about 3.5 km under the Pembrokeshire Coalfield. On record
section L(N) (Fig. 3b), there is a time-shift of 0.23s in the refracted phase Pz between
stations L11 and L13. In the ray-tracing interpretation (Fig. 4) this time-shift results in a

Downloaded from https://academic.oup.com/gji/article/78/3/661/648611 by guest on 24 May 2022

step of 600 m in the refractor, downthrowing to the south, just north of station L l l . To
account for the delay in arrival times of refracted rays at the southern shot-point and station
L1, another step feature is introduced in the ray-tracing model between stations L4 and L5
with a downthrow to the south of 1-2 km. Thus, at the southern end of line L, the depth to
the refractor is around 5.5 km.
The ray-tracing model for line K (Fig. 5 ) shows that the shallowest refractor depths of about
2 km occur beneath the western end of the Johnston-Benton fault block. Depths to the

" D I S T A N C E IN K M
Figure 6. Normalized, bandpass filtered record sections for line B2 for shots (a) 2, @) 3, (c) 5 , (d) 7 ,
(e) 8 and (0 9 to all stations, plotted with a reduction velocity of 6 km s-', and showing least-squares fits
of the P-wave arrivals.
 A seismic study in the Bristol Channel area 669
refractor increase to about 3.5 km beneath the eastern end of the Johnston-Benton fault
block. Further east, depths remain at 3.5-3.8 km before increasing to 4-4.5 km near the
eastern end of the line. The structure of the refractor beneath line K, as shown in the ray-
tracing model (Fig. 5 ) , is such that an updip apparent velocity would be observed beneath
the western part o f t h e line from the eastern shot-point while approximately the true velocity
would be observed beneath the eastern part of the line from the western shot-point. This is
in agreement with the apparent velocities of 6.19 and 6.82kms-' determined for phaseP2
on record sections K(W) and K(E) respectively (Fig. 3c,d).
Where both lines cross the Johnston-Benton fault block at its eastern end, the depth to
the refractor, which is correlated with the Precambrian crystalline basement (see above), is
about 3.5 km. Thus it seems that the Precambrian rocks of the Johnston-Benton fault block
are contained within an isolated thrust slice with a source at depth to the south.

Downloaded from https://academic.oup.com/gji/article/78/3/661/648611 by guest on 24 May 2022

At the western end of the Johnston-Benton fault block the Johnston thrust dips south
to SW at 25-45'. Thus, in one model (Fig. 4a) the Johnston thrust is shown to dip 45" S
at the surface. On this view it is possible that the southernmost step of 1-2 km in the basal
refractor represents the downward continuation, displaced along the Johnston thrust, of the
Benton fault. In this connection, while little or no movement took place along the Benton
fault at the end of the Carboniferous, substantial movement occurred of post lower ORS-
pre-Carboniferous age (Strahan et al. 1914). A downthrow to the south exceeding 1.5 km

CL

ta
\
X
I
e+
Y

W
E
CI

+
0
w
d 02
W
e

DISTANCE I N KM
Figure 6 - continued
670 J. Mechie and M. Brooks
(Hancock et al. 1983) is indicated where the fault is exposed about 5 km east of station L7.
This model requires about 3-4 km of vertical movement and 4-5 km of horizontal move-
ment along the Johnston thrust. An unknown but possibly substantial amount of movement
took place along the Johnston thrust at the end of the Carboniferous (Strahan et al. 1914;
George 1970).
In the other model (Fig. 4b) the Johnston thrust is shown to dip 30” S at the surface. In
this case, no correlation between the southernmost step feature in the basal refractor and the
Benton fault is possible. Moreover, greater horizontal and vertical movements than those
shown in Fig. 4(a) are required along the Johnston thrust. One possible major difficulty with
both models is, however, that at the eastern end of the Johnston-Benton fault block, about
5 km east of line L, the geological evidence for the eastward continuation of a major thrust is
debatable. Hancock et al. (1983) stated that at the eastern end of the block the Johnston

Downloaded from https://academic.oup.com/gji/article/78/3/661/648611 by guest on 24 May 2022

thrust ‘is represented by a swarm of subordinate splay faults’ and that ‘the total displacement
is reduced to 500 m’. On this basis they refute the views of Sanzen-Baker (1972) and Owen
(1974) that the Johnston thrust continues east of the block as a major structure with signifi-
cant displacement along it. Nevertheless, the present seismic interpretation suggests that the
structural setting for the Precambrian rocks of the block does involve major thrust faulting.

4 Lines in the Central South Wales/Bristol Channel area and the Mendips - B2, CY-A, C2,
J, I
In this area, five profiles were completed: four, B2, C2, I and J , using marine shots and one,
CY-A, using quarry blasts (Fig. 1).

4.1 L I N E B2

Line B2 was a N-S line, planned as an extension of earlier line B (Fig. 1). The line was 65 km
long, with nine marine shots fired across the Bristol Channel and 10 recording stations set
up across Cower and the northern limb of the South Wales Coalfield (Fig. 8). The line crossed
several major E-W trending structures. In the Bristol Channel, the profile first crossed, at
its southern end, Devonian rocks (Lloyd et al. 1973). It then crossed the Mesozoic rocks of
the Bristol Channel syncline (BCS) (Figs 2 and 8) and, further north, the Helwick syncline.
Estimates of the maximum thickness of Mesozoic strata preserved in the centre of the Bristol
*
Channel syncline include 1.85 km (Lloyd et al. 1973), 2.1 0.3 km (Brooks & James 1975),
and 2.47 ? 0.18km (Evans & Thompson 1979). Brooks & James (1975) and Evans &
Thompson (1979) postulated a southerly downthrow of about 1 km on the Bristol Channel
fault zone (BCFZ). The Helwick syncline contains about 300 m of Mesozoic strata beneath
line B2 (C. R. Price, private communication).
The five recording stations on Cower were sited on Carboniferous Limestone, although
the line also crossed the Cefn Bryn anticline (CBA) exposing ORS in its core. Immediately
north of Cower, the line crossed on to the steeply dipping southern limb of the South Wales
Coalfield. Where the line crossed the main synclinal axis of the coalfield, geological data
(Archer 1968) suggest a thickness of around 2.6 km of Upper Carboniferous rocks. After
crossing the gently dipping northern limb of the coalfield, the line ended on the north crop
of Carboniferous Limestone flanking the coalfield (Fig. 8).
Between shot 3 and station 3 phase Pb,with a velocity of 5.52 ? 0.03 km s-l , is observed
(Figs 6a, b and 7a). A similar phase was observed by Bayerly & Brooks (1980) and Llewellyn
(1981) at short range in the Carboniferous Limestone in the South Wales area. Bayerly &
Brooks (1980) attributed the phase t o dolomitized sections of Carboniferous Limestone and
 A siesmic study in the Bristol Channel area 67 1

Downloaded from https://academic.oup.com/gji/article/78/3/661/648611 by guest on 24 May 2022

0
w
u
a 3
0

DISTANCE I N KM
Figure 7. Normalized, band-pass fdtered record sections for line B2 for stations (a) 1, (b) 4, (c) 5, (d) 6,
(e) 9 and (0 10 to all shots, plotted with a reduction velocity of 6 km s-', and showing least-squares fits
of the P-wave arrivals.
672 J. Mechie and M. Brooks
suggested that wedging out or faulting out of the dolomitized zone might account for the
disappearance of the phase at greater ranges. North of shot 3, phase Po,with a velocity of
5.1 1 t 0.06 km s-I, is observed (Figs 6a and 7b, c, f). If the Po (and Ph) phases represent
arrivals from Carboniferous Limestone, it follows that the limestone exists beneath the
coalfield and extends south of Gower, under the Mesozoic rocks, at least as far south as shot
3. The amplitudes of the Po and Pi phases often decay quite rapidly to the point where the
phase is no longer observable as a first arrival (Figs 6a, b and 7a-c). Hence, between shots
3 and 9 no phase through the upper layers was observed. Thus, in the various interpretations
(see, e.g. Fig. 8), the velocity in these layers between shots 3 and 9 is only assumed.
Phase P, is observed with an interstation apparent velocity of 5.39 f 0.08 km s-' (Fig.
6a-f) and an intershot apparent velocity of 5.68 f 0.04 km s-' (Fig. 7a-d). Phase Pz has
an interstation apparent velocity of 6.09 f 0.03 km s-' (Fig. 6a, c-f) and an intershot

Downloaded from https://academic.oup.com/gji/article/78/3/661/648611 by guest on 24 May 2022

apparent velocity of 5.97 f 0.04 km s-' (Fig. 7c-f). True velocities of 5.52 f 0.09 km s-'
and 6.01 f 0.15 kms-' were derived for phases P I and Pz from a planar-layer interpreta-
tion and these velocities are approximately equal to the average velocities of the layers
propagating the P, and P, phases in the ray-tracing analysis discussed below (Fig. 8). The
planar-layer interpretation showed that the refractors propagating the PI and P, phases dip
north and south, respectively, and that the former wedges out north of station 7. Thus, these
features were included in the ray-tracing analysis.

N Coalfield Gower Bristol Channel S

CN A CZ C D BCFZ F2 F

b
Bkrn

Figure 8. Combined ray-tracing and geological model for line B2. Key: ray tracing interfaces
tracing velocities in km s-'; faults - -; other geological boundaries -; 0Mesozoic rocks; =Upper
- ; ray-

Carboniferous; Carboniferous Limestone; Old Red Sandstone; marine Devonian rocks;

Lower Palaeozoic/?Precambrian supra-basement rocks; =Precambrian crystalline basement rocks;
cross-over points with other lines are indicated (cf. Fig. 1). (b) Shows alternative interpretation at south
end of line (see Section 5).

For the final ray-tracing model (Fig. 8), the averave absolute differences between calcu-
lated and observed travel-times for the PI and P, phases are 0.03 and 0.05 s respectively
(Table 1). Results from marine seismic surveys were used to delineate the base of the Meso-
zoic sequence in the Bristol Channel syncline (Brooks & James 1975; C. R. Price, private
communication) and Geological Survey maps and borehole data (Archer 1968) were used to
delineate the base of the Upper Carboniferous succession in the South Wales Coalfield (see
above).
The thickness of the Carboniferous Limestone is 900 m in south Gower and decreases to
2 0 0 m north of the coalfield (George 1970) and t o nil between shots 3 and 8. These thick-
nesses are too small t o account for the delay times to the refractor propagating the PI phase.
 A seismic study in the Bristol Channel area 673
Thus the material above this refractor must include the ORS. Thickness estimates of the ORS
near the northern end of the line range from 1 km (Cope 1979) to 1.5 km (George 1970). Thus
it seems likely that under the coalfield low velocity Lower Palaeozoic and, possibly, Precam-
brian rocks are present to account for the large delay times to the refractors propagating the
PI and PZ phases (Fig. 8). In South Wales, velocities of 4.69-4.85 km s-l have been obtained
for the ORS from refraction profile measurements (see Section 3; James 1971; Bayerly
1978). In north Pembrokeshire a velocity of 4.86 km s-l for low velocity Lower Palaeozoic
rocks was also encountered (see Section 3). Thus in the ray-tracing analysis the low velocity
layer consisting of ORS, Lower Palaeozoic and, possibly, Precambrian rocks was assigned a
velocity of 4.8 km s-l . Across the Bristol Channel there is a lateral transition of ORS into
marine Devonian rocks which havea higher velocity (Brooks et al. 1977) and the low velocity
layer, mainly representing ORS, is thus truncated under the Bristol Channel (Fig. 8).

Downloaded from https://academic.oup.com/gji/article/78/3/661/648611 by guest on 24 May 2022

Under the southern Bristol Channel, from thickness estimates of the Devonian in nearby
north Devon (Edmonds, McKeown & Williams 1975), it is thought that the 5.5kms-'
refractor, occurring at a depth of about 2 km, represents approximately the top of the Lower
Palaeozoic. Between shots 6 and 7 the 5.5 km s-l refractor is interpreted to lie directly
beneath the Mesozoic rocks of the Bristol Channel syncline, suggesting that the syncline is
located over a pre-existing anticlinal structure from which the Upper Palaeozoic section has
been eroded. Under Cower and the northern Bristol Channel, it is uncertain whether the low
velocity layer includes a few hundred metres of Lower Palaeozoic rocks or whether the
5.5 kms-' refractor, at 1.5-3 km depth, represents the top of the Lower Palaeozoic succes-
sion. Beneath the coalfield, the 5.5 kms-l refractor falls to a depth of 4-5 km and eventually
wedges out north of station 7 (Fig. 8).
By virtue of its velocity and depth, the refractor propagating the P2 phase is thought to
represent Precambriancrystalline basement rocks(cf Section 3). Except for a local deepening
beneath the coalfield, this refractor dips southwards from about 2.5 km in the north to
around 7.5 km in the south (Fig. 8). Hence, although the 5.5 km s-l refractor may not
represent a constant geological horizon along line B2, it is safe to conclude from the above
discussion that there is a southwards thickening of the Lower Palaeozoic/?Precambrian
supra-basement sequence under Cower and the Bristol Channel (Fig. 8). To account for the
delay in arrivals times at shot 9, the basal refractor exhibits a step feature which may be the
continuation at depth of the Bristol Channel fault zone (Fig. 8). The only reason why
the model does not show the step feature in the 5.5 km s-l refractor is that the model does
not test its existence.
Line B2 crossed several E-W trending refraction lines (Figs 1 and 8) and except for the
case of lines F2 and F (see Section 5), depths to the refractors propagating the P1and P2
phases at crossover points with other lines are in agreement to within 1.5 km (Table 2).

4.2 L I N E c2

Line C2, 6 0 km long, was planned to complement earlier line C (Fig. 1). Nine marine shots
were fired in an E-W line across the northern Bristol Channel sited, as often as possible, on
Carboniferous rocks to avoid the thick Mesozoic sequences further south (Figs 2 and 11).
On land, 10 recording stations were set up across the western Vale of Glamorgan sited either
on Carboniferous Limestone or on a thin veneer of Mesozoic rocks (Fig. 11).
As on line B2 phase Pi,with a velocity of 6.00 f 0.15 km s-l, is observed at short range
(Fig. 9a, b). On line C2 it occurs between stations 1 and 5, near Cornelly Quarry, in the same
area that Bayerly & Brooks (1980) and Llewellyn (1981) found a similar phase associated
with the local Carboniferous Limestone. Between shot 2 and station 3 phase Po is observed
22
674 J. Mechie and M. Brooks
Table 2. Comparison of depths (in km) to the refractors propagating the P, and P , phases at crossover
points between the various seismic refraction lines.
Phase PI p2

Line Cross-profile Profde B2 Cross-profile Profile B2

Ba 2.i15 2.2-3 - -

CN - - 2.9- 3.1 2.2

A 1.6 -2.5' 2.7 5.g3 5.5
C2 2.4' 2.4 5.44 5.0
C 2.25 2.2 - -
D 2.2' 1.7 5.2' 5.5
F2 - - 2.34 ?7.5
F 0.9l 1.9 2.2' ?7.5

Downloaded from https://academic.oup.com/gji/article/78/3/661/648611 by guest on 24 May 2022

Pro file C2 Profile C2

C 2.25 2-2.5 - -
CY (- PN) - - 3.g5 4.2
CY (- AW) 1.ti3 2.3 4.7-5.3) 4.2
B2 2.4' 2.4 5.04 5.4

Profile J Profile J

D 2.2' 2.5 5.2' 5.8

CY-AW 0.9-1.72'3 2.4 4.1-5.42'3 5.5
LISPB - - 6' -

Profile I Profile I

E 1.25 1.6 6.7' 5.6

H - - 6.5- 10.2' 5.6
LISPB - - 66 5.6
' Brooks et al. (1 977). * Bayerly (1978).
Bayerly & Brooks (1980). This study.
Llewellyn (1981). K. R. Nunn (private communication).
a Depths do not include Upper Carboniferous.

with a velocity of 5.12 f. 0.09kms-' (Fig. lOa,b). If, as on line B2, the Po (and Pi) phases
are assumed t o propagate through Carboniferous Limestone, the limestone must exist under
the Mesozoic cover between shot 2 and station 3. Between shots 2 and 9 and stations 5 and
10 no phase through the upper layers was observed due to the rapid decay in amplitude of
the Po and PA phases. However, as these shots and stations were sited on either Carboniferous
Limestone or a thin veneer of Mesozoic rocks, and as a velocity for phase Po of 5.1 km s-'
through the local Carboniferous Limestone was observed on line B2 where it crossed line
C2, a velocity of 5.1 kms-' was used for the upper layers in the various interpretations (see,
e.g. Fig. 11).
Phases P I and P2 are also represented along line C2. Phase PI has an interstation apparent
velocity of 5.44 f 0.15 km s-' (Fig. 9a-c) and an intershot apparent velocity of
5.49 f 0.15 km s-' (Fig. 10a, c-f), and phase P2 has an interstation apparent velocity of
6.05 k 0.07 km s-' (Fig. 9d-f) and an intershot apparent velocity of 5.88 f 0.06
km s-l (Fig. 1Oa-f). Planar-layer interpretation yielded true velocities for these phases of
5.46 f 0.1 5 km s-' and 5.96 f 0.25 km s-' , which were subsequently used in the ray-tracing
analysis (Fig. 1 1). As the planar-layer interpretation indicated that the refractors propagating
 2

Downloaded from https://academic.oup.com/gji/article/78/3/661/648611 by guest on 24 May 2022

A

D I S T A N C E IN KM
Figure 9. Normalized, band-pass ffltered record sections for line C2 for shots (a) 1 , (b) 2, (c) 3 , (d) 5,
(e) 7 and (0 9 to all stations, plotted with a reduction velocity of 6 km s-', and showing least-squares
fits of the P-wave arrivals.
 J, Mechie and M. Brooks

-
tD
- 55 60 65
\
2-
X
I
et- 1 -
Y

Downloaded from https://academic.oup.com/gji/article/78/3/661/648611 by guest on 24 May 2022

W
E 0 I
2-
0
w
d05 1 -
W
OZ
0
5
D I S T A N C E I N KM
2-
u
% 1 -
z
c(
I I
- 0
5 ' l b 1 5 ' 20 525 30 35 4b 45 50 55
X

+ -

--
I
$20 25 30 36 40' 45' 50
DISTANCE IN K M
Figure 10. Normalized, band-pass filtered record sections for line C2 for stations (a) 2, (b) 3 , (c) 5 , (d) 6 ,
(e) 8 and (f) 10 to all shots, plotted with a reduction velocity of 6 km s-', and showing least-squares fits
of the P-wave arrivals.
 A seismic study in the Bristol Channel area 677
w Bristol Channel Vale of Glamorgan E
82 CY
I I
Shots 5.1 34 Stations
9 8 7 6 5 1 L 3 4 2 1 1 2 3 L 5 6 7 8910

u
Figure 11. Combined ray-tracing and geological model for line C2. Key as for Fig. 8.

the P1 and P2 phases dipped respectively east and west, these dips were incorporated in the

Downloaded from https://academic.oup.com/gji/article/78/3/661/648611 by guest on 24 May 2022

ray-tracing analysis.
The average absolute differences between calculated and observed travel-times for the
final ray-tracing model (Fig. 11) are 0.05 and 0.07 s for the P1 and P2 phases respectively
(Table 1). Thickness estimates of Mesozoic rocks beneath line C2, incorporated in the model,
were obtained from Bluck (1965), George (1970), Brooks & James (1975), and C. R. Price
(private communication).
Beneath line C2 the thickness of the Carboniferous Limestone, which decreases from
900m in SW Cower to 8 0 0 m in the Vale of Glamorgan (George 1970), is too small t o
account for the delay times t o the refractor propagating the PI phase. Thus, as for line B2
and the quarry blast line CY-AW in the Vale of Glamorgan (Bayerly & Brooks 1980), ORS
with a velocity of 4.8 km s-' was included in the sequence above this refractor. The thickness
of the low velocity layer from the ray-tracing analysis is 1.0-1.5 km and this thickness is in
agreement with local estimates of ORS thickness (George 1970) and measured thicknesses
in boreholes about lOkm north of line C2. Thus it is thought that beneath line C2 the
5.5 kms-' refractor at a depth of 2-2.5 km (Fig. 11) represents the top of the Lower
Palaeozoic sequence.
The refractor propagating the P2 phase dips west from 3.7 t o 6 km (Fig. 1 I), and because
of these depths and its velocity, it is taken to represent Precambrian crystalline basement
rocks (cf: Section 3). Thus the Lower Palaeozoic/?Precambrian supra-basement sequence is
interpreted to thicken from 1.4 km in the east to 3.5 km in the west.
Depth estimates to the refractors propagating the P1 and Pz phases at crossover points
with other seismic lines agree to within 1.I km (Table 2).

4.3 LINE C Y - A

In 1977, three shots from Cornelly Quarry in the Vale of Glamorgan were recorded on
Cower along E-W trending line A at ranges of 30-44km (Fig. 1). Cornelly Quarry is sited
on Carboniferous Limestone and the stations were sited on Carboniferous Limestone and
ORS (Fig. 2). On the record section (Fig. 12a), phase P2 can be recognized with an apparent
velocity of 6.47 f 0.22 kms-' . Thus the arrivals could be used in the regional time-term
analysis of the 6.0-6.2 km s-' refractor discussed below (Section 6).

4.4 LINE J

This SW-NE trending, 57 km long line (Fig. 1) was unreversed with one marine shot-point
and 2 4 recording stations on land. The south-western end of the line, including the shot-
point in the Bristol Channel, crossed Mesozoic rocks up to 300 m thick (Brooks & Al-Saadi
678 J. Mechie and M. Brooks

u
a: 1 -
-
z
n 1 I I I I
- "0 5 10 15 20 25 3d '35 ' 40' 45

Downloaded from https://academic.oup.com/gji/article/78/3/661/648611 by guest on 24 May 2022

DISTANCE I N KM
Figure 12. Normalized, band-pass filtered record sections for Lines (a) CY-A, (b) J and (c) I, plotted with a
reduction velocity of 6 km s - l , and showing least-squares fits of the P-wave arrivals.

1977; C. R. Price, private communication) resting unconformably on the Palaeozoic sequence

(Fig. 13). The line then crossed the Carboniferous Limestone and ORS outcrops immediately
SE of the South Wales Coalfield before crossing the western limb of the N-S Usk anticline
exposing Silurian rocks in its core (Figs 2 and 13).
Phase Po,with a velocity of 5.19 f 0.1 1 kms-I , is observed between the shot-point and
station 5 after which its amplitude decays rapidly (Fig. 12b). As on lines B2 and C2, this
phase is thought to represent arrivals through Carboniferous Limestone which emerges from
beneath a Mesozoic cover NE of station 6 (Fig. 13). Phases P1 and Pz are observed with
apparent velocities of 5.42 f 0.08 and 5.83 f 0.1 1 kms-I respectively (Fig. 12b), and
horizontal interpretation yields depths of 3.9 and 5.5 km to the relevant refractors.
Although line J had only one shot-point, a ray-tracing model (Fig. 13) which took into
account the varied surface geology, was computed, resulting in average absolute differences
between calculated and observed travel-times of 0.04 s for both phases PI and Pz (Table 1).
The thickness of the Carboniferous Limestone beneath line J decreases from 700 m under
the shot-point (George 1970) to zero (or almost zero) north of station 7. Hence, as on lines

sw
NEL/SPB CY;AW
'itntions 5.2 L.1
2 P
0
2
1
6
8krn

Figure 13. Combined ray-tracing and geological model for line J . Key as for Fig. 8.
 A seismic study in the Bristol Channel area 679
B2 and C2, much of the delay time to the refractor propagating the P1 phase must be
accounted for by ORS. From beneath the shot-point to beneath station 17 the 4.8 kms-'
layer has an almost constant thickness of 2 km (Fig. 13). Local estimates of ORS thickness
(George 1970) and thicknesses encountered in nearby boreholes range from 1-1.6 km. Thus
it is possible that the refractor propagating the P1 phase represents the top of the underlying
Lower Palaeozoic sequence (Fig. 13), although the possibility that the 4.8 km s-' layer
includes a few hundred metres of Lower Palaeozoic rocks cannot be ruled out. North-east
of station 17, the ORS wedges out until, between stations 21 and 23, Silurian rocks crop out.
Thus in the ray-tracing model the 4.8kms-' layer was thinned out and the 5.4kms-'
refractor brought to the surface where Silurian rocks crop out (Fig. 13). The low velocity
of 5.83 km s-l for the P2 phase was assumed to be a downdip apparent velocity and a true
velocity of 6kms-' was selected (Fig. 13), in accord with velocities on lines B2 and C 2 .

Downloaded from https://academic.oup.com/gji/article/78/3/661/648611 by guest on 24 May 2022

Assuming a local southerly dip of the refractor at the southern end of line J in agreement
with the surface geology (Figs 2 and 13), the refractor is interpreted to lie at depths of
5-7.5 km, the overall refractor geometry being shown in Fig. 13. As on lines B2 and C 2 ,
the 6 km s-l refractor is thought to represent Precambrian crystalline basement rocks and
on this basis the Lower Palaeozoic/?Precambrian supra-basement succession is interpreted
to thicken north-eastwards from 3 t o 5.5 km.
At crossover points with other seismic lines, depth estimates to the refractors propagating
the P1 and P2 phases agree t o within 1.5 km (Table 2). It should be noted that the LISPB
line (Bamford et al. 1976) crossed line J near its north-eastern end where there is no informa-
tion from line J about the structure of the 6.0kms-' refractor. Results from the LISPB
line (K. R. Nunn, private communication) indicate that the 6kms-' refractor shallows
north-eastwards from about 7.3 km depth beneath station 17 to around 6 km depth beneath
the crossover point (Fig. 13).

4.5 LINE I

Line I, 6 0 km long, extended WNW-ESE along the Mendip Hills and split into two branches
at its eastern end (Fig. 1). From one marine shot-point a single-ended seismic line was estab-
lished on land with 30 recording stations, of which 2 3 (1 -1 8 and 26-30) form the main line
and seven (19-25) occupy a subsidiary branch following the curve of the Mendip structural
axis (Fig. 1). In addition, an offend shot was fired along the trend of line I about 4 krn
WNW of the other shot-point. This shot was observed by the five most easterly recording
stations on the two branches of the line (Fig. 1).
The shots in the Bristol Channel were sited on Keuper Marl (Lloyd et al. 1973) with a
local thickness of 100-200 m (Brooks& Al-Saadi 1977). On land, line I ran along the Mendip
Hills which comprise a series of periclinal Variscan structures emerging from beneath a
Mesozoic cover (Figs 2 and 14). Although the main Palaeozoic rock-type exposed is Carboni-
ferous Limestone, ORS crops out in the cores of the four individual periclinal folds. Silurian
rocks, including volcanics, are also exposed in the core of the easternmost fold - the Beacon
Hill pericline. At its eastern end, line I crossed back on to Mesozoic rocks.
Interpretation was based mainly o n observations from the inner shot-point along the main
line (including the southern branch) for which the record station is shown (Fig. 12c). Phase
Po with a velocity of 5.1 kms-' is observed between the inner shot-point and station 2, after
which its amplitude decays rapidly (Fig. 12c). As stations 1 and 2 were sited on Carboni-
ferous Limestone, the phase is considered to represent this geological unit, as on lines B2,
C2 and J. Phases PI and P2 are observed with apparent velocities of 5.43 t 0.06 and 6.01
0.lOkms-' respectively (Fig. 12c). Horizontal interpretation yields a depth of 2.3 km to
680 J. Mechie and M.Brooks
M e n d i p H i l l s ESE
LISPB
I

S 0

w
u
,??.>0000000(7TF
0

Figure 14. Combined ray-tracing and geological model for line 1. Key as for Fig. 8.

the refractor propagating the P1 phase and a depth of 5.7 km, beneath both the northern

Downloaded from https://academic.oup.com/gji/article/78/3/661/648611 by guest on 24 May 2022

and southern branches of the lines, to the refractor propagating the P, phase. For arrivals
from the 6 kms-' refractor at the outer shot-point there is a time delay of about 0.1 s with
respect to arrivals at the inner shot-point. This time delay can be explained by an increase of
1-1.5 km in the depth to the 6 k m s-l refractor between the inner and outer shot-points.
Again, a ray-tracing model (Fig. 14) which took into account the varied surface geology
was computed, resulting in average absolute differences between calculated and observed
travel-times of 0.03 and 0.02 s for phases P1 and Pz respectively (Table 1).
Along Line I, a maximum thickness of 6 5 0 m of Carboniferous Limestone occurs beneath
station 11. In contrast, where line I crosses the periclinal folds of North Hill, Pen Hill and
Beacon Hill, the Carboniferous Limestone is absent and ORS and Silurian rocks crop out.
Thus, as on lines B2, C2 and J, in addition to Carboniferous Limestone, an ORS layer must
contribute to the delay times associated with the P1 phase. The 4.8kms-' layer has an
average thickness of 1.3 km and the 5.4km s-l refractor occurs at about 1.6km depth
(Fig. 14). Thickness estimates of about 5 0 0 m are given for the ORS in the area including the
Beacon Hill pericline, between stations 16 and 27 (Fig. 14), where the top and base of the
local ORS succession can be seen (Green & Welch 1965; Kellaway & Welch 1948). Thus,
unless significant variations of ORS thickness occur locally, the 5.4 km s-l refractor must lie
0.5-1 km below the top of the Lower Palaeozoic sequence (Fig. 14). Between stations 17
and 26, Silurian rocks crop out. However, as the most easterly sampled point giving rise to
a first arrival from the 5.4kms-' refractor occurs beneath station 14, it cannot be tested
whether or not the 5.4 k m s-l refractor comes to the surface between stations 17 and 26.
The refractor propagating the P, phase, interpreted as Precambrian crystalline basement
rocks as on profiles B2, C2 and J , occurs at a depth of 5.6 km (Fig. 14). The 5.4 km s-l layer
is about 4 k m thick and thus the total thickness of the Lower Palaeozoic/?Precambrian
suprabasement sequence is about 4.5-5 km.
Except for line H (Llewellyn 1981) depth estimates to the refractors propagating the PI
and P, phasesat the crossover point with other seismic lines agree to within 1.1 km (Table 2).

5 The line in SW England - F2

This 8 8 k m long, E-W trending line was planned as an extension of earlier line F (Fig. 1).
Eight marine shots were fired across the southern Bristol Channel from off the north Devon
coast towards Lundy while 10 recording stations were set up on land in west Somerset. At
its western end, line F2 crossed Tertiary rocks about 200m thick underlain by Mesozoic
rocks about 3 0 0 m thick (Brooks & James 1975). Eastwards the Tertiary and Mesozoic
sequences thin out and the eastern four shots were located on Devonian rocks (Fig. 17).
Between shot 1 and recording station 1, which were located 18 km apart, line F2 crossed the
 1 -

0
55 0

1 -

0 '

Downloaded from https://academic.oup.com/gji/article/78/3/661/648611 by guest on 24 May 2022

50

D I S T A N C E IN K M

I
JU -lU -lU uu uu uu uu
7n
I U
z-
A

1 -

0-
20
2-
25' so ' 35 ' 40' '45 50
I
55

1 -

0
15 20' - + 30
$5 '40 a 4 5' so
D I S T A N C E IN KM
Figure 15. Normalized, band-pass filtered record sections for line F2 for shots (a) 1 , (b) 2, (c) 4 , (d) 5 ,
(e) 7 and (f) 8 to all stations, plotted with a reduction velocity of 6 km s-I, and showing least-squares
fits of the P-wave arrivals. Phase PP arrivals are marked by -.
682 J. Mechie and M. Brooks

I I

- 135 40

2!
I c I I

Downloaded from https://academic.oup.com/gji/article/78/3/661/648611 by guest on 24 May 2022

35 '40 4% 5

D I S T A N C E IN K M
Figure 16. Normalized, band-pass filtered record sections for line F2 for stations (a) 2, (b) 4 , (c) 6 , (d) 7 ,
(e) 8, (f) 10 to all shots and (g) a selection of seismograms showing the reflected phase PP. Sections are
plotted with a reduction velocity of 6 km s-'and (a-f) show least-squares fits of the P-wave arrivals. In
(a-0, phase PP arrivals are marked by -.

fold axis of the north Devon anticline. The recording stations were located on the northern
limb of this anticline on Devonian rocks or on overlying Triassic rocks up to 2 5 0 m thick
(Thomas 1940).
The only refracted phase that can be observed in the record sections is phase P2 with an
interstation apparent velocity of 6.10 f 0.02kms-' (Fig. 15a-f) and an intershot apparent
velocity of 6.33 +0.03kms-' (Fig. 16a-f). Using a top layer velocity of5.54kms-',whichis
an average of the two velocities, 5.43 and 5.65 km s-', for the layers above the 6.2 km s-l
refractor on line F (Brooks e f al. 1977, fig. l), a planar-layer interpretation yielded a true
velocity of 6.21 f 0.04kms-' for phase P2.The refractor is interpreted to dip east from a
depth of 800 m in the west to 4.7 km in the east (Fig. 17). Where lines F2 and F approxi-
mately coincide, depth estimates to the 6.2 km s-' refractor of 2.3L2.8 km were obtained for
line F2, while for line F an average depth of 2.1-2.2 km was obtained (Brooks e f al. 1977).
Where line B2 crosses lines F2 and F, it appears that there may be a strong disagreement
between the interpretations (Table 2). A particular problem is whether or not the 6 km s-l
refractor occurring at 7-8 km depth under the southern part of line B2 represents the same
geological interface as the 6.2 km s-' refractor occurring at 2-3 km depth beneath the shot-
points of lines F2 and F. If the two refractors represent the same geological interface and
that interface is the top surface of the Precambrian crystalline basement (cf: Section 4.1, line
B2), the latter must occur at shallow depths under lines F2 and F (Fig. 8b). There must be a
rapid southward reduction in the thickness of the Lower Palaeozoic/?Precambrian supra-
basement sequence from about 5 km beneath the central Bristol Channel to only a few
hundred metres under the north Devon coast. Even if Fig. 8(b) shows the true model,
 A seismic study in the Bristol Channel area 683

Downloaded from https://academic.oup.com/gji/article/78/3/661/648611 by guest on 24 May 2022

0

---+-
e 1 I ! I1 I 4 1 I ! 1
0 '
35' 401 45 i50 6$ 70 175 !SO 85
1 I
DISTRNCE IN KM
Figure 16 - continued

W Bristol Channel West somerset E

2.1-3.7 ?2 Shots 3.7 Stations
8 7 6 5 L 3 2 1

-
5.5L

6.21
0 lOkm

Figure 17. Planar-layer model, including the Mesozoic and Tertiary rocks, the line F 2 . Velocities are in
km s-'.

however, line B2 does not continue far enough south to detect the shallow 6.0-6.2 km s-l
refractor beneath lines F2 and F. The ray-path ADB, which does not intersect the
6.0-6.2km s-l refractor at shallow depths, has approximately the same travel-time as the
ray-path ACB (Fig. 8b).
If the two refractors do not represent the same geological interface and if the 5.5-
5.6 km s-l refractor detected under lines F and B2 represents the top of the Lower Palaeo-
zoic sequence (cf Section 4.1, line B2), the shallow 6.2kms-' refractor beneath lines F2
and F probably represents a rock-type occurring a few hundred metres below the top of the
Lower Palaeozoic sequence and terminating a short distance north of line F2, possibly along
684 J. Mechie and M. Brooks
a major fault or thrust (see below). Alternatively, if the 5.5-5.6 kms-' refractor represents
a horizon in the Lower Devonian sequence, then the shallow 6.2 km sT1 refractor may repre-
sent a high velocity rock-type also occurring in the Lower Devonian sequence (Matthews
1981).
In addition to the first arrival phase P2 a reflected phase PP was recognized along line F2
(Fig. 16g). The interpretation of this reflected phase is made uncertain as no associated
refracted phase can be observed, possibly because of the 'screening' effect of the shallow
high velocity refractor or because line F2 was not long enough for the associated refracted
phase to occur as a first arrival. Nevertheless, assuming that velocity varies only with depth,
an average one dimensional interpretation of the reflected phase PP was attempted. Synthetic
seismograms (Fig. 19) using the ray theoretical method (cerveny, Molotkov & PInEik 1977)
and travel-time curves (Fig. 16g) for the velocity-depth ( V - 2 ) function (Fig. 18) are

Downloaded from https://academic.oup.com/gji/article/78/3/661/648611 by guest on 24 May 2022

presented. Differences between observed and theoretical travel-times are less than 0.17 s
(Fig. 16g), while the synthetic seismograms show, in agreement with the observed data, that
the amplitudes of phase PP are much greater than those of phase P2 out t o a distance of
about 75 km (Fig. 19).

V e l o c i t y (km/sl
5 6 7
0

Figure 18. Average velocity-depth function for line F2.

3 .O
-
'02.0 '

9
Yl.0 '
_-1
----
-
a
-0.0 '
-__ - ' , - -*- - - - ,
-'
.------,--
-* PP
.

-I .o.
5
-2.0,*0 20. 40 .O 80. 80. 100.0

In the I/-2 function (Fig. 18) the 6.2 km s-l layer has, at 1.5 km, a thickness which is
considered to be great enough to propagate the P2 phase, with a frequency of at least
8-10 Hz, over adistanceofabout 100km. Taking the critical distance for the reflected phase
PP to be about 45 km and assuming a thickness of 1.5 km for the 6.2 k m s-l layer, a low
velocity layer must occur between the 6.2 and 6.3 km s-l layers. The average velocity of the
low velocity layer is calculated to be 5.74kms-', which is considerably less than the true
velocities o f 5.96-6.21 km s-l derived in this region for Precambrian crystalline basement
 A seismic study in the Bristol Channel area 685
rocks. Thus, in this model (Fig. 18), the 6.2 and 5.74kms-' layers are taken to represent
part of the Lower Palaeozoic sequence. The 6.3 km s-l basal layer at a depth of 7.6 km is
taken to represent Precambrian crystalline basement rocks.
The postulated velocity structure below a depth of about 5 km along line F2 is in approxi-
mate agreement with the velocity structure determined beneath the southern part of line B2
and thus represents one possible way of resolving the apparent differences of structure
beneath line B2 and lines F2 and F.
In considering the possible geological significance of the V-2 function (Fig. 18), it is
interesting t o recall earlier attempts to explain the northward fall of Bouguer gravity values
across Exmoor, north Devon (Bott, Day & Masson-Smith 1958). It was suggested by Bott
et al. (1958) and by later workers (Bott & Scott 1964; Brooks & Thompson 1973) that the
gravity gradient may be caused by the overthrusting of the Devonian sequence of north Devon

Downloaded from https://academic.oup.com/gji/article/78/3/661/648611 by guest on 24 May 2022

over a concealed sedimentary sequence of lower density. The proposed V-2 function under
line F2 is entirely compatible with such a structural model and the base of the shallow high
velocity layer might represent the basal unit of the overthrust sheet. The underlying low
velocity layer would then represent autochthonous Palaeozoic strata and the basal high
velocity layer Precambrian crystalline basement, equivalent to the basal refractor encountered
in other parts of the region. Northward wedging out of a thrust sheet would account con-
veniently for the absence of the shallow high velocity layer under lines B2 and D.

6 Regional time-term analysis

A time-term analysis was attempted t o determine the depths to and average velocity of the
Precambrian crystalline basement across the region. In this analysis, data from quarry blasts
including those from lines CN/BN-GH, PN-CN, and CY-A (Bayerly & Brooks 1980), the
1975 cruise lines A and D (Brooks e t al. 1977; Bayerly & Brooks 1980), the 1976 cruise
lines B2, K and L, and the 1978 cruise were included. Mechie (1980) provided details of the
analysis, from which a contoured version of the map of Precambrian crystalline basement
depths (Brooks et al. 1983, fip. 10.4) is presented (Fig. 20). The average velocity for the

S O U T H W A L E S

. I d 20

B R I S T O L
C H A N N E L

6.2.

Figure 20. Contoured map showing Precambrian crystalline basement depths (in km), derived from time-
term analysis, under the South Wales/Bristol Channel area. Main basement highs are marked by +.
686 J. Mechie and M. Brooks
Precambrian basement obtained from the analysis was 6.06 km s-l, which is within the range
of 5.96-6.2 1 km s-l for true velocitiesdetermined by other methods for Precambrian crystal-
line basement in this region. In converting time-terms to depths an average top layer velocity
of 5.28 kms-' was used (cf: Bayerly & Brooks 1980).
The pattern of basement depths obtained from the time-term analysis is similar to that
obtained from the planar-layer and ray-tracing interpretations. The shallowest depths to the
basement occur in north Pembrokeshire, in the vicinity of the Hayscastle anticline where
Precambrian rocks crop out (Fig. 20), and the basement deepens to 4-4.5 km beneath south
Pembrokeshire. Under the Johnston-Benton fault block, where Precambrian intrusive
igneous rocks crop out, the time-term analysis yields depths of 2-5 km to the basement
(cf: Section 3). In Carmarthen Bay, the basement shallows eastwards from 4-4.5 km depth
near Caldy t o about 3 km in central Carmarthen Bay.

Downloaded from https://academic.oup.com/gji/article/78/3/661/648611 by guest on 24 May 2022

Shallow basement depths of about 3 km occur beneath the central and north-western
parts of the South Wales Coalfield. From this area, the basement deepens southwards to
around 6 km beneath the south-western part of the coalfield, 5-7 km beneath Cower, and
4-4.5 km beneath the Vale of Glamorgan. Under the Bristol Channel basement depths are
generally in excess of 6 km, while a depth of 7.5 km is encountered off the north Devon
coast. Beneath the central part of the coalfield where the basement occurs at a depth of
about 3 km, thicknesses of about 2 km of Upper Carboniferous rocks are preserved. Thus
the pre-Upper Carboniferous supra-basement sequence must thicken southwards from
about 1 km beneath this area to more than 4 km in Cower and the Bristol Channel area.

7 Discussion
The seismic refraction results show that beneath lines B2, C2 and J in the central South
Wales/Bristol Channel area and line I in the Mendip Hills the structure is similar. Below the
Carboniferous Limestone with a normal velocity of 5.1-5.2 kms-', the ORS, with a velocity
of 4.7-4.85kms-', and sometimes parts of the Lower Palaeozoic succession act as a low
velocity layer. Below this low velocity layer, a refractor propagating phase P1 with a velocity
of 5.4-5.5 kms-' is commonly detected. This refractor is considered to represent a horizon
at or near the top of the Lower Palaeozoic sequence and it could possibly come to the surface
at the northern end of line J (Fig. 13) where Silurian rocks crop out in the Usk anticline.
Beneath the northern end of line B2, the 5.4-5.5 kms-' layer wedges out which is in agree-
ment with the absence of a 5.4-5.7kms-' refractor under the South Wales Coalfield lines
CN/BN-GH and CN-PN (Bayerly & Brooks 1980). Thus, the refractor seems to be absent
below the central and northern parts of the coalfield and it was also undetected along lines
K and L in Pembrokeshire, where top layer velocities of 4.9-5.2 km s-', representing very
varied sequences o f Palaeozoic and Precambrian rock-types, were obtained.
Beneath most lines in the region a basal refractor, propagating phase P2 with a velocity
of 6.0-6.2 km s-l, has been detected (Table 2). Except under lines F2 and F in SW England,
it is considered safe t o correlate this refractor with the top surface of Precambrian crystal-
line rocks, forming a seismic/geological basement beneath the region. Broadly speaking, the
refractor deepens southwards (Fig. 20). Under lines F2 and F , the shallow 6.2 km s-' refractor
could possibly represent Precambrian rocks (Fig. 8) but, taking account of the results from
line B2 and interpretation of the reflected phase PP observed on line F2, it is considered
more likely that the refractor represents a horizon within the Palaeozoic sequence.
In north Pembrokshire, Precambrian igneous rocks crop out in the Hayscastle anticline
(Fig. 2). The seismic experiment has shown that in the vicinity of the Hayscastle anticline the
Precambrian crystalline basement occurs at a depth of about 1 km (Figs 4 and 20). As
 A seismic study in the Bristol Channel area 687
Upper Carboniferous rocks crop out next to the Precambrian rocks of the Hayscastle anti-
cline without a major fault boundary being involved (Hancock et al. 1983, fig. 3.3) this
basement high must have existed during the late Carboniferous. In the central part of the
South Wales Coalfield, the Precambrian crystalline basement occurs at about 3 km depth
(Fig. 20; Bayerly & Brooks 1980). As the Upper Carboniferous sequence is about 2 km thick
in this area, the pre-Upper Carboniferous supra-basement sequence must be only about 1 km
thick. Thus in the central part of the main coalfield, a basement high must similarly have
existed in the late Carboniferous.
Combining geological, seismological and aeromagnetic evidence, Brooks et al. (1983) pro-
posed the existence in South Wales of a series of basement highs in an E-W zone, of which
the Hayscastle anticline and the central part of the South Wales Coalfield can be shown to
have existed before the main movements of the Variscan orogeny. As shown by the seismic

Downloaded from https://academic.oup.com/gji/article/78/3/661/648611 by guest on 24 May 2022

experiments, the basement deepens beneath south Pembrokeshire, Cower, the Vale of
Clamorgan, and the Bristol Channel and, consequently, the pre-Upper Carboniferous supra-
basement sequence thickens (Figs 20, 4 and 8). This zone of shallow basement was perhaps
responsible for controlling the position of the Variscan orogenic 'front' which runs through
South Wales t o the south of the basement highs (Rast 1983, fg. 1.3; Hancock et al. 1983)
and which is used to delineate the northern boundary of the Variscan orogenic belt.
From thickness estimates of the Upper Palaeozoic sequence in the central South Wales
area (see Sections 1 and 4), it can be deduced that the pre-Upper Palaeozoic sequence,
possibly including Precambrian strata, also thickens southwards from almost zero where the
basement high exists in the central part of the coalfield (Bayerly & Brooks 1980; Brooks et
al. 1983) to 1-3 km beneath the Vale of Glamorgan (Figs 11 and 13; Bayerly & Brooks
1980), 2-5 k m beneath Cower and the northern Bristol Channel (Figs 8 and 11;Brooks et
al. 1983), and 3-5 km beneath the middle and southern Bristol Channel (Fig. 8 ; Brooks
et al. (1 983). In Pembrokeshire, the pre-Upper Palaeozoic supra-basement sequence,
including Precambrian volcanic rocks, thickens southwards from a few hundred metres where
the basement high exists in the vicinity of the Hayscastle anticline to around 3.5 km at
the northern edge of the Pembrokeshire Coalfield (Fig. 4). In south Pembrokeshire, because
of the complicated geological structure and the absence of a refractor at or near the top of
the Lower Palaeozoic succession (cfi central South Wales), it is difficult to separate, on the
basis of the seismic results, the Upper Palaeozoic sequence from the pre-Upper Palaeozoic
sequence. However, to explain the position of the Precambrian crystalline rocks in the
Johnston-Benton block, the seismic interpretation requires at least 3 km of vertical move-
ment and 4-5 km of horizontal movement along the Johnston thrust during the Variscan
orogeny (Brookset al. 1983, Fig. 10.6).

Acknowledgments
This research was carried out while JM was at University College, Swansea, supported by a
NERC Research Studentship. Research grants to MB from the NERC and the Royal Society
were awarded for the purchase of explosives. The Captain, officers and crew of RRS John
Murray and MV Edward Forbes, the managers of Crwbin and Cornelly quarries, and many
land owners are thanked for their help and cooperation. The Universities of Cardiff and East
Anglia, the NERC geophysical pool, and the IGS Global Seismology Unit are thanked for
loaning seismic equipment. For their useful discussions and help, we would like to mention
Drs M. Bayerly, D. J. Llewellyn, P. Styles, A. M. Sykes, J. Cope, the late Mr P. Corles, Mr
D. Lancy and Mr M. Punter. The calculations were carried out at the Computer Centres of
the University College, Swansea and the Universities of Manchester and Karlsruhe. M. Bigott
typed the manuscript.
688 J. Mechie and M. Brooks
References
Archer, A. A., 1968. The geology of the South Wales Coalfield. Special Memoir. Gwendraeth Valley and
adjoining areas, Mem. geol. Surv. G.B.
Bamford, D., Faber, S., Jacob, B., Kaminski, W., Nunn, K., Prodehl,C., Fuchs, K., King, R. & Willmore, P.,
1976. A lithospheric seismic profile in Britain - 1. Preliminary results, Geophys. J. R . astr. SOC.,
44, 145-160.
Bayerly, M. A., 1978. Quarry blast seismic studies in South Wales, unpublished PhD thesis, University of
Wales.
Bayerly, M. & Brooks, M., 1980. A seismic study of deep structure in South Wales using quarry blasts,
Geophys. J. R . astr. SOC.,60, 1-19.
Berckhemer, H., 1976. Standard equipment for deepseismic sounding, in Explosion Seismology in
Central Europe, eds Giese, P., Prodehl, C. & Stein, A., Publication of the Deutsche Geophysikalische
Gesellschaft, Springer-Verlag. Berlin.
Berry, M. J. & West, G. F., 1966. An interpretation of the fist-arrival data of the Lake Superior experi-

Downloaded from https://academic.oup.com/gji/article/78/3/661/648611 by guest on 24 May 2022

ment by the time-term method, Bull. seism. SOC.A m . , 56, 141-171.
Bluck, B. J., 1965. The sedimentary history of some Triassic conglomerates in the Vale of Glamorgan,
South Wales, Sedimentology, 4, 225-245.
Bott, M. H. P., Day, A. A. & Masson-Smith, D., 1958. The geological interpretation of gravity and mag-
netic surveys in Devon and Cornwall, Phil. Trans. R . SOC.A . , 251, 161-191.
Bott, M. H. P. & Scott, P., 1964. Recent geophysical studies in south-west England, in Present Views on
Some Aspects of the Geology of Cornwall and Devon, pp. 25-44, Blackford, Truro, Cornwall.
Brooks, M. & Al-Saadi, R. H., 1977. Seismic refraction studies of geological structure in the inner part of
the BristolChannel, J. geol. SOC.London, 133,433-445.
Brooks, M., Bayerly, M. & Llewellyn, D. J., 1977. A new geological model to explain the gravity gradient
across Exmoor, north Devon, J. geol. SOC.London, 133,385-393.
Brooks, M. & James, D. G., 1975. The geological results of seismic refraction surveys in the Bristol
Channel, 1970-73, J. geol. SOC.London, 131,163-1 82.
Brooks, M., Mechie, J. & Llewellyn, D. J., 1983. Geophysical investigations in the Variscides of southwest
Britain, in The Variscan Fold Belt in the British Isles, ed. Hancock, P. L., Hilger, Bristol.
Brooks, M. & Thompson, M. S., 1973. The geological interpretation of a gravity survey of the Bristol
Channel, J. geol. SOC.London, 129, 245-274.
eerveng, V., Langer, J. & PbnELk, I., 1974. Computation of geometric spreading of seismic body waves in
laterally inhomogeneous media with curved interfaces, Geophys. J. R . astr. SOC., 38.9-19.
&men$, V., Molotkov, I. A. & PXenBk, I., 1977. R a y Method in Seismology, University of Karlova,
Prague.
Cope, J. C. W., 1977. An Ediacara-type fauna from South Wales, Nature, 268,624.
Cope, J. C. W., 1979. Theearly history of the Tywi anticline in the area south of Carmarthen, South Wales,
in The British Caledonides Reviewed, eds Holland, C. H., Leake, B. E. & Harris, A. L., Geological
Society of London, Scottish Academic Press, Glasgow.
Edmonds, E. A., McKeown, M. C. & Williams, M., 1975. South-West England, Br. reg. Geol., 4th edn.
Evans, D. J. & Thompson, M. S., 1979. The geology of the central Bristol Channel and the Lundy area,
South Western Approaches, British Isles, Proc. geol. Ass., 90(1), 1-14.
Fletcher, B. N., 1975. A new Tertiary basin east of Lundy Island, J. geol. SOC.London, 131,223-235.
George, T. N., 1970. South Wales, Br. reg. Geol., 3rd edn.
Green, G. W. & Welch, F. B. A., 1965. Geology of the country around Wells and Cheddar, Mem. geol.
Surv. G. B.
Hancock, P. L., Dunne, W. L. & Tringham, P. L., 1983. Variscan deformation in southwest Wales, in
The Variscan Fold Belt in the British Isles, ed. Hancock, P. L., Hilger, Bristol.
James, D. G., 197 1. Seismic refraction studies in parts of South Wales and the Bristol Channel, unpublished
PhD thesis, University of Wales.
Johnson, S. H., 1976. Interpretation of split-spread refraction data in terms of planedipping layers, Geo-
physics, 41,418-424.
Kellaway, G. A. & Welch, F. B. A., 1948. Bristol and Gloucester District, Br. reg. Geol., 4th edn.
Llewellyn, D. J., 1981. Geophysical investigations of the deep structure of the Bristol Channel and South
Wales, unpublished PhD thesis, University of Wales.
Lloyd, A. J., Savage, R. J. G., Stride, A. H. & Donovan, D. T., 1973. The geology of the Bristol Channel
floor, Phil. Trans. R . SOC.A , 274, 595-626.
Matthews, S. C., 1981. A cross section through southwest England, Geologie Mijnb., 60,145-148.
 A seismic study in the Bristol Channel area 689
Mechie, J., 1980. Seismic studies of deep structure in the Bristol Channel area, unpublished PhD thesis.
University of Wales.
Owen,T. R., 1974. The Upper Palaeozoic and Post-Palaeozoic rocks of Wales, University of Wales Press.
Rast, N., 1983. Variscan orogeny, in The Variscan Fold Belt in the British Isles, ed. Hancock, P. L.,
Hilger, Bristol.
Sanzen-Baker, I., 1972. Stratigraphical relationships and sedimentary environments of the Silurian-early
Old Red Sandstone of Pembrokeshire, Proc. geol. Ass., 83, 139-164.
Strahan, A., Cantrill, T. C., Dixon, E. E. L., Thomas, H. H. & Jones, 0. T., 1914. The geology of the
South Wales Coalfield. Part XI. Haverfordwest, Mem. geol. Sum. G.B.
Thomas, A. N., 1940. The Triassic rocks of north-west Somerset, Proc. geol. Ass., 51, 1-43.
Willmore, P. L. & Bancroft, A. M., 1960. The time-term approach to refraction seismology, Geophys. J.
R . astr. Soc., 3 , 4 19-432.

Downloaded from https://academic.oup.com/gji/article/78/3/661/648611 by guest on 24 May 2022