Pure Appl. Geophys.

169 (2012), 2031–2046

Ó 2012 Springer Basel AG
DOI 10.1007/s00024-012-0456-x Pure and Applied Geophysics

Understanding the 3D Structure of a Thermal Water Fissured Granite Aquifer

by Use of Geophysical Studies
E. C. RAMALHO,1 J. P. CARVALHO,1 R. GONÇALVES,2,3 and F. A. MONTEIRO SANTOS3

Abstract—The thermal hot springs of Caldas de Monção, detect favourable areas for thermal water borehole drilling to
located in Northern Portugal, are supplied by a thermal water replace the hot springs with the objective of protecting the aquifer
aquifer that must be preserved from pollution and anthropic effects. from pollution.
Enhancement of knowledge about the aquifer is therefore needed,
so sustainable management and water quality preservation can be Key words: Geophysics, hydrogeology, thermal water,
conducted, namely by replacing the springs by boreholes, which granites.
are less vulnerable to pollution. Several investigations have been
conducted with this purpose, to review the available geological and
hydrogeological information from the area. For this reason, several 1. Introduction
geophysical surveys were reprocessed and others were acquired
and planned, so that a 3D conceptual groundwater flow model of The thermal hot waters of Caldas de Monção have
could be drawn. Three seismic reflection profiles were acquired
with the objective of locating deep faults capable of transporting
been used for several centuries because of their reco-
groundwater to the surface and to complement information about gnized therapeutic virtues, and there is information
shallow features available before the seismic survey. This infor- about their use since Roman times (HENRIQUES, 1726).
mation included interpretation of a dipole–dipole survey conducted
They are hyperthermal, with temperature of approxi-
in 2001, that reached an investigation depth of approximately
40 m, and two shallow boreholes drilled, after a deeper geoelec- mately 50°C; their chemical characteristics are
trical survey performed in the 1980s, to replace the springs. Besides hypossaline, sulfureous sodium-bicarbonate, pH 7.8,
the seismic reflection profiles, seven vertical electrical soundings 560 mg/l total dissolved solids, and electric conduc-
were also acquired and refraction interpretation of the first reflec-
tion data to arrive was also undertaken to complement the tivity of 628 lS/cm (15.9 X m) (CRUZ, 1992).
information. This paper reports integrated interpretation of the Thermal water aquifer management is the best
geophysical and geological data to obtain a 3D conceptual model to way to conduct an exploitation plan that simulta-
achieve the above mentioned objectives. After interpretation and
lateral correlation of the structures found in the geophysical pro-
neously preserves the hydromineral resource for the
files, the deep 3D conceptual model integrating the geological, future and will protect the aquifer from anthropic
hydrogeological, and geophysical information was constructed. effects. One way to protect the aquifer is replace-
This model was the basis for planning the location of an explora-
ment of the springs by boreholes. This replacement
tory borehole that was drilled afterwards. From two initial locations
appearing as the strongest candidates for a successful thermal water will prevent pollution in the aquifer but needs a
production well, one was chosen to drill a third borehole to supply good knowledge of its characteristics. However, the
the spa facilities. The third borehole results were very positive— proximity of the effect of the Minho River may
250 m deep, temperature approximately 50°C, and water flow
12.5 l/s. This example shows the utility of using geophysical jeopardize the integrity of the aquifer, with regard
methods to increase knowledge about groundwater circulation and to overexploitation and contamination. Knowledge
of the hydromineral circuit of Caldas de Monção is
inferred from previous investigations conducted by
LIMA (2001) and CARREIRA et al. (2004, 2005,
1
Laboratório Nacional de Energia e Geologia, LNEG, 2006), and use of silica geothermometry (AIRES-
Estrada da Portela, Bairro do Zambujal, Apartado 7586, 2720-866
Amadora, Portugal. E-mail: elsa.ramalho@lneg.pt
BARROS, 1981) has revealed reservoir temperatures
2
Instituto Politécnico de Tomar, ESTT, Área Fı́sica, Quinta of approximately 115°C. The discharge is condi-
do Contador, 2300-313 Tomar, Portugal. tioned by the hydraulic connection between the
3
Universidade de Lisboa, Centro de Geofı́sica da Univer-
granitoid rocks and the alluvium deposits of the
sidade de Lisboa-IDL, Campo Grande, Ed. C8, 1749-016 Lisbon,
Portugal. Minho River.
2032 E. C. Ramalho et al. Pure Appl. Geophys.

Drilling of two shallow thermal wells in Caldas de 48°C (CARREIRA et al., 2004), are framed in a contact
Monção in the 1980s was based on a previous elec- domain between the biotitic chalco-alkaline porphi-
trical prospecting survey reaching 80 m depth that roid granites (post-Estefanian) and the two micas
detected three faults (A. CAVACO, 1982), but further alkaline granites (ante-Vestefalian). However, they
geophysical surveys were subsequently carried out. A emerge from the Hollocenic alluvium deposits of the
dipole–dipole survey was conducted that reached a River Minho and are affected by the above mentioned
depth of investigation of approximately 40 m in 2001 NNW fault system, occurring as three parallel align-
(GEOSONDA, 2001). In 2005, more geophysical sur- ments and their conjugates. A. CAVACO (1982) defined
veys were conducted to determine the best location three conductive alignments with apparent resistivi-
for a new well. Three seismic reflection profiles were ties ranging from 30 to 40 X m oriented toward the
therefore acquired to locate deep-rooted faults capa- NNW. One of the alignments passes into Spain, in a
ble of water transport to the surface. Seven vertical zone where local information suggests that hot water
electrical soundings (VES) and the refraction inter- springs emerged decades ago (on the right side of the
pretation of the first reflection data to arrive were also Minho River) and in Portugal (left side) (A. CAVACO,
undertaken to complement the information for the 1982). Another alignment passes by the Nossa Sen-
shallowest layers. hora da Saúde spring and the third alignment passes
by the S. Lázaro, S. Bento, and Sta. Maria springs,
the last of which dried up on opening of the borehole
2. Geological and Hydrogeological Conditions AC-2 (IGM, 1999).
Groundwater circulation is discontinuous with
Caldas de Monção is located in the Centro-Iberian dominant sub-vertical flow. Phreatic circulation of
Zone of the Portuguese Hesperian Massif, in an area water locally contaminates deep circulation thermo-
dominated by granitic rocks (RIBEIRO and MOREIRA, mineral waters; this is apparent from their chemical
1986), where the predominant rock, the Monção variability during the year. According to geological
Granite, is highly fractured (Fig. 1). It is affected by and well data, the thickness of the alluvium cover and
fault systems with predominant NNW, NNE and NE fluvial terraces is no more than 10–13 m, and these
regional orientations and tectonic mechanisms which lie directly over the granitic bedrock.
are similar to those of other thermal hot water springs in Based on isotopic analysis (CARREIRA et al., 2004)
Northern Portugal located in the same tectonic provi- and precipitation records (LIMA, 2001) the recharge
nce (Fig. 1), for example Caldas das Taipas, Caldas de area of the hydromineral system must be located at
Vizela, and Caldas da Saúde (CRUZ et al., 1996). topographic elevations from about 300–600 m and
In the studied zone are outcrops of Hollocenic located S of Caldas de Monção, and it seems to have
alluvium deposits, composed of sandy material with residence time greater than 60 years. Its TDS values
pebbles (quartz or quartzite), and Pleistocenic fluvial are approximately 500 mg/l and calculated tempera-
terraces, composed of cobbles and pebbles with clayey ture from silica geothermometers is approximately
and sandy intercalations that form conglomerates, 115°C (AIRES-BARROS, 1981).
often strongly cemented. These are located near the
geological contact with the Gerês and Monção Granite.
The studied zone is limited at E and W by two probable 3. Geophysical Surveys
faults; however neither of the faults has been mapped in
situ, because they are covered by the alluvium. The A set of geophysical surveys, comprising several
main fracturing system in the region is NNW oriented, methods, for example electrical and seismic, was
in tectonic nodes with the E–W system. conducted in the area of Caldas de Monção. These
HENRIQUES (1726) mentioned several hot springs surveys were acquired in several phases. The location
emerging in the border of the river that often get of the geophysical surveys and lines that were con-
flooded. These springs of Caldas de Monção, emer- ducted near Caldas de Monção in a topographic map,
ging with an average temperature of approximately originally scaled 1:25,000, are depicted in Fig. 2.
Vol. 169, (2012) Understanding the 3D Structure of a Thermal Water Fissured Granite Aquifer with Geophysical Studies 2033

Figure 1
Location and geology of the study area (RIBEIRO and MOREIRA, 1986)

Besides the geoelectrical deeper survey conducted a reflection seismic survey reaching depths of approx-
by A. CAVACO (1982), using a rectangle array with imately 400 m and a vertical electrical sounding (VES)
AB/2 = 300 m and MN = 10 m that reached the survey reaching depths of approximately 100 m were
depth of 80 m, other older surveys using dipole–dipole conducted to overcome the lack of information about
arrays that reached shallower depths were studied the development of structural features at greater depth.
(GEOSONDA, 2001; CARREIRA et al. 2004). More recently, Improved knowledge about aquifer geometry will
2034 E. C. Ramalho et al. Pure Appl. Geophys.

Figure 2
Location of the geophysical surveys and lines that were conducted near Caldas de Monção based on a topographic map, originally scaled
1:25,000 (Instituto Geográfico do Exército, 1997)

enable determination of where the conductive fluids in two different surveys (Fig. 2). One comprised four
flow and enable confirmation of the presence of shallow profiles located near the springs, trying to identify
faults. fault zones and zones with mineral water circulation
(GEOSONDA, 2001). The other survey comprised two
profiles located about 1,200 m south of the springs
3.1. Electrical Surveys
(CARREIRA et al., 2005, 2006).
Because of the proximity of the railway power line, Dipole–dipole data were processed using
the urban environment, and highly resistive shallow Res2DInv software (LOKE and BARKER, 1996), with
layers, electromagnetic methods, for example time an algorithm 10 times faster than the conventional
domain electromagnetics (TDEM), could not be used, minimum-squares method for large volumes of data
because they generate too much cultural noise that and which requires less computing memory. Topog-
disturbs data quality (MCNEILL, 1980). Considering raphy was not considered, because it is not
these conditions, use of traditional geoelectric methods significant. Both surveys reached approximately the
was the only possibility in this area. Therefore, two same depth; inversion results, relative location, and
dipole–dipole surveys were conducted; one, near their interpretation regarding fault zones and con-
Caldas de Monção and the other on the southern side ductive zones probably related with hot mineralised
of the spa. Also a profile of seven VES was conducted water circulation are represented in Fig. 4.
near one of the dipole–dipole lines. Keeping in mind
the lithological log of the wells AC-1 and AC-2, 3.1.2 Vertical Electrical Soundings
located in Caldas de Monção (Fig. 3), electrical data
were related and interpreted. Seven VES were acquired with a Schlumberger
array over the dipole–dipole line P2 (reflection
3.1.1 Dipole–Dipole Surveys seismic line PS1) to study the first tens of metres of
the near surface and to achieve deeper penetration
Several geoelectrical profiles using the dipole–dipole than the dipole–dipole lines P1, P2, P3, and P4. The
array were conducted in the Caldas de Monção area, AB spacing (current injection) ranged from 2 to
Vol. 169, (2012) Understanding the 3D Structure of a Thermal Water Fissured Granite Aquifer with Geophysical Studies 2035

30–40 m (Fig. 6). Interpretation of geoelectric pro-

files has enabled attribution of high resistivity
values, about 2,000 X m, to the pebble alluvium
deposits and unaltered granites at depth. On the
other hand, averagely altered granites seem to have
resistivity values ranging from 400 to 1,000 X m.
The fault planes with resistivities lower than
200 X m were attributed to anomalous zones,
potentially conductive of groundwater. Anomalous
zones corresponding to faults were also identified at
several points, enabling identification of several
conductive alignments, some of them approximately
N–S and one with NW–SE direction.
All the VES field curves and one-dimensional
models are similar (Fig. 5), showing high to very
high resistivity, corresponding to the initial dry layers
(2,000–10,000 X m). These are followed by approx-
imately 500 X m resistivity values, the low altered
granite. At the granite complex a geoelectric layer
inclusion of 100 X m and 20–30 m thick is detected;
this may correspond to groundwater circulation. This
lowest value resistivity layer matches the dipole–
dipole P2 line observed layers.
Reaching higher depths than the dipole–dipole
survey, the two-dimensional inverted model of the
VES is represented in Fig. 6. As seen, there is a
lower-conductivity shallow block between VES5 and
VES3 that may correspond to shallowest expression
Figure 3 of F1, where thermal water circulation may occur.
Lithological logs and casing of the wells AC-1 and AC-2
3.1.4 Interpretation of Seismic First Arrival Data
400 m. The apparent resistivity curves were inverted
assuming a one-dimensional layered earth, using a Refraction interpretation of the first seismic reflection
non-linear algorithm based on singular value decom- data to arrive was also performed to improve the near
position and the Levenberg–Marquardt stabilization surface imaging. The interpretation methods included
of the solution (JOHANSEN, 1977; MONTEIRO SANTOS delay-times and ray-tracing (HAENI et al., 1987). The
et al., 1996) (Fig. 5). TD curves and refraction model for profile PS1 are
shown in Fig. 7.
3.1.3 Interpretation
3.2. Reflection Seismics Survey
The dipole–dipole surveys, located approximately
1,200 m south of the thermal springs, were con- 3.2.1 Acquisition
ducted with the objective of locating the southern
expression of the faults that seem to control the Seismic reflection data were acquired in three
hydrogeological system of Caldas de Monção. selected lines (Fig. 2) with an accelerated 250 kg
These surveys only gave information about the weight-drop source (Gisco AWD-550T) and 48 fixed
shallowest layers, because none reached deeper than 50 Hz vertical receivers spaced 4 m apart (Profile
2036 E. C. Ramalho et al. Pure Appl. Geophys.

Figure 4
Inversion results from the dipole–dipole surveys conducted by GEOSONDA (2001) and CARREIRA et al. (2004, 2006), schematic relative location
and their interpretation. Interpreted faults are projected as black dashed lines and zones with probable thermal water circulation are
highlighted with orange arrows

PS1 used 7.5 m receiver spacing). The sampling rate velocities, a target depth of approximately 350 m was
during the 800-ms record length was 0.5 ms. The reached. This fix-spread geometry compared with the
source was moved every 7.5 m for each profile PS1 classical CMP method results in much higher cover-
and 4 m for the other profiles inside. To increase the age at the central part of the profiles but varying
length of profile PS1 and the investigation depth coverage from the centre of the display to its ends. In
without the need to acquire more profiles, and the authors’ experience this varying coverage has not
therefore reduce costs, additional shots were fired caused major difficulties of interpretation of the
off-end (until 71.25 m). Only one shot off-end at both resulting processed stacks. The most important con-
ends of profiles PS2 and PS3 were fired (2 m source- sequence of this geometry, in particular the receiver
first receiver offset). spacing used, was that aliased surface waves could be
Profile PS1 was acquired directly over the Hollo- present in the shot gathers but these were not
cenic alluvium, whereas PS2 and PS3 were conducted observed in reality, possibly because of a compacted
over compacted embankments. Environmental noise upper layer.
was high during data acquisition, because of the
constant presence of people, machinery, and the river 3.2.2 Data Processing
itself. In some cases, this noise was difficult to
eliminate. To enhance the signal-to-noise ratio five Data were processed with commercial software
blows of the source were used at each source location. (Seismic Processing Software-SPW, version 1.8.17,
With this acquisition geometry and local subsurface 2004) using the following major steps: pre-
Vol. 169, (2012) Understanding the 3D Structure of a Thermal Water Fissured Granite Aquifer with Geophysical Studies 2037

Figure 5
VES resistivity inversion curves (black squares, apparent resistivity field data; black circles, smoothed apparent resistivity; blue line,
resistivity model)
2038 E. C. Ramalho et al. Pure Appl. Geophys.

W E
VES 7 VES 6 VES 5 VES 4 VES 3 VES 1 VES 2

0
-10
-20

Depth / m
-30
-40
-50
-60
-70
-80

45 m
1

1.2

1.4

1.6

1.8

2.2

2.4

2.6

2.8

3.2

3.4

3.6

3.8
Log10(rho) / ohm-m

Figure 6
Two-dimensional inverted model of the VES conducted in Caldas de Monção

Figure 7
TD curves (left) and velocity model (right) obtained by refraction interpretation of the first reflection data to arrive (profile PS1). The
velocities and the depth of the interface determined have a good correlation with expected lithology. The first layer velocity is similar to that
for unconsolidated sediments (alluvium) and the second layer velocity clearly corresponds to altered and fractured granite

processing, deconvolution and trace balance, velocity The profiles, migrated in time and respective depth
analysis (constant velocity panels), surface consistent conversion, are presented in Figs. 8 and 9.
automatic residual statics, integral DMO (YILMAZ,
2001), velocity analysis, stack, and FK migration 3.2.3 Interpretation
(STOLT, 1978) with stretch. Finally, post-stack spatial
and coherence filtering and spectral whitening A fault interpretation is presented in Fig. 9. In the
(20–100 Hz) were used to reduce noise and increase interpretation of seismic attributes, several colour
data frequency. density, variable area and wiggle displays, and
This processing chain is important to highlight the geographical scales were used. In Fig. 9 we show
tenuous reflections that occur inside the granite the raw seismic amplitudes, an AGC window of
compared with the intra-alluvium reflections and 300 ms with a variable area/wiggle display and a
those originating at the contact between the alluvium vertical exaggeration of approximately 1/3. In the
and the granite. With careful processing, satisfactory horizontal axis it is shown the distance to the first
consistent reflections can be obtained in crystalline shot point for each profile. The faults were marked
rocks (KIM et al., 1994; RADZEVICIUS and PAVLIS, after discontinuities in the reflectors or seismic
1999). Depth conversion was achieved by use of patterns. It should be kept in mind that depth sections
interval velocities calculated from stacking velocities. were converted using stacking velocities because well
Vol. 169, (2012) Understanding the 3D Structure of a Thermal Water Fissured Granite Aquifer with Geophysical Studies 2039

Figure 8
Seismic reflection profiles PS1, PS2, and PS3 presented in the variable area-wiggle form. The coverage for each profile is also shown

Figure 9
Depth-converted seismic reflection profiles, with major interpreted faults (red lines). Red dashed lines indicate probable faults
2040 E. C. Ramalho et al. Pure Appl. Geophys.

A1 A3 A5 A7

P2

A1 A3 A5 A7
Depth (m)

PS1

Distance (m)

Figure 10
Two-dimensional schematic model (middle) based on seismic profile PS1 (below) and electrical profile P2 (above)

logs or a vertical seismic profile where well above the 4. Discussion

available funds for this work. Maximum offset
relative to the first receiver is 71.25 m for profile An analysis comprising simultaneous seismic and
PS1, because the first sample point of the subsurface electrical profiles located near Caldas de Monção was
(CMP) is located at 35.6 m only. conducted (Figs. 10, 11, 12; Tables 1, 2, 3). The
The horizontal resolution (Fresnel zone) of these points are indicated in Fig. 2.
profiles is 52 m at 200 ms, resulting in a dominant
frequency of 75 Hz and an average P-waves
4.1. Profile PS1
velocity of 2,000 m/s. This results in 13–26 CMP
per Fresnel zone for profile PS1 and for profiles Profile P2 partially overlaps profile PS1, between
PS2 and PS3, respectively. These values are quite geoelectric points A1–A8 (Fig. 2). However, the
satisfactory (STEEPLES, 1997), considering data shallow basin is visible in both profiles. From the
quality. The vertical resolution, according to the faults detected in PS1, only the fault near point A7 is
WIDESS (1973) criteria and assuming similar values observed in profile P2 from GEOSONDA (2001). The
of frequency bandwidth, dominant frequency, and other seismically detected faults are observed at
seismic wave velocity is 6.6 m. In the sedimentary depths greater than the depth of investigation in the
column the resolution will be approximately half GEOSONDA (2001) survey. An exception is the western-
this value. most fault that reaches the first 20 m of the near surface
Vol. 169, (2012) Understanding the 3D Structure of a Thermal Water Fissured Granite Aquifer with Geophysical Studies 2041

D3 D4 D5 D6

P4
Depth (m)

Distance (m)

PS2

Figure 11
Two-dimensional schematic model (middle) based on seismic profile PS2 (below) and electrical profile P4 (above)

but this area was not covered by profile P2. However, as reflectors with good continuity in the first 20 m.
interpreted from Fig. 6, the faults F1 and F2 can also be According to well data AC-1 and AC-2 (Fig. 3) the
interpreted as acting as barriers for hot thermal water uppermost reflector corresponds to an intra-alluvium
circulation, reaching about 50 m depth. interface (clay–gravel) located approximately 4–6 m
The seismic units in this profile deepen slightly deep whereas the second represents the gravel–
from E to W. The central sector (320–180 m) has granite contact, which is approximately 12 m deep.
collapsed because of normal faulting. This subsi- Above this reflector interval velocities found vary
dence increases with depth. The profile shows two between 550 and 1,175 m/s. Below, velocities range
2042 E. C. Ramalho et al. Pure Appl. Geophys.

P1

C1
C2 C3 C4

PS3

Figure 12
Two-dimensional schematic model (middle) based on seismic profile PS3 (below) and electrical profile P1 (above)
Vol. 169, (2012) Understanding the 3D Structure of a Thermal Water Fissured Granite Aquifer with Geophysical Studies 2043

Table 1 changes inside the unaltered granite. This area is

Physical properties attributed to the materials of the conceptual characterised by velocities of 3,848–4,447 m/s. Bel-
model of profile PS1 low 150–200 m, there is an area of strong,
Material Characteristic q (X m) Vp (m/s)
discontinuous, low-frequency reflectors. Besides the
usual attenuation versus depth effect, this zone may
1 Alluvium \2,000 550–1,175 also represent distinct geological characteristics,
2 Altered granite 400–1,000 2,244–4,440
3 Weakly altered granite [2,000 3,848–4,447
probably another facies change in the granite. Its
velocity ranges from 4,467–6,304 m/s.
Large variations in the seismic P-wave velocity of the 2nd layer are
attributed to the degree of fracturing
4.2. Profile PS2
Table 2
Profile PS2 partially overlaps profile P4, from
Physical properties attributed to the materials of the conceptual
model of profile PS2 points D2 to D6 (Fig. 2). The profiles are rotated at
20° to each other. In profile PS2 four faults were
Material Characteristic q (X m) Vp (m/s)
interpreted, two of these have been marked as
1 Alluvium \2,000 304–746 probable, because they are not clear in the time
2 Altered granite 400–1,000 1,975–3,147 sections and may result from the depth conversion
3 Weakly altered granite [2,000 4,866–4,902
process. These two faults, located close to point D5 at
depths of approximately 20–80 m, correspond to
Table 3 heterogeneities observed in profile P4 between points
Physical properties attributed to the materials of the conceptual D5–D6, but apparent resistivity values in the area do
model of profile PS3 not suggest water circulation. The two other inter-
Material Characteristic q (X m) Vp (m/s)
preted faults in the seismic section, the westernmost
between D3–D4 and the easternmost around points
1 Alluvium \2,000 644–1,711 D5 to D6, are also visible in profile P4. In profile PS2
3 Weakly altered granite [2,000 3,012–3,420
they are deep rooted and also reach the near surface
(20 m). Only the westernmost (D3–D4) appears to be
from 2,244–4,440 m/s. This information agrees with water conductive, with very low apparent resistivity
the refraction interpretation of this profile (Fig. 7), values.
which shows a sharp change of velocity from This seismic line shows some differences from the
approximately 600 m/s to more than 4 km/s at a previous profile but the three seismic patterns
depth of approximately 12 m. The first velocity is observed in PS1 also show up in PS2. The shallowest,
typical of unconsolidated sediments whereas the composed of good, continuous reflectors, reaches
second layer velocity corresponds to altered and approximately 30 m in depth and is characterised by
fractured granite. Below these two reflectors, a zone low interval velocities ranging from 304–746 m/s. It
of tenuous, high-frequency, sometimes discontinuous corresponds to intra-alluvium and alluvium–granite
reflectors is found until a depth of 150–200 m. Its interfaces. The 2nd pattern composed of weak
continuity is good in an area between horizontal 280 amplitude and continuity reflectors has velocities
and 100 m in which a good reflector approximately between 1,975 and 3,147 m/s. Toward the west of the
40 m deep can easily be followed. Toward the W of 130 m position, after the location of a fault, the
280 m, it loses its continuity, probably because of continuity and strength of the reflectors increases.
faulting. According to the AC-1 geological log, this Below 150–200 m depth the third seismic pattern is
seismic horizon may correspond to the appearance of composed of discontinuous, low frequency reflectors,
fresh granite, which reaches a depth of 36 m in that possibly indicative of a facies change in the granites,
well. Between the W end of the profile and 300 m whose velocity ranges from 4,866–4,902 m/s. Two
position two reflectors can be followed at a depth of deep faults at approximately 60–50 m and approxi-
80–120 m. These two horizons probably signal facies mately 145–110 m, reach the near surface (at least
2044 E. C. Ramalho et al. Pure Appl. Geophys.

Figure 13
Three-dimensional schematic model of the area of Caldas de Monção using geoelectric, borehole, and seismic results. Black lines correspond
to interpreted faults. F1, F2, and F3 are the faults represented in Fig. 4. Red triangles correspond to the VES. Material 1 corresponds to
alluvium, Material 2 corresponds to altered granite, and Material 3 corresponds to weakly altered granite, as mentioned in Tables 1, 2 and 3

20–30 m). Two other faults at 85 m and 60 m geoelectric profile this layer is horizontal but in
positions affect the first 40–60 m of the subsurface. the depth converted seismic section a small basin
However the latter faults are not visible in the time shows up. Because, in the first 200 ms, the seismic
sections, and probably result from the depth-conver- signal is good and we trust the interval velocities
sion process. used, we believe that the difference observed is
because of the lower resolution of the geoelectric
survey. In profile P1 no fault is observed because the
4.3. Profile PS3
high resistivity shallow layer is quite homogeneous.
Profile PS3 and geoelectric profile P1 almost Therefore, the fault observed in the seismic section
overlap (between points C1–C5), with the latter slightly E of point C3 that reaches approximately
continuing east for further 50 m (Fig. 2). Two faults 180 m depth is not water conductive.
were observed in the seismic profile, although the After the interpretation and lateral correlation of
westernmost (point C2) is marked as probable, just the structures found in the geophysical profiles two
like the deeper part of the easternmost fault (point locations appear best for a thermal water production
C3). The high resistivity layer in this profile is well. They are both in profile PS1 (Fig. 4). The first in
thicker than in the other geolectric profiles, in the continuity of well AC-1 near point A7, and the
agreement with the seismic sections. In the other inside the collapsed area between points A1–A3.
Vol. 169, (2012) Understanding the 3D Structure of a Thermal Water Fissured Granite Aquifer with Geophysical Studies 2045

This profile shows major differences from the the collapsed area between points A1–A3. For
other seismic lines. Not only are the reflectors more logistical and technical reasons, Well G1 site was
irregular, because the second seismic pattern (com- chosen (Figs. 2, 13). The well was drilled down to
posed of weak discontinuous reflectors) is absent, but 250 m depth, as projected. The geological sequence
the thickness of the first pattern here is approximately crossed by the well comprises 15 m alluvial and
60 m, because of the greater thickness of the pebble deposits followed by altered granite, highly
embankment. Its velocity varies between 644 and altered grey/rose granite, and low altered grey/rose
1,711 m/s. In the depth-converted section a small granite. Water flow is 12.5 l/s. Well design included
basin is seen between 110 and 30 m, limited by a casing and grouting of the first 130 m, and keeping
probable deep rooted fault to the west and a fault to the well in an open-hole in the lower section to
the east. Below this pattern, reflectors of reasonable improve hydraulic performance. Geothermal circu-
amplitude and continuity are observed. These have lation has already been detected in the alluvium. The
steep dips because of the significant faulting men- granite was highly faulted and with evidence of
tioned above that affects the uppermost 20–30 m of mineral water circulation, with a hydraulic head of
the near surface. The interval velocities found for this approximately 5 m above the ground. The flowing
seismic package are from 3,012–3,420 m/s. Two discharge at the well head is 5 l/s with a water tem-
major, deep-rooted faults are observed in the depth perature of 47°C. Transmissivity, evaluated from the
sections. Their presence in the time sections is less specific capacity, is ca 100 m2/day, therefore, in the
pronounced so they were signalled as probable faults. upper magnitude of the thermomineral water of the
The eastern one, near the 120 m position (at the Portuguese Hercynian Basement, (CARVALHO 1993,
surface), reaches the shallow surface (at least 2006) whereas the average transmissivity for wells up
20–30 m) and is also observable in the time section. to 150 m deep with ‘‘normal water’’ is 1.7 m2/day
The westernmost, located close to the 50 m position, (CARVALHO, 2006). This work shows the importance
is not so clearly observed in the time section. of geophysical methods in learning more about the
With the acquired seismic, geoelectrical (either geometry of mineral water aquifers.
dipole–dipole or VES), and well information data, a
conceptual three-dimensional model based on
inferred physical and geometric properties from each Acknowledgments
section was produced (Fig. 13).
The authors wish to thank Luı́s Carvalho, head of
Geopesquisa, and Isabel Nascimento, from the
5. Conclusions Municipality of Monção, for allowing publication
of the data presented here. Thanks are also due to
The replacement of the Caldas de Monção springs Eng. Ricarda Moura, from Geosonda and to Carla
by boreholes less vulnerable to pollution was a task Lourenço, from DGEG. We are also grateful to
involving investigation of the aquifer by several EuroGeol José Martins Carvalho for his helpful
methods. The schematic models that were built with suggestions and for sharing with us his knowledge
the available information enabled lateral correlation of the geology and hydrogeology of Caldas de
of the structures that led to the three-dimensional Monção. The authors also wish to thank both
conceptual model represented in Fig. 13. reviewers of this paper.
With the models made for each section and fol-
lowing the mineral water aquifer whose geometry REFERENCES
was drawn on the basis of seismic and electrical
prospecting, two locations were the strongest candi- A. CAVACO (1982). Estudo hidrogeológico das nascentes minerais
dates for the production well; both were in seismic de Monção. Plano de trabalhos.
AIRES-BARROS, L. (1981). Sobre as potencialidades em geocalor de
profile PS1. The first was located in the continuity of alguns pólos geote´rmicos do continente portugueˆs. Boletim de
well AC-1, near point A7, and the other was inside Minas, 18, n. 2, 8p.
2046 E. C. Ramalho et al. Pure Appl. Geophys.

CARREIRA, P. M., MARQUES, J. M., ANDRADE, M., NUNES, D. (2004). HAENI, F. P., DEBORAH, G., ELLEFSEN, K. (1987). Micro-computer
Groundwater flow patterns in Caldas de Monção thermomineral based version of SIPT—A program for the interpretation of
systems evaluated from isotopic and geochemical data—NW seismic refraction data. USGS, Open-File Reports, 87-103-A.
Portugal. Proceedings of the International Workshop on the HENRIQUES, F. F. (1726). Aquile´gio Medicinal. Edição fac-similada
Application of Isotopes in Hydrological and Environmental de 1998. Instituto Geológico e Mineiro. Lisboa. 288p.
Studies, UNESCO-IAEA, Paris, France, 6-8 September 2004, IGM (1999). Catálogo de Recursos Geotérmicos. CD-Rom de
p. 47. divulgação.
CARREIRA, P. M., MARQUES, J. M., CARVALHO, M. R., MONTEIRO JOHANSEN, H. K. (1977). A Man/Computer interpretation system for
SANTOS, F. A., MATIAS, H., LUZIO, R., NUNES, D. (2006). Fluid/ resistivity soundings over a horizontally stratified earth. Geo-
mineral equilibrium calculations and chemical geothermometry physical Prospecting, 25: 667–691.
of Monção thermal waters (NW-Portugal): a tool to estimate the KIM, J. S., MOON, W. M., LODHA, G., SERZU, M., SOONAVVALA, N.
deep thermal waters composition. Abstracts of the International (1994). Imaging of reflection seismic energy for mapping shallow
Symposium ‘‘Aquifer Systems Management’’ ? CD with Full fracture zones in crystalline rocks, Geophysics, 59, 753–765.
Texts, Theme 4, 4–15. LIMA, A. S. (2001). Hidrogeologia de Terrenos Granı́ticos. Minho—
CARREIRA, P. M., MARQUES, J. M., MONTEIRO SANTOS, F. A., Portugal. PhD Thesis. Universidade do Minho, 451p.
ANDRADE, M., MATIAS, H., LUZIO, R., NUNES, D. (2005). Role of LOKE, M. H., BARKER, R. D. (1996). Rapid least-squares inversion
geophysics, geochemistry and environmental isotopes in the of apparent resistivity pseudosections by a quasi-Newton
assessment of Caldas de Monção low-temperature geothermal Method. Geophysical Prospecting, 44, 131–152.
system (Portugal). Geothermal Resources Council Transactions, MCNEILL, J. D. (1980). EM34-3 Survey interpretation techniques.
29, 209–213. Technical Note TN-8. 6p.
Instituto Geográfico do Exército (1997) Carta Militar de Portugal, MONTEIRO SANTOS F. A. M., DUPIS, A., AFONSO A. R. A., MENDES-
Continente—Série M888, n. 3. VICTOR, L. A. (1996). An audiomagnetotelluric survey over the
CARVALHO J. M. (2006). Prospecção e pesquisa de recursos hı́dricos Chaves geothermal field (NE Portugal). Geothermics 25 (3):
subterrâneos no Maciço Antigo Português: linhas metodológicas. 389–406.
Dissertação de doutoramento na Universidade de Aveiro. 292 RADZEVICIUS, S. J., PAVLIS, G. L. (1999). High-frequency reflections
p. ? annexes ? hydrogeological map. in granite? Delineation of the weathering front in granodiorite at
CARVALHO, J. M. (1993). Mineral and thermal water resources Piñon Flat, California, Geophysics, 64, 1828–1835.
development in the Portuguese Hercynian Massif. Hydrogeology RIBEIRO, M. L. and MOREIRA, A. (1986). Notı´cia explicativa da
of Hard Rocks. International Association of Hydrogeologists, Folha 1-B—Monção. Serviços Geológicos de Portugal, 46p.
IAH, Oslo, 24 (1): 548–561. STEEPLES, D. W. (1997). Near-Surface Seismology, Continuing
CRUZ, J. (1992). Termas e águas engarrafadas em Portugal. Edição Education Short Courses, Soc. Explor. Geophys.
DGGM. STOLT, R. H. (1978). Migration by Fourier transform, Geophysics,
CRUZ, J., LOURENÇO, M. C., RAMALHO, E. C., RODRIGUES DA SILVA, A. 43, 23–48.
M., CORREIA, A. (1996). Breve descrição das ocorreˆncias termais WIDESS, M. B. (1973). How thin is a thin bed, Geophysics, 38,
com temperatura superior a 20°C em Portugal Continental. 1176–1180.
Boletim de Minas, 32, 132–134. YILMAZ, Ö. (2001). Seismic Data Analysis, Investigations in Geo-
GEOSONDA (2001). Trabalhos de Prospecção Geofı´sica para a imp- physics, 10, S. M. DOHERTY (Ed.), Soc. Explor. Geophys, 2027p.
lantação de furos de captação de água termo-mineral.
9p. ? annexes.

(Received March 3, 2011, revised December 6, 2011, accepted December 8, 2011, Published online February 16, 2012)