GEOPHYSICS, VOL. 86, NO. 3 (MAY-JUNE 2021); P. B181–B191, 9 FIGS., 2 TABLES.

10.1190/GEO2020-0336.1
Downloaded 11/09/21 to 80.217.127.130. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

Case History

Innovative seismic imaging of volcanogenic massive sulfide deposits,

Neves-Corvo, Portugal — Part 2: Surface array

George A. Donoso1, Alireza Malehmir1, Bojan Brodic1, Nelson Pacheco2, João Carvalho3, and
Vitor Araujo2
DOI:10.1190/geo2020-0336.1

ABSTRACT this, a carefully tailored processing algorithm was developed

Seismic methods are an affordable and effective way of study-
ing the subsurface for mineral exploration. With the goal of
testing new technologies for mineral exploration in highly chal-
lenging mining areas, in early 2019, an innovative seismic sur-
vey was conducted at the Neves-Corvo mine, south Portugal.
We have focused on the data and results from the surface array
data, whereas other work deals with the underground seismic
data. The surface seismic survey consisted of two perpendicular
2D profiles positioned above the known world-class tier-1 Lom-
bador deposit. Simultaneously, a survey inside the active under-
ground mine took place, being unique because it included the
testing of a prototype system that enabled accurate GPS-time
(microsecond accuracy) synchronization inside the mine tun-
nels, approximately 650 m below the surface profiles. Due to the
active mining operations, the surface data are noisy. To handle
and applied to enhance reflections in the data, interpreted to
originate from lithologic contacts and the Lombador deposit.
The results and interpretations from 2D processing were vali-
dated taking advantage of the known deposit geometry using 3D
exploding reflector modeling and pseudo-3D cross-dip analysis.
These analyses suggest that there is an out-of-plane signature of
the Lombador deposit on the surface data. Additionally, source
points activated in the exploration tunnels and simultaneously
recorded on the surface profiles allowed for the creation of a
2D velocity model that was used for migration and time-to-
depth conversion, providing a reliable 2D seismic section of
the subsurface under the surface profiles. We determine that lim-
ited surface coverage 2D surveys and a velocity model derived
from the tunnel-to-surface seismic recordings allow for imaging
of key subsurface geologic structures and delineating mineral
deposits of economic interest.

INTRODUCTION Yavuz et al., 2015) helping to guide mineral exploration decisions

In the past decade, seismic reflection methods have increasingly
been used for mineral exploration applications in hardrock environ-
ments. This success may be because of the high-density contrasts of
most metallic deposits relative to their host rocks (Salisbury et al.,
2000; Dehghannejad et al., 2012; Malehmir et al., 2012, 2013;
at various mining sites (Tryggvason et al., 2006; Urosevic et al.,
2012; Manzi et al., 2015; Place and Malehmir, 2016). Most early
works that examined the use of seismic reflection methods for direct
targeting of volcanogenic massive sulfide (VMS) deposits come
from Canada (Verpaelst et al., 1995; Milkereit et al., 1996; Adam
et al., 1997; Malehmir and Bellefleur, 2009; Bellefleur et al., 2018),

Manuscript received by the Editor 21 May 2020; revised manuscript received 26 January 2021; published ahead of production 10 February 2021; published
online 21 April 2021.
1
Uppsala University, Department of Earth Sciences, Villavägen 16, Uppsala 75236, Sweden. E-mail: george.donoso@geo.uu.se (corresponding author);
alireza.malehmir@geo.uu.se; bojan.brodic@geo.uu.se.
2
Somincor (Lundin Mining), Neves-Corvo Mine, Santa Bárbara de Padrões, Castro Verde 7780-409, Portugal. E-mail: nelson.pacheco@lundinmining.com;
vitor.araujo@lundinmining.com.
3
LNEG, Estrada da Portela-Zambujal, Apartado 7586, 2610-999 Amadora, Portugal. E-mail: joao.carvalho@lneg.pt.
© 2021 Society of Exploration Geophysicists. All rights reserved.

B181
 B182 Donoso et al.
whereas early attempts in Europe focused on imaging of the struc-
tural environment hosting the deposits (Dehghannejad et al., 2010,
Downloaded 11/09/21 to 80.217.127.130. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms
2012; Juhlin et al., 2012; Malehmir et al., 2012; Abdi et al., 2015).
With the large increase in computing processing power in the past
decade, it has been possible to revisit and reprocess legacy data us-
ing today’s technologies, particularly from European sites (Koivisto
et al., 2015; Balestrini et al., 2020; Bräunig et al., 2020; Markovic
et al., 2020; Papadopoulou et al., 2020). Legacy refers to the data
from before the systematic and standardized storage of digital data
became common practice in the late 1990s, and they are usually
lacking complete information regarding acquisition equipment,
field geometry setup, or support observation logs, or they are found
in a nonstandardized digital format.
Donoso et al. (2020) present the reprocessing of a 2D seismic
data set acquired in 1996, also from the Neves-Corvo mine. This
1996 data set is likely the first study in Europe to not only prove
the capability of directly imaging massive sulfide deposits using the
seismic method, but also to image various geologic structures such
as the ore-bearing Neves main thrust and associated lithostrati-
graphic units. As a result of the Donoso et al. (2020) study, the
world-class Lombador massive sulfide and other smaller deposits
were imaged with improved continuity and resolution. Furthermore,
a few never-before-seen shallow and steeply dipping reflections
were imaged, highlighting the value of legacy seismic data and
DOI:10.1190/geo2020-0336.1
the importance of taking advantage of state-of-the-art data process-
ing algorithms in hardrock environments.
With the motivation to continue the aforementioned study and
complement other ongoing geophysical studies (Marques et al.,
2019), and also with the purpose of testing new technologies for
deep targeting of massive sulfides, a pioneer study was conducted
at the beginning of 2019 at the Neves-Corvo mine in southern Por-
tugal, in the ongoing quest to apply seismic methods in an im-
proved, sustainable, and efficient manner, in highly noisy and
logistically challenging mining environments. The survey is consid-
ered challenging because the acquisition had to be performed ac-
commodating limited time, site accessibility, and activities of the
operating Neves-Corvo mine. The data and the work shown in this
article are expected to extend this effort by studying the area to the
north of the Lombador deposit, while at the same time showing the
imaging potential of a small-scale survey (i.e., short offsets and con-
ventional accelerated drop-hammer) implemented together with in-
novative acquisition and data processing technologies.
The search for mineral deposits is increasingly targeting greater
depths because many of the shallow resources, usually greater than
500 m depth (Schodde, 2020), have already been discovered or are
being exploited. Consequently, modern mineral exploration is be-
coming more focused in extending the knowledge of known depos-
its and helping to find new mineralization at depth for the future
(Dentith and Mudge, 2014) by increasingly turning to the applica-
tion of geophysical methods, in particular reflection seismic, which
is used routinely for hydrocarbon exploration. Seismic methods can
also complement electromagnetic (EM) surveys, in particular because
EM methods suffer greater losses in terms of resolution at depth and
also because they are affected by the fact that nonmineralized con-
ductive bodies may shield regions underneath from being resolved.
The Neves-Corvo mine, southern Portugal, has proven to be an
example of the use of seismic methods and the testing of new tech-
nologies for deep targeting of massive sulfides (Yavuz et al., 2015;
West and Penney, 2017; Donoso et al., 2020). In the present study,
two new 2D perpendicular profiles were acquired on the surface
above the known Lombador deposit, using an accelerated weight-
drop (AWD) as seismic source and wireless recorders as receivers;
the source points and receivers were set 10 m apart. In addition to
the surface profiles, the survey also tested a prototype system that
enables accurate global positioning system (GPS) time (microsec-
ond accuracy) synchronization in areas with no satellite signal; to do
this, four seismic profiles were also deployed inside mine galleries
located approximately 650 m below the surface profiles (Malehmir
et al., 2019a). Another unique aspect of the survey was the use of an
electrically driven, linear synchronous motor-based (Noorlandt
et al., 2015; Brodic et al., 2019) vibrator (E-vib) as a seismic source.
The E-vib was used along one of the mine tunnel seismic profiles
together with the new GPS-time system; this provides a common
time base to successfully obtain the corresponding data from the
source points activated inside the tunnel and recorded by the wire-
less seismic recorders along the two surface profiles. Both the GPS-
time system and the E-vib seismic source were developed within the
EU-funded Smart Exploration™ project (Malehmir et al., 2019b).
This paper presents the results of reflection seismic data processing
along the surface profiles, complemented with additional techniques
such as 3D exploding reflector forward modeling for verifying the
results and cross-dip analysis of the main reflections allowing us to
take into account the out-of-plane nature of the deposit that may not
be projected when stacking inline data only (Wu et al., 1995). Also,
the study benefits from the shots activated inside the mine and re-
corded on the surface and the velocity model obtained between
the tunnel and surface lines for migration and time-to-depth conver-
sion. Given the rich and unique nature of the data set, this study fo-
cuses primarily on the surface profiles while an accompanying paper
by Brodic et al. (2021) presents the tunnel seismic data.
GENERAL GEOLOGIC SETTING
The Neves-Corvo VMS deposits, located in southern Portugal, are
considered to be the richest known mineralized bodies of the Iberian
Pyrite Belt (IPB). The IPB, which stretches through southern Spain
into Portugal, itself is arguably the largest and one of the most im-
portant VMS metallogenic provinces in the world (Relvas et al.,
2002). The Neves-Corvo VMS deposits, up to 140 m thick (Oliveira
et al., 2013), typically occur as massive sulfide lenses that host base
metals (copper, zinc, tin, and lead) and stockwork mineralization that
formed at or near the seafloor in submarine intracontinental volcanic
environments. Seven major deposits have been discovered in the area,
namely, Neves, Corvo, Graça, Zambujal, Lombador, Semblana, and
Monte Branco. With the discovery of Semblana in 2010 and Monte
Branco in 2012, the area is still considered highly prospective with
the expectation that more lenses are still to be found in the area.
The VMS deposits can be found in the Late Devonian, at depths
ranging from 200 to 1300 m, within a footwall of autochthonous
volcanosedimentary rocks of mainly shales and felsic volcanic
rocks (360–342 Ma). The dominant minerals are pyrite, chalcopy-
rite, sphalerite, galena, and cassiterite in a lesser amount (West and
Penney, 2017).
The mineralization is overlain by a repetition of volcanic-sedi-
mentary and flysch units (approximately 350 m thick in the Lom-
bador deposit region), which, together with the basement, are folded
into a moderate anticline orientated northwest–southeast, plunging
to the southeast. The deposits have been subject to subvertical and
low-angle thrust faults, resulting in tectonic repetition and thicken-
 Seismic imaging of VMS deposits — Part 2
ing of the massive sulfides (Lundin Mining, 2017, NI 43-101 tech-
nical report).
Downloaded 11/09/21 to 80.217.127.130. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms
The zinc-rich tier-1 Lombador deposit is the main research target
of this work, being the largest of the seven massive sulfide deposits
(approximately 150 Mt) at Neves-Corvo situated on the
northeastern flank of the anticline, on the northern side of the
Neves-Corvo mine lease. (Note that tier-1 refers to the highest qual-
ity deposits regarding high ore concentration and size, long life-of-
mine of more than 10 years, and economic value of billions of dol-
lars.) It was discovered in 1988 following a gravimetric anomaly
(approximately 0.2 mGal), located at approximately 400 m depth
at its western end and extending down to a depth of 1200 m below
the surface along a northwest–southeast strike. The sulfide lens has
dimensions of up to 150 m in thickness and extends for approxi-
mately 1400 m downdip and at least 1600 m along strike (Lundin
Mining, 2017, NI 43-101 technical report). Figure 1a shows a sche-
matic geologic cross section over the Neves and Lombador depos-
its, which were deformed into separate ore lenses by a set of thrust
faults, and it shows how repeated lenses of mineralization occur at
different depths due to at least three sets of thrust faults, which dip
between 20° and 40° toward the northeast. Although the area is
stratigraphically well studied, several thrust faults (e.g., Neves
Thrust) complicate the general structural architecture of the area,
making deep targeting at the site a challenging task.
DOI:10.1190/geo2020-0336.1
SURFACE DATA ACQUISITION
The data acquisition was carried out from the last week of Janu-
ary until the first week of February of 2019; the two surface profiles,
namely, SP6 (2130 m long) and SP7 (1000 m long), were deployed
perpendicular to each other, with the former oriented in a north to
a)
1200
Mining level (m)
1000
800
600
Figure 1. (a) Geologic cross section over the Neves and Lombador deposits. Vertical mining level refers to the real elevation plus 1000 meters.
(b) Base map shows the location of the Lombador and Neves deposits (the projected red surfaces) together with the 2019 surface (cyan) and in-
mine (magenta) 2D seismic lines. The cross-section location is represented by line southwest–northeast (dark blue). (Modified from Donoso
et al., 2020; courtesy of Somincor.)
B183
south direction and the latter from west to east (Figure 1b). These
surface lines were designed to take advantage of an orthogonal
common depth point (CDP) binning grid, and to cover the less-ex-
plored northern area. For example, SP6 is ideal in terms of seismic
illumination (reflection angle), favoring imaging the downdip and
along-strike extension of the Lombador deposit.
A 250 kg vertically AWD was used as the seismic source (Fig-
ure 2a), with the source points located every 10 m. At every source
position, five records were made and later vertically stacked to im-
prove the signal-to-noise ratio (S/N). Both profiles were acquired
using 10 m spaced wireless seismic recorders connected to
10 Hz geophones (Figure 2b). This spacing was chosen given
our earlier experiences from similar VMS settings and the equip-
ment available to the project. Table 1 summarizes other relevant
acquisition parameters.
The AWD seismic source used for the surface lines was chosen as
an alternative to the broadband E-vib seismic source (Noorlandt
et al., 2015) used in the underground tunnels. Given the time
and logistical constraints associated with a survey inside an opera-
tional mine it became necessary to modify the experiment towards
using an accelerated drop-hammer source (brought initially as a
backup source) for the surface profile recording. Because of the
time and access given to the crew, shots were only possible on
the northern half of SP6, with 120 shot points made in total. No
shots were made along SP7; however, those done along SP6 were
also simultaneously recorded by all stations on SP7. As a result, a
higher CDP fold (approximately 120) in the northern part of SP6
was obtained, with a lower CDP fold (approximately 70) at the in-
tersection with SP7.
The survey was conducted with a fixed spread, with 214 stations
using 3C receivers on SP6 and 101 stations (mixed 1C and 3C) on
b)
 B184 Donoso et al.

SP7. All of the wireless recorders were set to operate in an autono- nal to an array of receivers inside different tunnels (Malehmir et al.,
mous mode during the entire survey. A trigger geophone placed 2019a). This system permitted us to synchronize the surface and
Downloaded 11/09/21 to 80.217.127.130. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

directly beside the source was used to mark each record with a
GPS timestamp, and this allowed us to harvest the raw data from
each wireless recording station with enough precision to define the
time range of the desired data. The GPS timestamps of every source
record were used as a base to download the corresponding 5 s record
length for each raw shot gather.
In addition to the surface lines, four underground seismic profiles
(GP2, GP3, GP4, and GP5) were deployed inside the mine drifts
located approximately 650 m below the surface lines (Figure 2c).
This was part of a prototype technology testing that enabled accu-
rate GPS-time (microsecond accuracy) synchronization inside the
denied GPS-time spaces (GPS-denied environment refers to any
environment where the GPS-time signal is blocked by natural or
man-made structures), which allowed us to transmit a GPS-time sig-
a) b) c)
SP6
DOI:10.1190/geo2020-0336.1
tunnel receiver arrays, simultaneously recording all of the shots ac-
tivated in the tunnels into the wireless receivers on the surface; these
data (see also Brodic et al., 2021) were additionally used in this
work to create a velocity model for better depth positioning of
the reflections imaged by the surface data.
DATA PROCESSING
The surface-acquired seismic data show a relatively low S/N due
to strong anthropogenic noise inside the mine, in particular, an ac-
tive raise boring whose strong coherent noise can be already seen in
the raw shot gathers (Figure 3a) together with other sources of noise
from active mining and rock crushing in the mine site, making the
data challenging to process. Standard hardrock seismic processing
sequences were applied. Additionally, a specific
processing sequence was tailored and applied to
In-mine array (2019) target the position and signature of localized
Surface array (2019)
noise, enabling two major sets of reflections to
1 km be imaged (Figure 3b). Table 2 summarizes
the relevant processing steps applied in
this work.
SP7

GP4
GP4
GP4
Prestack signal enhancement
GP2
GP3
SW
GP3
GP3 The prestack processing work started with a
review of the raw data and validation against
GP5 N the field logs to accurately create and assign
geometry files. For the conventional prestack
seismic data processing, only data recorded on
Figure 2. (a) Seismic source used for the surface profile, a 250 kg AWD. (b) Wireless SP6 were considered, given the 2D nature of
recorders connected to 10 Hz geophones spaced every 10 m were used as receivers for the offset distribution along this profile, which
the surface profiles. (c) Diagram of the surface lines and the four underground seismic enabled the application of standard refraction
profiles (GP2, GP3, GP4, and GP5) deployed inside the mine drifts located approxi- static corrections and resulted in improved imag-
mately 650 m below the surface lines.
ing. Thus, a 2D geometry with CDP spacing of
5 m was built and assigned with only the receiver

Table 1. Main acquisition parameters of the 2019 surface reflection seismic data set in the Neves-Corvo VMS deposit, Portugal.

Acquisition system Sercel 428 Lite

Acquisition dates 27 January 2019 to 9 February 2019

Acquisition type Fixed spread
Profiles: SP6 SP7
Profile length (m) 2140 1000
Number of active receivers (10 Hz) 214, 3C 101, mixed 1C-3C
Receiver interval (m) 10 10
Source interval 10 m (five impacts per point) only along profile SP6
Max. source-receiver offset (m) 2130 1255
Source 250 kg accelerated vertical drop hammer
Recording parameters:
Record length (s) 5 5
Sampling rate (ms) 2 2
Geodetic surveying:
Equipment Total station and differential global positioning system
 Seismic imaging of VMS deposits — Part 2 B185

stations located on SP6. Although the acquired raw data are 3C, ure 3a); all of the prestack tools and algorithms applied had to take
only the vertical component data are used in this study, assuming into account this source of coherent noise. Following this, bad traces
Downloaded 11/09/21 to 80.217.127.130. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

here that the source is fully vertical and not taking into consider- were selected and removed, and first arrivals were picked manually
ation wave-mode conversions (Bellefleur et al., 2004, 2012). where possible. For the traces directly on top of the raise-boring
A first view of the shot gathers shows relatively high background noise cone, first-break picking had to be done simultaneously with
noise levels, as expected from an active mining site, together with frequency attenuation through a set of band-pass filters, spectral
strong coherent noise source produced by active raise boring (Fig- balancing, and top-mute application where necessary. The picked
first arrivals were then used to build a velocity
a) b) model for refraction statics to correct for a shal-
0
N S N S low low-velocity layer. Afterward, the direct
wave from the source was removed by deleting
100 a few milliseconds below the first-break picks.
Upon executing these processing steps, two no-
200 ticeable reflections were recognized in the shot
records, one at approximately 200–300 ms and
Time (ms)

300 R1 a stronger one at approximately 600 ms (Fig-

ure 3b). After conventional processing methods
400 were applied, a targeted approach at removing
the raise-bring noise cone was done in the off-
500 Lombador set-gather domain via offset manipulation in
the header values: iterative median-filters were
600
designed and applied to target the specific drill-
Raw shot Processed shot ing-cone and to deal with coherent sources of
−500 0 500 1000 1500 −500 0 500 1000 1500
Offset (m) Offset (m)
noise from far offsets. For this, the noise cone
of the coherent source (given its fixed position)
DOI:10.1190/geo2020-0336.1

Figure 3. (a) Example of a raw shot gather strongly contaminated with noise from a
raise boring machine operating during the survey within one of the exploration tunnels,
approximately 650 m below the surface profile. (b) Processed shot gather showing that
most of the coherent noise has been removed, allowing us to resolve a shallow reflection
(R1) and a deeper one likely from the Lombador deposit.
was identified and redefined as a zero-offset lo-
cation; then, for all of the other receivers, a new
offset was calculated. The median filter then took
advantage of the new offset headers for attenuat-
ing the noise from the raise boring that was tak-

Table 2. Principal 2D processing steps applied to the 2019 surface reflection seismic data set (SP6) in the Neves-Corvo VMS
deposit, Portugal.

Process Parameters

Data input SEGD, converted to SEGY for processing

Geometry assignment From trace headers and field logs
CDP binning Straight line: CDP bin size of 5 m
Trace editing Kill noisy traces
First-break picking Automatic, manual refinement
Refraction static corrections Two-layer model based on first breaks
Top mute Using first breaks
Air-wave attenuation Velocity of 330 m/s
Low-pass filter 130–150 Hz
Velocity analysis Iterative
Header math Offset manipulation
Median filter 900–1200 m/s
NMO corrections 40% stretch mute
Surface-consistent residual statics Twice and iterative
Band-pass filter 10–30–130–180 Hz
AGC Window: 200 ms
Stack Unity
Migration Poststack finite-differences algorithm: velocity model from tomographic inversion
Time-to-depth conversion 1D constant velocity
 B186 Donoso et al.

ing place at the same level as the underground tunnels at the 650 m that would maximize the reflection coherency on the final stacked
depth (specifically on the GP4 tunnel) at the same time as the sections. The appropriate stacking velocity was found to be approx-
Downloaded 11/09/21 to 80.217.127.130. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

survey. imately 5200 m/s, and this was used as a starting velocity model.
Residual statics were then calculated using the NMO-corrected
Poststack processing gathers, further improving the continuity of the reflections on
the stacked sections. Additional processes applied on the poststack
After successfully enhancing the reflections in the shot gathers, data for coherency enhancement are FX-deconvolution, band-pass
constant-velocity stacks (CVS) were created to inspect the imaging filtering, automatic gain control (AGC), and high-pass filtering for
potential and to help define an appropriate NMO velocity function low-frequency noise removal.
An iterative process of velocity analysis and surface-consistent
N S residual statics was used to further enhance the two observed major
0
reflections. The final unmigrated stacked profile along SP6 is pre-
sented in Figure 4. Two main reflections are observed, one at ap-
100 proximately 300 ms attributed to the base of the Upper Flysch
lithologic group and a deeper one that begins at approximately
200 600 ms and extends up-dip, likely originated by the Lombador de-
posit. The vertical resolution was estimated using Rayleigh’s limit
Time (ms)

300 theorem (one quarter of the wavelength) assuming an average veloc-

ity of 5000–5500 m/s and a dominant frequency of 45–55 Hz, to be
25–30 m. This means that although the Lombador deposit is at pla-
400
ces thicker than 25–30 m, it has only been detected rather than ver-
tically resolved. This is consistent with the fact that only one single
500 reflection is observed, as was also seen in the reprocessed legacy
data shown in Donoso et al. (2020).
600 Although seismic data were recorded and processed along line
DOI:10.1190/geo2020-0336.1

SP7, due to the very low CDP fold, it was not possible to obtain
a suitable seismic section. This is because all of the shots were made
0 500 1000 1500
along SP6 and none were made along SP7, making it unreliable to
Distance along profile (m)
calculate refraction statics on SP7 because no near-offset data are
Figure 4. Unmigrated stacked section of SP6. The two strong re- available. A practical solution that we found for this was to incor-
flections marked by arrows are interpreted to originate from the porate SP7 data in the cross-dip analysis where the midpoint cover-
base of the Upper Flysch group (allochthonous) and the Lombador age from the cross-profile recording helped to obtain some
deposit. information about the out-of-plane nature of the reflections.

a) b) c) Tunnel-to-surface profile and P-wave

first-arrival tomography
The E-vib shots activated in the underground
tunnels, primarily on GP4 (Figure 2c), and re-
corded by the surface receiver stations on SP6
provided a way of measuring the direct compres-
sional wave (P-wave) traveltimes from the sub-
surface (tunnels) to the receivers on the
surface profiles. With the objective of placing
the observed reflections at their true positions
at depth, an appropriate time-to-depth velocity
model was needed. To obtain this, an initial
approximation was done by picking the first
arrivals on the surface data corresponding to
the shots in the tunnels (the green squares in Fig-
ure 5a), similar to the way in which a walkaway
vertical seismic profile (VSP) survey would be
executed. In this work, the particular setup of
the tunnel-to-surface profiling method (TSP)
was introduced. From the TSP, a 1D velocity
model was built assuming straight raypaths be-
Figure 5. (a) Example shot gather generated inside one of the exploration tunnels at tween the source and receiver positions (Fig-
650 m depth, recorded on SP6 with direct arrivals picked (the green squares).
(b) The 1D velocity model estimated using the direct arrivals from 20 shots located ure 5b), and it was later used to analyze the
20 m apart. (c) Surface elevation profile for SP6. The local low-velocity zone in the velocity structures between the tunnel and sur-
southern part of SP6 correlates to a drop in the elevation (the blue rectangle). face profiles. Interestingly, toward the southern
 Seismic imaging of VMS deposits — Part 2 B187

portion of SP6, there is a reduction in the direct P-wave velocity; MODELING AND ANALYSIS
although it is an unexpected result, it does match the known geology
3D exploding reflector modeling
Downloaded 11/09/21 to 80.217.127.130. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

of the study area.

To maximize the value of the tunnel-to-surface data, P-wave first- Despite the noisy nature of the data, it was still possible to obtain
break traveltime tomography was done using a diving-wave tomog- clear reflections in the unmigrated and migrated sections. A time-to-
raphy algorithm (Tryggvason et al., 2002). The velocities obtained
from analysis of direct and refracted arrivals on the shot gather of
surface profiles were used to build the initial 2D velocity model N SP7 S
300
centered along the SP6 profile. This velocity model was then used
as the input for ray tracing with the inversion done in an iterative 200
manner using a least-squares conjugate gradient solver (Paige and 100
Saunders, 1982). Cell sizes of 5 m were set in the inline, crossline,

Elevation (m)
0
and depth directions. For the inversion purpose, first breaks of the
E-vib source points activated in the underground tunnels (Brodic –100
et al., 2021) and the surface activated AWD sources and recorded –200
on all receivers (tunnel and surface profiles) were used. The total
–300
number of first breaks used was 81,402 picked on 770 receivers and GP4
GP4
220 source locations. For every iteration, smoothness constrains –400 1000 2000 3000 4000 5000 6000

were applied and gradually relaxed to prevent extreme velocity var- –500
P-wave (m/s)
iations, and, after seven iterations, a root-mean square error of 0 500 1000 1500 2000
2.1 ms was obtained, which is nearly similar to the sampling rate Distance along profile (m)
of the wireless recorders and our estimate of the error of the picked
first breaks. The obtained velocity model (Figure 6) was cross va- Figure 6. The 2D projection along profile SP6 of the 3D velocity
lidated against a priori geologic and structural information and then model obtained from the first-break traveltime tomography. Inter-
section with profile SP7 in the middle (the upper black arrow) and
taken as representative of the velocity distribution of the structures
DOI:10.1190/geo2020-0336.1

located between the tunnels and the surface profiles. A 2D slice
along the surface profile SP6 was extracted from the final 3D tomo-
graphically obtained P-wave velocity model and
then smoothed using a median filter to be used
for poststack migration and depth conversion. a)
position of the in-mine tunnel array GP4 (the lower black arrow) are
also shown. The velocity model, after smoothing, was used for mi-
gration and time-to-depth conversion.
SP7
SP6

Migration
For the surface data, first prestack time migra-
tion algorithms, full and partial (i.e., Kirchhoff
and dip moveout), were tested, but the results
showed a decrease in reflection continuity and
the creation of artifacts rather than an improve-
ment, most likely due to the limited source cover-
age because it was only possible to shoot along
half of the SP6 extension, effectively making the
offset and fold values irregular in particular to-
ward the southern part of the survey. Hence,
no prestack migration was included in the final b) SP7 SP6
processing sequence and the decision was made
to use poststack migration. Various algorithms
were tested, of which finite-differences poststack
time migration was selected because it presented
more continuous reflections and less migration
artifacts, and it was applied together with the
smoothed velocity model from the traveltime
tomography inversion, thus providing, according
to the deposit model (derived from boreholes), a
better positioning for the Upper Flysch group con-
tact and the Lombador deposit reflections (Fig-
ure 7). We will later discuss that the downdip
continuation of the Lombador, as seen in Figure 7,
cannot reliably be interpreted given the possible Figure 7. A 3D view of the migrated stacked section of SP6. The Lombador deposit
out-of-plane effect and cross-dip component in model is shown as a red surface. The green lines represent the mine tunnels. The total
the resulting section. depth of the section shown is 1500 m, with no vertical exaggeration.
 B188 Donoso et al.
depth conversion was applied to the unmigrated section using a con-
stant velocity to begin seismic interpretation relative to the known
Downloaded 11/09/21 to 80.217.127.130. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms
geology. Note that unmigrated sections should always be looked at
with care because the dip and depth extent are not true for dipping
reflections. In the unmigrated section shown in Figure 8a, it appears
that the shallowest reflection coincides with the base of the Upper
Flysch group (the light blue arrow, Figure 8a) and the deeper set of
the reflections is associated with the Lombador deposit (the red
arrow, Figure 8a), for which one cannot reliably discuss the depth
extension from an unmigrated section.
As a way of validating if the reflection imaged in the unmigrated
section corresponds to the Lombador known position, 3D exploding
reflector modeling (i.e., to assume the sources are located along the
a)
DOI:10.1190/geo2020-0336.1
b)
c)
Figure 8. (a) A 3D view of the unmigrated 2D stacked section of
SP6. (b) A 3D exploding reflector modeling (the yellow rays).
(c) Projected response (the yellow dots) on the 2D section. The
Lombador massive sulfide lens is shown as a red surface. The green
lines represent mine tunnels and shafts. Total depth of the sections
shown is 1800 m, with no vertical exaggeration.
surface of interest) (White et al., 2012; Ahmadi et al., 2013) is done
for the known Lombador surface model (assuming that the whole
surface is still an intact mineralization) and the receivers on the sur-
face (see the yellow rays in Figure 8b) and then projected into the
2D unmigrated stacked section (the yellow dots, Figure 8c), show-
ing a relatively good match between the reflection observed on the
unmigrated stacked section and the 2D projection (the yellow dots,
Figure 8c) of the Lombador deposit’s known position at depth. The
exploding reflector modeling methodology uses depth information
only, and it is essentially a constant-velocity straight-ray approach.
The Lombador deposit surface model is highly reliable because it
has been thoroughly constrained by numerous surface and under-
ground boreholes.
Cross-dip analysis
In regions where complex geologic structures are present, cross-
dip analysis is a useful tool that brings valuable structural informa-
tion that, together with the image obtained in an inline stacked sec-
tion, makes it possible to estimate the true dip and strike direction of
the geologic features imaged in a stacked section. The cross-dip
time delay Δt is defined as the component of the reflection dip
in the vertical plane perpendicular to the seismic profile (Larner
et al., 1979; Wu et al., 1995; O’Dowd et al., 2004), and it can
be calculated from Δt = (2ΔY/v)sinφ, given a cross-dip angle φ
and medium velocity v (assumed to be constant in most cases);
the reflection times in the seismic traces within a CDP gather will
vary according to the distance ΔY between the midpoint and the
effective stacking line.
For the Neves-Corvo 2019 surface survey processing, some out-
of-plane dip contribution into the SP6 profile from the Lombador
deposit is expected. Cross-dip analysis requires spread midpoints,
and, with only one straight CDP line, such as line SP6 data, this
calculation is not possible. Thus, for this analysis, the data from
the perpendicular line SP7 was also included to build a pseudo-
3D midpoint cloud and determine the cross-dip component of
the reflections identified in the unmigrated stacked section.
Because of the cross-dip angle dependence on the stacking veloc-
ity, the first step toward the analysis was to generate a series of CVS
panels to find the most adequate stacking velocity (dip dependent in
this case). A stacking velocity range of 4900–5300 m/s was tested,
and a 5200 m/s constant NMO velocity was found to provide the
best results.
Subsequently, cross-dip analysis was applied to constant velocity
NMO-corrected (5200 m/s) CDP gathers, which include the mid-
point contributions from SP7; these were plotted as CVS with dif-
ferent cross-dip correction angles (Figure 9). Positive angles
indicate a northwestern cross-dip component, and negative angles
a northeast one; by visually comparing these sections, finding the
most coherent and strongest reflections, the optimal cross-dip angle
is determined.
As indicated by the results shown in Figure 9, the reflection
attributed to the Lombador deposit stacks more coherently after
a cross-dip correction angle of approximately −30°, implying that
there is a lateral contribution from reflections on the western side
of SP6. This may explain why the Lombador reflection appears
a bit longer at depth than the modeled deposit because an out-
of-plane feature also implies later arrivals, hence a false downdip
continuation.
 Seismic imaging of VMS deposits — Part 2 B189

DISCUSSION there may be a further downdip (vertical extension) continuation of

the deposit, beyond what is known from borehole data. Given the
The processed seismic data were analyzed to validate their reli-
Downloaded 11/09/21 to 80.217.127.130. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

out-of-plane nature of the reflection, this depth extension may only

ability and also to support their interpretations, including the use of
TSP survey data, 3D exploding reflector modeling, and cross-dip
analysis. The TSP data and first-break traveltime tomography
velocity model, which were obtained uniquely in this study, were
used for a more accurate time-to-depth conversion, as well as to
indicate a reduction in the direct P-wave velocity toward the
southern portion of SP6 (Figures 5b and 6). The location of the
low-velocity zone matches a lower topography and is consistent
with what would be expected from the presence of a large fault
or a system of closely spaced faults, such as the Neves main thrust
(Figure 1a) that verges southwest and becomes shallower toward the
south of line SP6 overlaying GP4 where the shots were made, which
means that the direct waves slowed down when passing through one
of the thrust faults. This observation strongly indicates that the TSP
setup is ideal, where possible, for proper depth imaging applica-
tions, and also for providing by-product results and information that
could be important for improving the understanding of the geology
at depth in and around a mine site. TSP data could also be treated
like VSP data for reflection identifications, removing the upcoming
wavefield and shot migration and also for carrying out proper 2D
tomographic imaging between the tunnels and surface as illustrated
in this study.
DOI:10.1190/geo2020-0336.1
be due to the 2D nature of the data and the fact that a reflection from
the deposit is delayed in time, hence providing a false implication
in depth.
The seismic data acquisition carried out in this work was done
with a conventional impact source and basic surface receiver setup,
and it was already possible to obtain suitable images of the miner-
alization and contact lithologies. These are encouraging results and
illustrate the capability of seismic methods for direct deposit target-
ing in this region and possibly similar regions elsewhere. However,
geologic complexities and variable background velocity highlight
the necessity for good control on the velocity model to avoid mis-
leading interpretations and false positioning at depth. The tunnel-to-
surface seismic data provided a comparable scale seismic velocity
model further illustrating why such as a setup is ideal in under-
ground mining sites such as the Neves-Corvo, opening possibilities
for improved and more cost-effective imaging solutions.
CONCLUSION
Two surface seismic reflection profiles were acquired over the
world-class Lombador deposit in the Neves-Corvo mine, southern

The relevance of lateral contributions into the unmigrated stacked
2D profile became clear after the cross-dip analysis results, which
are consistent with what was already shown by the 3D exploding
reflector modeling. The reflection from the Lombador deposit in the
unmigrated stacked section, at approximately 500 ms (Figure 8),
has a strong out-of-plane influence. It is interesting to emphasize
that, when visualizing the migrated stacked seismic section together
with the deposit model (Figure 7), the seismic section suggests that
Portugal. The surface seismic data acquisition was partly synchron-
ized with an in-mine seismic experiment, taking advantage of a
GPS-time synchronization system developed for such an applica-
tion. The obtained unmigrated stacked section of the surface profile
supports the assertion that it is possible to detect near-surface re-
flections, which can be associated with mineral deposits of eco-
nomic interest, in an active mining environment with a low S/N.
In this regard, these observations encourage the use of new technol-

0
–45° –30° –15° 0° 15° 30° 45°
100

200
Time (ms)

300

400

500

600

450 450
Time (ms)
Time (ms)

500 500

550 550

600 600
0 500 1000 1500 0 500 1000 1500
Distance along profile (m) Distance along profile (m)

Figure 9. Constant-velocity stacked panels (5200 m/s) corrected for a range of cross-dip angles along SP6 with lateral midpoint contributions
from SP7. A cross-dip angle at approximately −30° shows the most continuous reflection for the Lombador. A close-up of the Lombador
reflection is shown for cross-dip angle −30° and for the unmigrated stacked section equivalent to a 0° cross-dip angle (uncorrected).
 B190 Donoso et al.
ogies for mineral resource exploration that may result in increased
efficiency, optimizing exploration and mine planning procedures.
Downloaded 11/09/21 to 80.217.127.130. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms
In the final stacked section of SP6, the shallow reflection corre-
sponds to the known Upper Flysch group (allochthonous) contact
according to the known geology from the area, and the second
strong reflection matches the known Lombador massive sulfide de-
posit. The latter interpretation was further validated using 3D ex-
ploding reflector forward modeling.
The velocity model from the tunnel-to-surface survey was critical
to obtain improved migration and time-to-depth conversion results.
The TSP survey has the potential to provide information about the
localized geologic structure, and it was done as a proof-of-concept
where such a possibility exists. Albeit this type of study was only
possible after the development of the GSP-signal technology tested
in this survey and presented in the accompanying article from the
tunnel experiment.
The cross-dip analysis showed a strong lateral contribution from
approximately 30° west of SP6, generated from the Lombador de-
posit. The influence of out-of-plane reflections is not readily seen
on a 2D seismic section and can lead to false positives or inaccurate
target positioning at depth.
ACKNOWLEDGMENTS
We thank Somincor-Portugal (a subsidiary of the Lundin Mining
DOI:10.1190/geo2020-0336.1
Corporation) for providing access to the mining site and the logis-
tical support during the acquisition and for collaborating with us
through the Smart Exploration project. Smart Exploration has re-
ceived funding from the European Union’s Horizon 2020 research
and innovation program under grant agreement no. 775971. G. A.
Donoso’s Ph.D. work is supported by the project. We also thank
LNEG for collaborating with us in this study and their contribution
to the discussions and interpretation of the results. We appreciate the
critical reviews and comments by the anonymous reviewers and the
associate editor G. Tsoflias.
DATA AND MATERIALS AVAILABILITY
Data associated with this research are available and can be ob-
tained by contacting the corresponding author.
REFERENCES
Abdi, A., S. Heinonen, C. Juhlin, and T. Karinen, 2015, Constraints on the
geometry of the Suasselkä post-glacial fault, northern Finland, based on
reflection seismic imaging: Tectonophysics, 649, 130–138, doi: 10.1016/j
.tecto.2015.03.004.
Adam, E., G. Arnold, C. Beaudry, L. Matthews, L. B. Milkereit, G. Perron,
and R. Pineault, 1997, Seismic exploration for VMS deposits, Matagami,
Québec, in A. G. Gubins, Proceedings of Exploration 97: Fourth Decen-
nial International Conference on Mineral Exploration, 433–438.
Ahmadi, O., C. Juhlin, A. Malehmir, and M. Munck, 2013, High-resolution
2D seismic imaging and forward modeling of a polymetallic sulfide de-
posit at Garpenberg, central Sweden: Geophysics, 78, no. 6, B339–B350,
doi: 10.1190/geo2013-0098.1.
Balestrini, F., D. Draganov, A. Malehmir, P. Marsden, and R. Ghose, 2020,
Improved target illumination at Ludvika mines of Sweden through seis-
mic-interferometric noise reduction: Geophysical Prospecting, 68, 200–
213, doi: 10.1111/1365-2478.12890.
Bellefleur, G., S. Cheraghi, and A. Malehmir, 2018, Reprocessing legacy
three-dimensional seismic data from the Halfmile Lake and Brunswick
No. 6 volcanogenic massive sulphide deposits, New Brunswick, Canada:
Canadian Journal of Earth Sciences, 56, 569–583, doi: 10.1139/cjes-
2018-0103.
Bellefleur, G., A. Malehmir, and C. Müller, 2012, Elastic finite-difference
modeling of volcanic-hosted massive sulfide deposits: A case study from
Half Mile Lake, New Brunswick, Canada: Geophysics, 77, no. 5, WC25–
WC36, doi: 10.1190/geo2011-0445.1.
Bellefleur, G., C. Müller, D. Snyder, and L. Matthews, 2004, Downhole seis-
mic imaging of a massive sulfide orebody with mode-converted waves,
Halfmile Lake, New Brunswick, Canada: Geophysics, 69, 318–329, doi:
10.1190/1.1707051.
Bräunig, L., S. Buske, A. Malehmir, E. Bäckström, M. Schön, and P. Mars-
den, 2020, Seismic depth imaging of iron-oxide deposits and their host
rocks in the Ludvika mining area of central Sweden: Geophysical Pro-
specting, 68, 24–43, doi: 10.1111/1365-2478.12836.
Brodic, B., R. de Kunder, P. Ras, J. Van den Berg, and A. Malehmir, 2019,
Seismic imaging using electromagnetic vibrators-Storm versus Lightning:
25th European Meeting of Environmental and Engineering Geophysics.
Brodic, B., A. Malehmir, N. Pacheco, C. Juhlin, L. Dynesius, J. Carvalho, J.
van den Berg, R. de Kunder, G. Donoso, T. Sjölund, and V. Araujo, 2021,
Innovative seismic imaging of VMS deposits, Neves-Corvo, Portugal —
Part 1: In-mine Array: Geophysics, this issue, doi: 10.1190/geo2020-0565.1.
Dehghannejad, M., T. E. Bauer, A. Malehmir, C. Juhlin, and P. Weihed,
2012, Crustal geometry of the central Skellefte district, northern Swe-
den–Constraints from reflection seismic investigations: Tectonophysics,
524–525, 87–99, doi: 10.1016/j.tecto.2011.12.021.
Dehghannejad, M., C. Juhlin, A. Malehmir, P. Skyttä, and P. Weihed, 2010,
Reflection seismic imaging of the upper crust in the Kristineberg mining
area, northern Sweden: Journal of Applied Geophysics, 71, 125–136, doi:
10.1016/j.jappgeo.2010.06.002.
Dentith, M. C., and S. T. Mudge, 2014, Geophysics for the mineral explo-
ration geoscientist: Cambridge University Press.
Donoso, G. A., A. Malehmir, N. Pacheco, V. Araujo, M. Penney, J. Car-
valho, B. Spicer, and S. Beach, 2020, Potential of legacy 2D seismic data
for deep targeting and structural imaging at the Neves–Corvo massive
sulphide-bearing deposit, Portugal: Geophysical Prospecting, 68, 44–
61, doi: 10.1111/1365-2478.12861.
Juhlin, C., E. Sturkell, R. Ove, J. Ebbestad, O. Lehnert, A. E. S. Högström,
and G. Meinhold, 2012, A new interpretation of the sedimentary cover in
the western Siljan Ring area, central Sweden, based on seismic data: Tec-
tonophysics, 580, 88–99, doi: 10.1016/j.tecto.2012.08.040.
Koivisto, E., A. Malehmir, N. Hellqvist, T. Voipio, and C. Wijns, 2015,
Building a 3D model of lithological contacts and near-mine structures
in the Kevitsa mining and exploration site, northern Finland: Constraints
from 2D and 3D reflection seismic data: Geophysical Prospecting, 63,
754–773, doi: 10.1111/1365-2478.12252.
Larner, K. L., B. R. Gibson, R. Chambers, and R. A. Wiggins, 1979, Simul-
taneous estimation of residual static and crossdip corrections: Geophysics,
44, 1175–1192, doi: 10.1190/1.1441001.
Malehmir, A., M. Andersson, M. Lebedev, M. Urosevic, and V. Mikhaltse-
vitch, 2013, Experimental estimation of velocities and anisotropy of a
series of Swedish crystalline rocks and ores: Geophysical Prospecting,
61, 153–167, doi: 10.1111/j.1365-2478.2012.01063.x.
Malehmir, A., and G. Bellefleur, 2009, 3D seismic reflection imaging of
volcanic-hosted massive sulfide deposits: Insights from reprocessing
Halfmile Lake data, New Brunswick, Canada: Geophysics, 74, no. 6,
B209–B219, doi: 10.1190/1.3230495.
Malehmir, A., R. Durrheim, G. Bellefleur, M. Urosevic, C. Juhlin, D. White,
B. Milkereit, and G. Campbell, 2012, Seismic methods in mineral explo-
ration and mine planning: A general overview of past and present case
histories and a look into the future: Geophysics, 77, no. 5, WC173–
WC190, doi: 10.1190/geo2012-0028.1.
Malehmir, A., L. Dynesius, and T. Sjölund, 2019a, Time-synchronized geo-
physical investigations in denied GPS-time spaces: 1st Conference on
Geophysics for Infrastructure Planning Monitoring and BIM, 1–5.
Malehmir, A., P. Holmes, P. Gisselø, L. V. Socco, J. Carvalho, P. Marsden,
A. Verboon, and M. Loska, 2019b, Smart exploration: Innovative ways of
exploring for the raw materials in the EU: 81st Annual International
Conference and Exhibition, EAGE, Extended Abstracts, 1–5, doi: 10
.3997/2214-4609.201901668.
Manzi, M., G. Cooper, A. Malehmir, R. Durrheim, and Z. Nkosi, 2015, Inte-
grated interpretation of 3D seismic data to enhance the detection of the gold-
bearing reef: Mponeng Gold mine, Witwatersrand Basin (South Africa):
Geophysical Prospecting, 63, 881–902, doi: 10.1111/1365-2478.12273.
Markovic, M., G. Maries, A. Malehmir, J. von Ketelhodt, E. Bäckström, M.
Schön, and P. Marsden, 2020, Deep reflection seismic imaging of iron‐
oxide deposits in the Ludvika mining area of central Sweden: Geophysical
Prospecting, 68, 7–23, doi: 10.1111/1365-2478.12855.
Marques, F., J. X. Matos, P. Sousa, B. Spicer, M. Penney, P. Represas, V.
Araújo, J. Carvalho, I. Morais, N. Pacheco, L. Albardeiro, and P. Gon-
çalves, 2019, The role of legacy land gravity data in the Neves-Corvo
mine discovery and its use in present-day exploration and new target gen-
eration: First Break, 37, 97–102, doi: 10.3997/1365-2397.n0051.
Milkereit, B., D. W. Eaton, J. Wu, G. Perron, M. H. Salisbury, E. Berrer, and
G. Morrison, 1996, Seismic imaging of massive sulphide deposits — Part
2: Reflection seismic profiling: Economic Geology, 91, 829–834, doi: 10
.2113/gsecongeo.91.5.829.
 Seismic imaging of VMS deposits — Part 2
Noorlandt, R., G. Drijkoningen, J. Dams, and R. Jenneskens, 2015, A seis-
mic vertical vibrator driven by linear synchronous motors: Geophysics,
80, no. 2, EN57–EN67, doi: 10.1190/geo2014-0295.1.
Downloaded 11/09/21 to 80.217.127.130. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms
O’Dowd, C. R., D. Eaton, D. Forsyth, and H. W. Asmis, 2004, Structural
fabric of the Central Metasedimentary Belt of southern Ontario, Canada,
from deep seismic profiling: Tectonophysics, 388, 145–159, doi: 10.1016/
j.tecto.2004.07.041.
Oliveira, J., C. Rosa, Z. Pereira, D. Rosa, J. Matos, C. Inverno, and T. An-
dersen, 2013, Geology of the Rosário-Neves Corvo antiform, Iberian
Pyrite Belt, Portugal: New insights from physical volcanology, palynos-
tratigraphy and isotope geochronology studies: Mineralium Deposita, 48,
749–766, doi: 10.1007/s00126-012-0453-0.
Paige, C. C., and M. A. Saunders, 1982, LSQR: An algorithm for sparse
linear equations and sparse least squares: ACM Transactions on Math-
ematics Software, 8, 43–71, doi: 10.1145/355984.355989.
Papadopoulou, M., F. Da Col, B. Mi, E. Bäckström, P. Marsden, B. Brodic,
A. Malehmir, and L. V. Socco, 2020, Surface‐wave analysis for static cor-
rections in mineral exploration: A case study from central Sweden: Geo-
physical Prospecting, 68, 214–231, doi: 10.1111/1365-2478.12895.
Place, J., and A. Malehmir, 2016, Using supervirtual first arrivals in con-
trolled-source hardrock seismic imaging — Well worth the effort: Geo-
physical Journal International, 206, 716–730, doi: 10.1093/gji/ggw176.
Relvas, J., F. Barriga, A. Pinto, A. Ferreira, N. Pacheco, P. Noiva, G. Barriga,
R. Baptista, D. Carvalho, V. Oliveira, J. Munhá, and R. Hutchinson, 2002,
The Neves-Corvo deposit, Iberian Pyrite Belt, Portugal: Impacts and fu-
ture, 25 years after the discovery: SEG Special Publication, 9, 155–76.
Salisbury, M. H., B. Milkereit, G. Ascough, R. Adair, L. Matthews, D. R.
Schmitt, J. Mwenifumbo, D. W. Eaton, and J. Wu, 2000, Physical proper-
ties and seismic imaging of massive sulphides: Geophysics, 65, 1882–
1889, doi: 10.1190/1.1444872.
Schodde, R., 2020, The challenges and opportunities for geophysics for
making discoveries under cover: Annual Convention, Prospectors &
Developers Association of Canada.
Tryggvason, A., A. Malehmir, J. Rodriguez-Tablante, C. Juhlin, and P.
DOI:10.1190/geo2020-0336.1
Weihed, 2006, Reflection seismic investigation in the western part of
B191
the Paleoproterozoic VHMS-bearing Skellefte district, northern Sweden:
Economic Geology, 101, 1039–1054, doi: 10.2113/gsecongeo.101.5
.1039.
Tryggvason, A., S. T. Rögnvaldsson, and Ó. G. Flovenz, 2002, Three dimen-
sional imaging of P- and S-wave velocity structure and earthquake loca-
tions beneath southwest Iceland: Geophysical Journal International, 151,
848–866, doi: 10.1046/j.1365-246X.2002.01812.x.
Urosevic, M., G. Bhat, and M. Grochau, 2012, Targeting nickel sulfide de-
posits from 3D seismic reflection data at Kambalda, Australia: Geophys-
ics, 77, no. 5, WC123–WC132, doi: 10.1190/geo2011-0514.1.
Verpaelst, P., A. S. Peloquin, E. Adam, A. E. Barnes, J. Ludden, D. J. Dion,
C. Hubert, B. Milkereit, and M. Labrie, 1995, Seismic reflection profiles
across the “Mine Series” in the Noranda camp of the Abitibi belt, eastern
Canada: Canadian Journal of Earth Sciences, 32, 167–176, doi: 10.1139/
e95-014.
West, D., and M. Penney, 2017, Brownfields and beyond — Undercover at
Neves Corvo, Portugal: Sixth Decennial International Conference on
Mineral Exploration, 291–304.
White, D. J., D. Secord, and M. Malinowski, 2012, 3D seismic imaging of
volcanogenic massive sulfide deposits in the Flin Flon Mining Camp,
Canada — Part 1: Seismic Results: Geophysics, 77, no. 5, WC47–
WC58, doi: 10.1190/geo2011-0487.1.
Wu, J., B. Milkereit, and D. E. Boerner, 1995, Seismic imaging of the enig-
matic Sudbury structure: Journal of Geophysical Research, Solid Earth,
100, 4117–4130, doi: 10.1029/94JB02647.
Yavuz, S., J. Kinkela, A. Dzunic, M. Penney, R. Neto, V. Araújo, S. Ziramov,
R. Pevzner, and M. Urosevic, 2015, Physical property analysis
and preserved relative amplitude processed seismic imaging of
volcanogenic massive sulfides — A case study from Neves–Corvo, Por-
tugal: Geophysical Prospecting, 63, 798–812, doi: 10.1111/1365-2478
.12269.
Biographies and photographs of the authors are not available.