Catena 137 (2016) 572–580

Contents lists available at ScienceDirect

Catena

journal homepage: www.elsevier.com/locate/catena

Deep weathering of granitic rocks: A case of tunnelling in NW Spain

Daniel Arias, Luis Pando ⁎, Carlos López-Fernández, Luis M. Díaz-Díaz, Álvaro Rubio-Ordóñez
Department of Geology, University of Oviedo, Jesús Arias de Velasco, 33005 Oviedo, Spain

a r t i c l e i n f o a b s t r a c t

Article history: This paper highlights a case of extensive development of saprolith from granitic bedrock in NW Spain. The
Received 9 March 2015 weathering front at the study area extends to more than 250 m in depth. The drilling of a railway tunnel
Received in revised form 26 October 2015 (Madrid–Galicia high-speed line) through the bedrock and saprolith resulted in significant technical issues and
Accepted 27 October 2015
economic losses. Cross-sections revealing the 3D complexity of saprolith development are mainly based on field-
Available online xxxx
work, boreholes, laboratory tests, and the excavation works. Saprock is always observed overlying unweathered
Keywords:
igneous bedrock. Two saprolite facies, one clay-rich and the other clay-poor overlie the saprock; the recognition
Deep saprolite is based on several criteria. These deposits gradually thicken from the boundaries between the granitoid and
Granitoid metasedimentary roof pendants. Factors controlling saprolith development are discussed, and the chemical
Roof pendant and mineralogical changes and trends that occur are described. Variations in some geotechnical properties and
Tunnelling the loss of quality of the rock masses from an engineering perspective are also quantified.
Weathering © 2015 Elsevier B.V. All rights reserved.

1. Introduction From the environmental viewpoint, landslides associated with shal-

The development of residual soils covering granitic rocks is a widely
known process (e.g., Girty et al., 2013; Ng et al., 2001; Oliva et al., 2003;
Ollier, 2010; Sequeira Braga et al., 2002). Diverse changes in physical
properties and petrographic features of granite occur during weathering.
Overall, three main stages are usually distinguished in this process:
(a) fresh rock, (b) saprock (i.e., partially weathered granite) and (c) sap-
rolite (Power and Smith, 1994; Taylor and Eggleton, 2001; Graham et al.,
2010). Saprock and saprolite, also collectively known as saprolith, are
usually characterized by preserving the original fabric and structure of
the granitic lithotype.
Mineralogically, biotite plays an important role in the alteration of
granite rocks. For example, it may produce a local volume increase as it
is transformed into hydrobiotite, vermiculite or interstratified biotite–
vermiculite (Isherwood and Street, 1976; Jeong and Kim, 2003;
Nettleton et al., 1970; Parizek and Girty, 2014; Taboada and García,
1999a). In turn, hydrolysis and hydration reactions induce the transfor-
mation of feldspars into clays and colloids which may migrate from the
rock (Carrasco and Girty, 2015; Taboada and García, 1999b; White et al.,
2001). On the other hand, the alteration progresses slowly in K-feldspar
and quartz grains, so they commonly form an unaltered framework in
saprolite. Hence, two main trends are differentiated in the weathering
of plutonic rocks: (a) the biotite-dominated trend (Girty et al., 2013)
and (b) the plagioclase-dominated trend (Nesbitt and Young, 1984,
1989).
⁎ Corresponding author.
E-mail addresses: darias@geol.uniovi.es (D. Arias), lpando@geol.uniovi.es (L. Pando),
lopezcarlos@uniovi.es (C. López-Fernández), diazluis.uo@uniovi.es (L.M. Díaz-Díaz),
arubio@geol.uniovi.es (Á. Rubio-Ordóñez).
low covers composed of saprolite and soil are a well-known hazard
(Bacellar et al., 2005; Durgin, 1977; Lan et al., 2003). The geotechnical
properties of the weathered materials can also result in excavation
and support issues during underground construction (De Abreu et al.,
2007; Grasso et al., 2003; Guatteri et al., 2013; Romero and Clayton,
2006; Sousa and Einstein, 2012). Moreover, saprolith occasionally
forms aquifers (e.g., Dewandel et al., 2006) whose size is greatly in-
creased when weathering extends in depth. The presence of both
groundwater and loose saturated deposits occasionally causes severe
engineering problems during the excavation in railway projects, as oc-
curred in the Nakayama tunnel (Jōetsu Shinkansen line, Japan) that
needed important drainage works (Egger et al., 1982) or in the Ayas
tunnel in Turkey (Mahmutoglu et al., 2006).
This research relates to a case of weathered granitoids tunnelled in
NW Spain that caused significant technical issues and economic losses
during the excavation works, especially due to the presence of saprolith.
Ground investigations linked to the project (Madrid–Galicia high-speed
railway) enabled the study of the weathering environment whose geo-
logical setting is presented. To do this, the main features of the geolog-
ical units have been analysed from structural, chemical, mineralogical,
geotechnical, and hydrogeological viewpoints.
2. Background
Two main subsections of high-speed railway are differentiated in
NW Spain (Fig. 1): (a) “Puebla de Sanabria–Ourense” (currently under
construction) with a length of 240.8 km, including 136.2 km projected
through 38 tunnels; (b) “Ourense–Santiago de Compostela” (completed
in 2011) reaching 87.5 km in length, with 31 tunnels (29.3 km of

http://dx.doi.org/10.1016/j.catena.2015.10.026
0341-8162/© 2015 Elsevier B.V. All rights reserved.
 D. Arias et al. / Catena 137 (2016) 572–580
Fig. 1. Location of the main granitoids mapped in NW Spain. Tunnels where deep saprolite
has been excavated at the Madrid–Galicia high-speed railway line are also indicated.
excavation). Altogether, 49 tunnels traversed mostly granitic rocks and
a diverse suite of weathered rocks. The saprolith and soil rarely exceed
10–20 m in thickness, but in the Burata tunnel the weathering process
was more intensely developed than in other areas. This tunnel,
constructed between June 2006 and May 2010, is 3998 m in length
and has a cross-sectional area of 85 m2. It reaches a maximum depth
of 150 m below the surface and 93 m on average. An emergency gallery
(with a cross-sectional area of 13 m2) transversally connected to the
tunnel was built every 250 m along the length of the tunnel.
Both tunnel portals in Burata crossed saprolith and soil that were
thicker than expected; despite this precedent, the weathering profile
was not reviewed. At several sections throughout the projected tunnel,
the preliminary geological surveys interpreted saprolite as material
linked to fault zones with restricted development. Nevertheless, when
drilling began using the New Austrian Tunnelling Method (NATM),
the weathered rocks exhibited a very poor geotechnical quality. The
saprolite behaved similarly to highly deformable soil, so the construc-
tive solutions (i.e., steel ribs, shotcrete and bolts) were not as effective
as expected. The measurements of convergences during construction
revealed a reduction of the theoretical tunnel profile of over 10%
(in area) on critical sections excavated in saprolite. It was necessary to
build in some sections a concrete inverted arch at the tunnel bottom,
and the support was reinforced (Fig. 2) with umbrellas of subhorizontal
micropiles, injections and jet-grouting (Castells et al., 2006). Addition-
ally, the Engineering Geology Research Group based in the University
of Oviedo conducted an urgent geological survey. New boreholes and
fieldwork should help to identify unknown roof pendants and weath-
ered rock masses before they were crossed by the tunnel.
Saprolite saturation also significantly complicated the tunnelling
due to the ground permeability and the hydrostatic pressure reached
at depth (occasionally greater than 1 MPa). During the excavation,
water drained into the tunnel (N20 l/s) mainly along granitoid–roof
pendant boundaries. It was necessary to drill horizontal drainage wells
to drawdown the groundwater table. The piezometric fall triggered a
subsidence process that was monitored; a surface settlement of up to
10 cm was measured, although it did not cause structural damage
(López-Fernández et al., 2013). The tunnel construction was delayed
for two years and the cost increased by 46 million Euros.
573
As a result of the constructive experience gained during the tunnel-
ling of Burata, the geological explorations were performed more rigor-
ously in other areas of the railway. This enabled in an additional four
tunnels (La Canda, Vilavella, Padornelo, and Prado) the identification
of saprolite at depth, making it possible to optimize the technical solu-
tions from the start of drilling.
3. Geological setting
The bedrock in northwestern Spain corresponds to internal zones of
the Iberian Massif, a cratonic block related to the Hercynian orogeny.
Specifically, in this study area, the batholiths are the dominant rock
masses (Fig. 1) and intruded (over the course of several events) into a
basement that consists of Precambrian and Cambro-Ordovician units
(Capdevila et al., 1973; Dallmeyer et al., 1997; Pérez Estaún et al.,
1991). The metasedimentary roof pendants are largely composed of
mica schist, quartz schist, quartzite and paragneiss, with interbedded
volcanic or subvolcanic layers. In these rocks, up to three deformation
phases have been differentiated (Marcos, 1973).
Three associations of granitoids are distinguished depending on the
compositional characteristics and their syntectonic or post-tectonic
emplacements (Bellido et al., 1992; Corretgé, 1983): (a) syntectonic
calc-alkaline granitoids; (b) syntectonic to late-tectonic peraluminous
granites; and (c) post-tectonic calc-alkaline granites. Syntectonic
tonalite–granodiorite–monzogranite (TGM) plutons are the oldest
granitoids. They are composed of amphibole–biotite and quartz diorites,
tonalites, biotite granodiorites (the most common lithology) and bio-
tite–muscovite monzogranite. This lithological association is character-
ized by high K-content and represents I-type magmatic series (Galán
et al., 1996). Syntectonic peraluminous granites are the most prevalent
granitoids in NW Iberia. They include peraluminous muscovite–biotite
to muscovite monzogranites, with minor feldspathic granites and ap-
lites; all autochthonous, subautochthonous and allochthonous anatectic
granites are included. This group corresponds to peraluminous rocks
characterized by low-Ca and Ti content, and represents S-type magmat-
ic series (Cuesta and Gallastegui, 2004). Concerning the post-tectonic
association, calc-alkaline and subalkaline granitoids form the largest
plutons. They have features closely related to the syntectonic T–G–M.
In this case, the Burata tunnel extends through the “Ourense–
Carballino–Rodeiro” batholith. This intrusive body represents a typical
example of the late Variscan post-kinematic magmatism in NW Iberia,
which was intruded across the boundary of two different belts of defor-
mation (Barrera et al., 1989). It is mainly composed of peraluminous
(two mica; usually muscovite N biotite) calc-alkaline granitoids
Fig. 2. Excavation procedure in the Burata tunnel based on the top-heading and bench
method (with a micropile umbrella in this photo).
574 D. Arias et al. / Catena 137 (2016) 572–580
(Capdevila and Floor, 1970), derived from magnesian alkali-calcic series
(Cuesta and Gallastegui, 2004). The batholith also includes aplite and
pegmatite dikes whose thicknesses reach just a few centimetres.
4. Methods
This research is supported by data obtained from fieldwork, trial pits,
rotary boreholes, geophysics, underground observations, and laboratory
testing. Fieldwork was addressed to improve the surface geological map
of the study area (1:2000) and preparing an inventory of water points.
Geophysical exploration involved the use of sonic logging, seismic
refraction, and electrical resistivity tomography (ERT) whose logs
have been reviewed. Meanwhile, 31 boreholes carried out along the
Burata tunnel alignment were collected (21 drilled in preliminary sur-
veys and 10 during the excavation works), providing 3870 m of logs
that could be checked in detail. Such data were supplemented by field
tests: standard penetration test, pressuremeter and hydraulic con-
ductivity. In terms of rock mass quality for engineering purposes,
the most common classifications were used: the Rock Mass Rating
(Bieniawski, 1989) and the Q-system (Barton et al., 1974). The RMQR
system was also implemented to assess the physical state of the rock
mass in relation to the intact rock (Aydan et al., 2014); it provides a
sound interpretation of water incidence and takes into account the
degradation degree, two essential aspects of this case study. All these
data, along with the information provided by the tunnelling progress,
allowed a new geological cross section of the Burata tunnel to be
constructed.
The boreholes and trial pits were also exploited for sampling in both
the bedrock and the weathered rocks. Laboratory tests derived from
collected samples were used to study the main properties of the mate-
rials and to evaluate the weathering process. At first, thin sections cut
from rock samples were studied. Next, the chemical compositions of
the weathered rocks were determined by X-ray fluorescence spectrom-
etry (XRF Philips PW2404 equipped with a rhodium anode tube). The
mineralogical composition for each geological unit was identified by
X-ray diffraction analysis (XRD Philips X'Pert Pro equipped with a
cupper anode tube) performed on bulk samples. The quantitative
phase analysis was based on the Rietveld method.
Regarding the geotechnical properties, more than 150 samples
(altered and unaltered) were subjected to indexing and mechanical
laboratory tests. They were carried out mostly according to UNE
standards (Spanish technical specifications). Features determined dur-
ing this phase of our study included particle size distribution (UNE
103101:1995) and soil classification (USCS); mass-volume relation-
ships (UNE 103300:1993, UNE 103301:1994, UNE 103302:1994);
Atterberg limits (UNE 103103:1994, UNE 103104:1993); uniaxial com-
pressive strength (UCS) and elastic constants (UNE 22950-1:1990, UNE
22950-3:1990, UNE 103400:1993); Point Load Test (UNE 22950-
5:1996); and rock abrasiveness (NF P94-430-1).
5. Results
5.1. Main features
The area in and around Burata tunnel is underlain by four geological
units identified during both fieldwork and borehole core inspection
(Fig. 3). Three of them are rock units and the other includes saprolite
and soil. Going from bottom to top the four units are: (i) unaltered or
very slightly weathered bedrock (I–II grades according to ISRM, 1981)
of medium-grained granite in which textures range from equigranular
to porphyritic; (ii) fractured massive granite including isolated fresh
pieces, having brown-red coloured patches due to moderate weathering
(III or even IV grade); (iii) saprolite with the original fabric of the granite
conserved, enclosing some corestones (i.e., spheroidal shaped enclaves
of unaltered bedrock); and (iv) masses of metasedimentary country
rock enclosed by the batholith. In preliminary surveys seismic refraction
exhibited a perceptible variation of velocities between the bedrock
(Vp N 2500 m/s) and the saprolite (b 1,000 m/s), and ERT profiles pro-
vided differentiated values for the bedrock (resistivity N1.5 kΩ·m) and
the saprolith (b 1 kΩ·m).
From the tectonic viewpoint, two major NE-striking faults were
identified dipping 50–60° SE. These faults controlled the geometry of
the hydrographic network in the surroundings of the tunnel alignment.
None were crossed by the excavation. Fieldwork and borehole core in-
spection enabled the study of the systematic fracturing of rock masses.
Three major joint sets were differentiated: (i) NE–SW, dipping 70–80°
SE, spacing 1–3 m; (ii) NE–SW, dipping 80–85° NW, spacing 5–10 m;
(iii) ESE–WNW, dipping 20–30° NNE, spacing 4–8 m. In the meta-
sedimentary rocks, the most common orientation of the foliation is
NE–SW, dipping 40–80° SE.
When the geological setting in the area of the Burata tunnel
was completely updated, it showed an irregular and noticeably indent-
ed weathering front occasionally reaching more than 250 m in depth
(Fig. 4). Above that, the saprock thickness usually ranges between 10
to 100 m, with significant lateral variations. Within the overlying
saprolite unit, two facies were differentiated depending on the clay
content and location. A clay-poor saprolite (C-P) extends to the land
surface and along the boundaries between intrusive rocks and roof pen-
dants, sometimes reaching 100 m in thickness. The clay enriched facies
(C-R), when present, are located below the clay-poor saprolite, and
their thicknesses usually do not exceed 20 m (Fig. 4a). This overall ge-
ometry of weathering lithotypes was recognized throughout most of
the area encompassing the Burata tunnel. Finally, the roof pendants ex-
tend to varying depths, from a few metres to over 150 m (Figs. 3 and 4).
5.2. Mineralogy
The mineralogical content was determined in four samples of fresh
rock and five for each altered unit. Bedrock in the Burata tunnel is chief-
ly composed of quartz and feldspars (85%) and, as explained earlier, two
micas. Among the accessory minerals are apatite, zircon, sillimanite,
andalusite, garnet, and tourmaline. The mineralogical content of the
saprock is quite similar to the unaltered rock (Fig. 5). However the pro-
portion of biotite is lower and kaolinite is identified (≈3%). These min-
erals along with textural evidences of weathering were also observed in
thin sections (Fig. 7). In the C-R saprolite, the percentages of quartz and
muscovite are higher, whereas the contents of potassium feldspar and
plagioclase are lower. The proportion of biotite is minimal in the C-R
saprolite, but the proportion of clay minerals (including montmorillon-
ite) reaches 20% on average. The C-P saprolite exhibits an even higher
proportion of quartz (N 50%) and muscovite, with a noticeable loss of
clay minerals. This regolith unit is therefore dominated by weathering-
resistant minerals.
5.3. Chemical data
Fifteen samples from the three units that form the saprolith were
geochemically evaluated (Table 1). Data for bedrock compositions
(see Appendix A) are from Barrera et al. (1989). Within the regolith,
weathering processes control element redistribution patterns. The
degree of weathering can be evaluated by tracing the mass transfer of
labile elements (e.g., Parker, 1970; Nesbitt and Young, 1982). Another
tool is the use of simple molecular ratios (e.g., Nesbitt and Young,
1982). In order to study the evolution of the weathering that affected
the collected samples, the A–CN–K system was used. Samples were
plotted in a centred ternary diagram (von Eynatten et al., 2002), and
two different trends can be distinguished (Fig. 6). The first trend is
defined by the C-P saprolite, while the second is defined by the C-R
samples. These trends are characteristics of the incongruent dissolution
of plagioclase, where Ca and Na are leached and removed from the
system (Nesbitt and Markovics, 1997; Nesbitt and Young, 1984, 1989).
Both trends are directed away from the p(CN) apex, with the gradient
 D. Arias et al. / Catena 137 (2016) 572–580
Fig. 3. Metasedimentary roof pendant surrounded by the underlying intrusion and bore-
hole cores of the geological units identified in the studied area: (a) fresh rock,
(b) saprock, (c) saprolite, (d) roof pendant.
of the trend defined by the C-R samples being greater than that related
to the C-P specimens.
To obtain a mass-balance model, an immobile element or elements
must be identified. Titanium and zirconium are well-known to be little
affected by solutions in the regolith, and as a result they are not com-
monly mobilized during weathering. Assuming that Ti is immobile, cal-
culation routines for mass-change were based on the log-ratio model of
Woronow and Love (1990). The assumption that Ti was immobile was
verified using the Ague (1994) bivariate plot. All uncertainties are
expressed at the 95% confidence level.
Transformation from unaltered bedrock to saprock indicates a statis-
tically significant addition of K and LOI mass. In addition, Si, Al, Fe and
P are also increased, whereas significant loss of Mn mass is evident
Fig. 4. Geological cross sections showing the extreme depth reached by the weathering.
575
(Table 2). Conversion to C-P saprolite involved the addition of Si, K,
LOI, Al and Fe mass. Additionally, significant losses of Ca, Na and P
mass likely occurred. In contrast changes in the C-R saprolite are quite
different. While the mass loss of Ca, Na and P is maintained, the other
elements (Si, Al, Fe, K, Mn, Mg) and especially the LOI suffer substantial
increases. The greater loss of Ca and Na mass in the C-R saprolite relative
to the C-P samples is consistent with p(A)–p(CN)–p(K) data shown in
Fig. 6.
5.4. Physical and mechanical properties
5.4.1. Rock units
Fresh bedrock in the Burata tunnel has an average dry unit weight of
25.6 kN/m3 (Fig. 8), and the porosity usually varies from 0.4 to 1.5%
(three out of four samples). The bulk density of the saprock is slightly
lower (b3%) compared to the unaltered granitoid. From a geome-
chanical viewpoint, the average compressive strength in the fresh rock
corresponds to the “moderate” class (ISRM, 1978). However, the aver-
age value in saprock falls below 25 MPa (low). Young's modulus in
fresh rock ranges mostly between 3 and 17 GPa (1st and 3rd quartiles;
37 samples), and Poisson's ratio has an average value of 0.21. Modulus
ratio is medium or low in 95% of the samples (b500). Finally, the values
of the Cerchar Abrasion Index, ranging from 2 to 4.5 (12 samples), are
typical of granitic rocks.
Regarding the rock mass quality, the RMR reaches values up to 75
(good rock) in the fresh rock. This advantageous qualification is
confirmed by the Q-system because the values estimated in this classi-
fication are over 20. In sharp contrast, the saprock is classified as “poor
rock” because RMR usually ranges between 30 and 40, and Q-values
are always below 1. According to RMQR the bedrock is qualified as
“good” (RMQR ranges mainly between 69 and 76), while the values
fall to 48 in the saprock (“fair or medium”).
5.4.2. Saprolite
All of the samples of saprolite (50) are coarse-grained sands (Unified
Soil Classification System, USCS). This means that more than 50% by
weight is retained by a No. 200 ASTM sieve (75 μm). Additionally,
sand is the most common grain size in all cases (approximately 70–
75% by weight). In detail, the granulometric analysis differentiates
several USCS classes, although two of them are dominant. The main
class corresponds to sandy silt (SM)/clayey sand (SC) and has the
highest silt–clay content, reaching approximately 20% whereas the
gravel percentage is very low (b 4%). This grain size class is abundant
in the C-R saprolite. The second most frequent class dominates the C-P
saprolite and is well-graded fine to coarse sand (SW)–silty sand (SM)/
SW–SC. This class exhibits an opposite trend, as the gravel content
increases to over 20% whereas the silt–clay proportion is only 7% on av-
erage. The remaining samples were classified as SP–SM/SP–SC, SP, SW
576 D. Arias et al. / Catena 137 (2016) 572–580

Fig. 5. Mineralogical content of the samples collected.
and silty gravel (GM). Relating to plasticity, most samples do not have
(nine out for ten) or it is very low (liquid limit b 34; plasticity index b 6).
Because of the difficulties in taking undisturbed samples in slightly
cohesive saprolith, physical and mechanical tests were performed only
on the C-R saprolite. In this facies, the values of dry unit weight are usu-
ally between 17 and 21.5 kN/m3. This represents an estimated decrease
from 15 to 33% in relation to the bedrock. The average moisture content
is approximately 13%, and the void ratio reaches values above 0.45
Fig. 6. Perturbed and centred p(A)–p(CN)–p(K) diagram showing the distribution of par-
ent rock, saprock, C-R saprolite, and C-P saprolite. Note the higher gradient weathering
trend displayed by C-R samples relative to C-P specimens, and the fact that both trends di-
verge from the saprock and parent samples. They are directed at a high angle away from
the p(CN) apex. Such relationships are consistent with plagioclase-dominated weathering
(Girty et al., 2013; Nesbitt and Young, 1984).
(porosity N30%). From a geomechanical viewpoint, uniaxial compres-
sive strength is 100 kPa on average, representing a severe reduction
with respect to the saprock (Fig. 8). The pressuremeter modulus has
typical values of granular sediments (b25 MPa) even at depths greater
than 65 m, while SPT provided N-values always exceeding 30 (medium

Table 1
Chemical composition of the analysed samples.

Fresh rock Saprock C-P saprolite C-R saprolite

(N = 7)

Sample Av. 1σ SR1 SR2 SR3 SR4 SR5 CP1 CP2 CP3 CP4 CP5 CR1 CR2 CR3 CR4 CR5

Oxides (wt.%)
SiO2 72.69 0.42 71.21 72.09 69.91 72.53 72.14 79.96 81.23 78.79 78.21 79.31 70.55 70.14 72.27 71.89 72.62
TiO2 0.28 0.05 0.26 0.24 0.23 0.21 0.20 0.14 0.14 0.11 0.14 0.14 0.26 0.14 0.13 0.15 0.14
Al2O3 14.79 0.22 14.70 14.70 14.19 14.60 14.70 10.65 9.99 11.29 11.18 11.07 15.09 16.18 15.89 15.74 15.70
Fe2O3 1.27 0.20 1.43 1.30 1.19 1.15 1.12 0.74 0.70 0.60 0.96 0.95 1.80 1.95 1.17 1.16 0.97
MnO 0.04 0.07 0.01 0.01 0.01 0.02 0.01 0.01 0.01 0.01 0.01 0.01 0.02 0.06 0.04 0.03 0.10
MgO 0.41 0.05 0.41 0.38 0.40 0.34 0.33 0.22 0.22 0.18 0.22 0.23 0.84 0.76 0.45 0.66 0.46
CaO 0.70 0.08 0.63 0.59 3.33 0.56 0.54 0.03 0.08 0.13 0.59 0.08 0.66 0.07 0.02 0.17 0.09
Na2O 3.17 0.28 2.65 2.71 2.52 2.84 2.88 0.69 0.61 1.55 1.08 1.08 1.50 0.27 0.27 0.74 1.18
K2O 5.11 0.23 5.99 6.07 5.88 6.00 6.12 5.31 5.12 5.69 5.57 5.50 4.63 4.51 4.84 5.10 5.06
P2O5 0.36 0.05 0.40 0.39 0.37 0.34 0.33 0.10 0.10 0.12 0.11 0.11 0.32 0.07 0.06 0.06 0.08
LOI 1.17 0.14 1.76 1.63 1.58 1.48 1.44 1.49 1.50 1.09 1.55 1.50 4.30 5.54 4.34 4.32 3.61
Total 99.45 100.10 99.61 100.07 99.81 99.34 99.70 99.56 99.63 99.98 99.97 99.69 99.49 100.03 100.01
A 55.22 1.04 55.08 54.85 46.14 54.51 54.36 60.55 59.89 55.79 55.74 58.43 63.49 74.79 73.53 69.07 67.43
CN 24.16 1.39 20.63 20.64 33.17 21.24 21.15 6.76 6.89 13.77 14.21 10.15 15.43 2.64 2.22 6.70 9.04
K 20.69 0.77 24.29 24.52 20.70 24.25 24.50 32.68 33.22 30.44 30.06 31.42 21.08 22.57 24.24 24.23 23.53

Ions (ppm)
Ba 436.3 448.8 451.7 457.2 449.9 308.7 288.9 335.4 314.2 316.0 266.3 165.4 164.7 138.8 155.2
Co 66.3 47.1 54.3 70.6 58.0 74.8 91.5 83.1 88.5 84.6 29.7 50.8 38.7 56.4 46.4
Cr 9.4 7.4 7.7 5.1 7.4 6.4 3.8 4.0 7.6 6.0 6.1 4.3 12.2 1.6 6.1
Cu 0.0 0.0 0.0 0.0 0.0 0.0 0.0 14.7 34.6 33.4 2.7 0.0 0.0 0.0 0.0
Nb 8.3 7.9 7.7 7.0 7.0 8.2 7.9 6.7 7.9 8.0 11.8 9.8 9.1 8.9 8.4
Ni 0.3 0.0 0.5 1.4 1.2 0.0 0.0 0.0 6.6 5.1 0.5 1.0 0.0 0.0 5.7
Pb 45.0 43.0 44.8 42.5 43.5 30.4 31.5 40.9 37.4 33.4 28.9 48.3 34.0 23.6 32.2
Rb 230.9 230.2 230.1 224.2 227.6 268.5 254.7 279.1 276.6 274.3 279.0 321.6 353.7 364.0 369.9
Sr 102.4 101.8 103.5 105.8 106.9 52.4 49.0 63.0 59.8 58.5 60.5 32.3 27.7 32.5 30.5
Th 5.8 5.2 5.3 5.0 4.9 2.7 2.7 1.4 2.7 2.9 5.2 5.4 5.1 5.0 4.7
U 7.8 2.2 6.0 8.2 8.1 3.1 2.6 2.0 3.6 3.4 8.4 17.3 8.2 4.8 4.9
V 12.2 9.8 10.7 10.5 7.8 8.0 5.8 6.8 7.0 6.5 12.3 4.2 2.8 3.3 4.6
Y 7.0 7.0 6.4 5.9 5.5 3.2 2.8 1.6 2.9 3.2 7.7 26.3 16.0 17.3 13.6
Zn 55.4 50.2 47.3 46.0 43.7 39.5 35.3 23.5 58.4 58.6 90.5 113.4 73.2 102.6 89.4
Zr 126.2 119.6 120.6 125.4 119.3 82.4 79.9 61.3 79.4 81.9 126.5 79.9 79.9 73.5 69.3
 D. Arias et al. / Catena 137 (2016) 572–580 577

Fig. 7. Photomicrographs of the granite excavated in Burata. The thin sections were treated with sodium cobalt nitrite to emphasised the K-minerals. (A, B) Bedrock with a slight alteration
of the Ca-plagioclase (Pl) and Ti exsolution in biotite (Bt). (C, D) Saprock with K-feldspar (Kf) partially altered, biotite has disappeared, and plagioclase is more weathered. The rock has
undergone microfracturing, and as a result muscovite (Mu) near these fractures presents oxides along the exfoliation planes.

compactness); penetration refusal was obtained in 75% of tests per- 6. Discussion

formed on the saprolite.
5.5. Hydrogeology
According to the on-site permeability tests performed on the bed-
rock (N = 7), the hydraulic conductivity is very low when the rock
mass is unaltered or slightly weathered (Lugeon coefficients ≤ 1). In
the saprolite, the values of permeability provided by Lefranc tests
(N = 3) do not exceed 1 · 10−7 m/s at the C-R facies (poor aquifer).
At the C-P saprolite, the conductivity was indirectly estimated from
granulometric data (Hazen, 1892) because field tests were not available.
The permeability usually ranges between 8 · 10−5 and 2 · 10−4 m/s
(1st and 3rd quartiles; 18 samples), so this facies qualifies as a fair to
good aquifer.
Three main aquifers linked to the saprolite are differentiated along
the tunnel alignment. They are limited in depth and laterally by the un-
altered granitoid and are partially confined by the roof pendants. The
hydrodynamics are controlled by the granite joints and the hydrograph-
ic network, achieving a higher level of development of these aquifers in
The Valley areas. The seven piezometers installed enabled the determi-
nation of the preferred orientation of the natural flow (NE) and con-
trolled the water table depletion during the excavation works.
From a chemical viewpoint, the analysis carried out on groundwater
samples (N = 82) taken from wells indicates pollution in the shallower
saprolite (up to 20 m deep). In total, 88% of samples contain high con-
centrations of faecal bacteria and nitrates, qualifying this water as
non-potable according to Spanish legislation (RD 140/2003). Deeper
samples were not available, but it is known that the current water sup-
ply for human consumption in this area comes from deep aquifers. The
flow through the sandy saprolith most likely reduces the concentrations
of water pollutants.
The distribution and physical and chemical properties of weathered
rock masses within the Burata tunnel prior to work reported here were
insufficiently known to address tunnelling. Preliminary geological
surveys provided a good approximation to the physical heterogeneity
of weathered granitic rocks (e.g., Olona et al., 2010). However their
usefulness was restricted given the great depth of the weathering
front. In the case of the Burata tunnel it can be said that the integration
of the preliminary data were of little use, and that the factors controlling
weathering were not studied. This absence of critical information lead to
an incomplete understanding of the geological setting and the geotech-
nical behaviour of the materials lying in the path of the tunnel. There-
fore this study focused on defining geological units, their geometry,
and their physical and chemical properties. The number of boreholes
and their depth was increased in areas where weathering was more de-
veloped. In addition, it was decided to excavate the gallery prior to tun-
nel drilling. This decision lead to a greater understanding of the
engineering properties of the material that the tunnel would encounter,
and thus identified some technical issues that arose as drilling ensued.
The rock mass encountered by the Burata tunnel includes fresh rock,
saprock, saprolite, and soil. Detailed laboratory analysis revealed intense
redistribution of elemental mass during in situ weathering. Results from
geotechnical tests indicate that the values of unit weight and uniaxial
compressive strength decrease, while porosity increases. This is ob-
served in other comparable substrates (e.g., Begonha and Sequeira
Braga, 2002). Regarding the rock masses, the main rating systems reveal
the loss of quality of the saprock with respect to the unaltered bed-
rock. However, two aspects stand out in this case: the high depth that
weathering reaches and the clear distinction between two types of sap-
rolite (here called C-R and C-P).
The orientation of the weathering front seems linked to the primary
foliation within the metasedimentary rocks. It dips approximately 50–
578 D. Arias et al. / Catena 137 (2016) 572–580

+556 (+306/−209)
+151 (+69/−54)
+64 (+45/−35)
LOI

+25 (+40/−30)
−37 (+26/−18)
−55 (+28/−17)
P

+120 (+59/−46)
+43 (+36/−29)

+65 (+63/−46)
K

−38 (+36/−23)
−66 (+41/−19)
+4 (+35/−26)
Na

+44 (+216/−86)
−68 (+84/−23)
−73 (+63/−19)
Ca

+164 (+71/−56)
+11 (+16/−14)
+9 (+15/−14)
Fig. 8. Two of the most interesting geotechnical parameters: dry unit weight (γd) and uni-
axial compressive strength (σc).

Mg
60° towards the SE. The granite lying in the path of the Burata tunnel
likely intruded parallel to this anisotropy, and as a result the weathering
+264 (+865/−256)
front is controlled by this penetrative feature. Additionally, the folia-
−32 (+103/−41)
+1 (+145/−59)

tion determines the geometry of the boundaries between granite and

metasedimentary roof pendants. These surfaces control the infiltra-
Mn

tion of surface water, which is the main agent that promoted in situ
alteration. Such fluids upon reaching the brittly fractured granitic bed-
rock, promoted development of saprock during the early stages of
weathering. The volume increase that occurs when biotite is altered to
+19 (+10/−10)
+28 (+21/−18)
+88 (+78/−55)

vermiculite or smectite induces microfracturing and thus additional

pathways for fluids (Dewandel et al., 2006; Eggler et al., 1969; Ledger
Fe

and Rowe, 1980). In this regard, the importance of the micropore sys-
tem to the formation of secondary products during weathering is
well-known (Calvo de Anta et al., 1983).
During the early of weathering, plagioclase destabilization and the
+20 (+30/−24)
+51 (+40/−32)
+85 (+71/−51)

incipient transformation of potassium feldspar occurred. This alteration

can be recognized in thin sections and in geochemical data. The masses
Al

of K, Si, and Al are increased, and an increase in volatiles is indicated by

increasing LOI values. Though unclear, such increases in K, Si and Al
mass may reflect the incongruent dissolution of biotite within the
metasedimentary rocks and their subsequent illuviation within the
+126 (+52/−43)
+20 (+28/−23)

+71 (+61/−45)

saprolith. Furthermore, if strong oscillations of the groundwater table

occur, then the weathering process is certainly enhanced, as the granit-
Mass balance calculations (%).

Si

oid is subjected to alternating conditions of unsaturation and water

flow that favour the development of saprolite.
The differentiation of two saprolite facies is supported by several
criteria: location, chemical and mineralogical composition, and grain
C-R saprolite
C-P saprolite

size distribution. This structural zonation is not usually clear in shallow

Saprock

residual covers where a mixture of saprolites has been identified (Arel

Table 2

and Tugrul, 2001; Sequeira Braga et al., 2002). Evaluating the mobility
of cations, it is found that both saprolite facies show statistically
 D. Arias et al. / Catena 137 (2016) 572–580 579

significant losses of Ca and Na mass. This is consistent with the dissolu-
tion of plagioclase by acidic solutions and the partial dissolution of apa-
tite. In C-R and C-P facies there is also enrichment in K, which likely was
leached from biotite in overlying roof pendants. However, there are dif-
ferences between the two saprolite facies. The C-R saprolite is enriched
in Si, Al, Fe, K, Mg, Mn and LOI, indicating that fluids entering this sapro-
lite were either carrying these elements as ions or transported particles
rich in them. During transport through the C-R unit, these elements or
particles containing them were adsorbed or deposited. Similar, but
less intense processes operated within the C-P saprolite.
Regarding the location, the C-R saprolite is usually closer to
the saprock and lies beneath the C-P saprolite. This, together with geo-
chemical evidence discussed above, indicates that both facies represent
two different products developed during the progression of the
weathering. This setting suggests that the development of the C-P facies
could be related to the flow of groundwater. Downward-percolating
water would have moved the finer particles of neoformed clay minerals,
resulting in a clay-poor saprolite dominated by quartz and muscovite. In
this context, the underlying C-R saprolite would be the destination of
the eluviated clay particles.
An extended period of time is essential for weathering patterns like
those described above to develop, probably millions of years (Wyns,
1991), and is required for intensive feldspar and biotite hydrolysis and
the formation of new clay minerals (Gi and Hye, 2003; Tardy et al.,
1973). The development of intensely weathered profiles also requires
stable geodynamic conditions (weathering rate widely exceeds the
erosion rate) and alternating cycles of dry and rainy periods. These
conditions occur in NW Iberian Peninsula. Since the Variscan orogeny,
this area has undergone a peneplanation process, resulting in a smooth to-
pography at the present (Pérez Alberti, 2001). Neither the Cenozoic exten-
sion nor Alpine compression had a significant influence on the evolution
of relief. Dating the weathering of the granitoids is very difficult in NW Ibe-
rian Peninsula and even more so for the deep saprolite. Its development in
this area is more likely linked with climate changes that have taken place
in the Cenozoic. In NW Portugal, several shallow saprolites fossilized by
Pliocene deposits were described (e.g., Sequeira Braga et al., 1990).
This research highlights the difficulties in obtaining geotechnical
data from highly weathered materials. Such difficulties include studying
rock masses properties through empirical relationships, and estimating
them from geophysical exploration. Core samples of fresh rock and
saprock can be subjected to laboratory tests (e.g., UCS, PLT), and on
site tests using the Schmidt hammer allows evaluating the weathering
state of the rock mass. In highly weathered saprock, the needle penetra-
tion test (NPT) could be used. Nevertheless, collecting undisturbed sam-
ples of saprolite is complicated, especially in the C-P facies. In such
conditions, the evaluation of geomechanical behaviour should be
based on remoulded samples or the use of field tests, although these
also have operational limitations. SPT is fast, cheap, and versatile in un-
consolidated deposits dominated by sand. Its usefulness here however
was limited by the appearance of isolated pieces of fresh rock, and as a
result, this test overestimated the stiffness of saprolite. In Burata, the
pressuremeter was more useful for determining the strength of the
saprock and the saprolite. However difficulties arose from the instability
of the borehole walls when drilling low-cohesive saprolith.
in the progression of the alteration at depth, which is initially favoured
by the presence of biotite in both the granite and the metasedimentary
roof pendants. From this stage, the evolution from saprock to saprolite is
strongly controlled by the alteration of plagioclase.
The saprock, with a thickness of tens of metres, consists of a rock mass
in which oxides of iron and incipient argillization of feldspars appears. The
saprolite is coarse-grained and preserves original igneous textures, and
two facies were differentiated according to several criteria such as location,
composition, and grain size distribution. Of the two facies, the C-R saprolite
is located deeper and does not exceed 20 m in thickness. It consists essen-
tially of quartz, microcline, muscovite, albite and neoformed clay minerals
(kaolinite and montmorillonite). The C-P facies is shallower, usually occur-
ring above the clay-rich saprolite and along the boundaries with
metasedimentary roof pendants, occasionally reaching 100 m in thickness.
It has a higher percentage of quartz and muscovite (weathering-resistant
minerals), but a lower content of clay minerals probably because they
have been removed by downward percolating water.
From a geotechnical viewpoint, the weathering caused the observed
loss of bulk density and rock strength of the saprock, so the rock quality
is significantly less than that of the fresh rock. Bulk density decreases
even more in the saprolite (N30%), and the mechanical strength is
lowered until reaching values similar to that of soft soil. Furthermore,
the large increase in porosity promotes permeability so that the sapro-
lite behaves as an aquifer, and this is especially well developed in the
case of the C-P facies. These conditions complicated the excavation of
the tunnel and made it necessary to conduct new geological surveys
to investigate the weathering process and its effects on the bedrock.
The main lesson learned from this tunnelling endeavour is the im-
portance of obtaining appropriate knowledge of the geological setting
when weathered basements are investigated. The preliminary surveys
must provide the properties of the materials to be excavated and accu-
rate geological cross sections (i.e., units and contacts between them are
correctly identified). This should prevent future technical issues and
economic losses when the rock masses are tunnelled. Nevertheless,
the first step is to not underestimate the extent of weathering in
terms of depth, as occurred here.
Acknowledgements
The authors are grateful to M. Callizo, C. Prada and E. Castells for their
technical assistance during the sampling campaigns. We also thank E.
Ariño for laboratory testing. The tunnel photo was provided by the en-
gineering and construction company. This research received financial
support from the University of Oviedo Foundation through the project
FUO-EM-012-12. Special thanks to the editor and the anonymous re-
viewers, whose constructive comments and suggestions helped us to
improve the quality of the manuscript.
Appendix A
Chemical composition (%) of the protolith samples (see Table 1).
Sample SiO2 TiO2 Al2O3 Fe2O3 MnO MgO CaO Na2O K2O P2O5 LOI

FR1 71.92 0.24 15.15 1.03 0.01 0.43 0.70 3.38 5.41 0.42 1.32
7. Conclusions FR2 72.37 0.33 14.64 1.46 0.01 0.43 0.81 3.12 5.19 0.33 1.26
FR3 72.75 0.23 14.81 1.03 0.01 0.31 0.62 3.37 5.13 0.36 1.36
FR4 72.80 0.30 14.54 1.36 0.02 0.45 0.68 3.44 4.93 0.31 1.05
This case study of tunnelling in NW Spain enabled the identification
FR5 72.81 0.21 15.00 1.16 0.02 0.38 0.72 3.34 5.04 0.43 1.11
of a strongly weathered granitic basement. Above an irregular and in- FR6 73.03 0.32 14.70 1.36 0.02 0.41 0.58 2.77 5.30 0.31 1.00
dented weathering front that reaches up to 250 m depth, it was possible FR7 73.16 0.32 14.67 1.49 0.19 0.44 0.76 2.80 4.74 0.35 1.09
to clearly distinguish a layered structure of saprock and saprolite.
Weathering in the studied section is primarily controlled by the location
of the boundary between the roof pendant and granitoids. This bound- References
ary constitutes the main access for the supply of groundwater, and cor-
Ague, J.J., 1994. Mass transfer during barrovian metamorphism of pelites, south-central
responds approximately to the orientation of the primary Variscan Connecticut. I: evidence for changes in composition and volume. Am. J. Sci. 294,
foliation. The joint network in the bedrock also has a major influence 989–1057.
580 D. Arias et al. / Catena 137 (2016) 572–580
Arel, E., Tugrul, A., 2001. Weathering and its relation to geomechanical properties of
Cavusbasi granitic rocks in northwestern Turkey. Bull. Eng. Geol. Environ. 60,
123–133.
Aydan, Ö., Ulusay, R., Tokashiki, N., 2014. A new rock mass quality rating system: rock
mass quality rating (RMQR) and its application to the estimation of geomechanical
characteristics of rock masses. Rock Mech. Rock. Eng. 47 (4), 1255–1276.
Bacellar, A.P., Coelho Netto, A.L., Lacerda, W.A., 2005. Controlling factors of gullyings in the
Maracuja Catchment, southeastern Brazil. Earth Surf. Process. Landf. 30, 1369–1385.
Barrera, J.L., Farias, P., González Lodeiro, F., Marquínez, J., Martín Parra, L.M., Martínez
Catalán, J.R., del Olmo, A., de Pablo Maciá, J.G., 1989. Mapa geológico y memoria
explicativa de la hoja 17/27 (Ourense/Verín), escala 1:200.000. 1ª edición, Madrid.
Barton, N., Lien, R., Lunde, J., 1974. Engineering classification of rock masses for the design
of tunnel support. Rock Mech. 6 (4), 189–236.
Begonha, A., Sequeira Braga, M.A., 2002. Weathering of the Oporto granite: geotechnical
and physical properties. Catena 49, 57–76.
Bellido, F., Brandle, J.L., Lasala, M., Reyes, J., 1992. Consideraciones petrológicas y
cronológicas sobre las rocas graníticas Hercínicas de Galicia. Cuad. Lab. Geol. Laxe
17, 241–261.
Bieniawski, Z.T., 1989. Engineering Rock Mass Classifications. Wiley & Sons, New York.
Calvo de Anta, R., García-Rodeja, E., Macías, F., 1983. Mineralogical variability in
weathering microsystems of a granitic outcrop of Galicia (Spain). Catena 10, 225–236.
Capdevila, R., Floor, P., 1970. Les différents types de granites hercyniens et leur distribu-
tion dans le Nord de l'espagne. Bol. Geol. Min. 81 (2–3), 215–225.
Capdevila, R., Corretgé, L.G., Floor, P., 1973. Les granitoides varisques de la Meseta
Ibérique. Bull. Soc. Geol. Fr. 15, 209–228.
Carrasco, T.L., Girty, G.H., 2015. Identifying a reference frame for calculating mass change
during weathering: a review and case study utilizing the C# program assessing ele-
ment immobility and critical ratio methodology. Catena 125, 146–161.
Castells, E., Perea, J.L., Bujalance, F., Varela, F., 2006. Túnel de Formigueiro y túnel de
Burata. Ingeopres 154, 44–49.
Corretgé, L., 1983. Las Rocas Graníticas y Granitoides del Macizo Ibérico. In: Comba, J.A.
(Ed.), Geología de España. Libro Jubilar J.M. Ríos. IGME n°1, pp. 115–128.
Cuesta, A., Gallastegui, G., 2004. Galicia Occidental. In: Vera, J.A. (Ed.), Geología de España.
SGE-IGME, Madrid, pp. 96–100.
Dallmeyer, R.D., Martínez Catalán, J.R., Arenas, R., Gil Ibarguchi, J.L., Gutiérrez Alonso, G.,
Farias, P., Bastida, F., Aller, J., 1997. Diachronous Varisca tectonothermal activity in
NW Iberian Massif: evidence from 40Ar/39Ar dating of regional fabrics. Tectonophysics
277, 307–337.
De Abreu, A., De lima Rocha, M.A., Sozio, L.E., Bilfinger, W., 2007. Rio de Janeiro Subway —
challenges of the access tunnel to Cantagalo Station. Proceedings of The World Tun-
nel Congress 2007 and 33rd ITA/AITES Annual General Assembly. Prague, Austria,
pp. 851–855.
Dewandel, B., Lachassagne, P., Wyns, R., Maréchal, J.C., Krishnamurthy, N.S., 2006. A gen-
eralized 3-D geological and hydrogeological conceptual model of granite aquifers
controlled by single or multiphase weathering. J. Hydrol. 330, 260–284.
Durgin, P.B., 1977. Landslides and the weathering of granitic rocks. Rev. Eng. Geol. 3,
127–131.
Egger, P., Ohnuki, T., Kanoh, Y., 1982. Bau des Nakayama-Tunnels Kampf gegen
Bergwasser und vulkanisches Lockergestein. Rock Mech. (Suppl. 12), 275–293.
Eggler, D.H., Larson, E.E., Bradley, W.C., 1969. Granites, gneisses and the Sherman erosion
surface, southern Laramie Range, Colorado, Wyoming. Am. J. Sci. 267, 510–522.
Galán, G., Pin, C., Duthou, J.L., 1996. Sr–Nd isotopic record of multi-stage interactions be-
tween mantle-derived magmas and crustal components in a collision context. The ul-
tramafic–granitoid association from Vivero (Hercynian belt, NW Spain). Chem. Geol.
131, 67–91.
Gi, Y.J., Hye, B.K., 2003. Mineralogy, chemistry, and formation of oxidized biotite in the
weathering profile of granitic rocks. Am. Mineral. 88, 352–364.
Girty, G.H., Colby, T.A., Rayburn, J.Z., Parizek, J.R., Voyles, E.M., 2013. Biotite-controlled lin-
ear compositional weathering trends in tonalitic to quartz dioritic saprock, Santa
Margarita Ecological Reserve, southern California, USA. Catena 105, 40–51.
Graham, R.C., Rossi, A.M., Hubbert, K.R., 2010. Rock to regolith conversion: producing hos-
pitable substrates for terrestrial ecosystems. GSA Today 20, 4–9.
Grasso, P., Xu, S., Del Fedele, M., Russo, G., Chiriotti, E., 2003. Particular failure mechanisms
of weathered granite observed during construction of Metro Tunnels by TBM.
Proceedings of 29th ITA World Tunnelling Congress. Amsterdam, Netherlands,
pp. 497–503.
Guatteri, G., Koshima, A., Lopes, J.R., Pieroni, M.R., 2013. Summary of the Brazilian experi-
ence on ground improvement with jet grouting for soft ground tunnelling. Proceed-
ings of World Tunnel Congress 2013. Geneva, Switzerland, pp. 2340–2347.
Hazen, A., 1892. Some physical properties of sands and gravel. Mass. State Board of
Health. Ann. Rept., pp. 539–556
Isherwood, D., Street, A., 1976. Biotite-inducted grussification of Boulder Creek granodio-
rite, Boulder County, Colorado. Geol. Soc. Am. Bull. 87, 366–370.
ISRM, 1978. Suggested methods for determining the uniaxial compressive strength
and deformability of rock materials. Int. J. Rock Mech. Min. Sci. Geomech. Abstr. 16,
137–141.
ISRM, 1981. In: Brown, E.T. (Ed.), Rock Characterization Testing and Monitoring ISRM
Suggested Methods. Pergamon Press, London.
Jeong, G.Y., Kim, H.B., 2003. Mineralogy, chemistry and formation of oxidized biotite in
the weathering profile of granitic rocks. Am. Mineral. 88, 352–364.
Lan, H.X., Hu, R.L., Yue, Z.Q., Lee, C.F., Wang, S.J., 2003. Engineering and geological charac-
teristics of granite weathering profiles in south China. J. Asian Earth Sci. 21, 353–364.
Ledger, E.B., Rowe, M.W., 1980. Release of uranium from granitic rocks during in situ
weathering and initial erosion (central Texas). Chem. Geol. 29, 227–248.
López-Fernández, C., Arias, D., Fernández-Viejo, G., Pando, L., 2013. Surface subsidence in-
duced by groundwater drainage tunnelling in granite residual soils (Burata railway
tunnel,Spain). J. Geotech. Geoenviron. Eng. 139 (5), 821–824.
Mahmutoglu, Y., Vardar, M., Kocak, C., Sans, G., 2006. Tunnelling difficulties under squeez-
ing and flowing conditions at the Ayas tunnel. Felsbau 24 (5), 44–50.
Marcos, A., 1973. Las series del Paleozoico inferior y la estructura herciniana del occidente
de Asturias (NW de España). Trab. Geol. 6, 3–107.
Nesbitt, H.W., Markovics, G., 1997. Weathering of granodioritic crust, long-term storage of
elements in weathering profiles, and petrogenesis of siliciclastic sediments. Geochim.
Cosmochim. Acta 61, 1653–1670.
Nesbitt, H.W., Young, G.M., 1982. Early Proterozoic climates and plate motions inferred
from major element chemistry of lutites. Nature 299, 715–717.
Nesbitt, H.W., Young, G.M., 1984. Prediction of some weathering trends of plutonic and
volcanic rocks based on thermodynamic and kinetic considerations. Geochim.
Cosmochim. Acta 48, 1523–1534.
Nesbitt, H.W., Young, G.M., 1989. Formation and diagenesis of weathering profiles. J. Geol.
97, 128–147.
Nettleton, W.D., Flach, K.W., Nelson, R.E., 1970. Pedogenic weathering of tonalite in south-
ern California. Geoderma 4, 387–402.
Ng, C.W.W., Guan, P., Shang, Y.J., 2001. Weathering mechanisms and indices of the igne-
ous rocks of Hong Kong. Q. J. Eng. Geol. Hydrogeol. 34, 133–151.
Oliva, P., Viers, J., Dupré, B., 2003. Chemical weathering in granitic environments. Chem.
Geol. 202, 225–256.
Ollier, C.D., 2010. Very deep weathering and related landslides. Geological Society,
London. Eng. Geol. Spec. Publ. 23, 5–14.
Olona, J., Pulgar, J.A., Fernández-Viejo, G., López-Fernández, C., González-Cortina, J.M.,
2010. Weathering variations in a granitic massif and related geotechnical properties
through seismic and electrical resistivity methods. Near Surf. Geophys. 8, 585–599.
Parizek, J.R., Girty, G.H., 2014. Assessing volumetric strains and mass balance relationships
resulting from biotite-controlled weathering: implications for the isovolumetric
weathering of the Boulder Creek granodiorite, Boulder County, Colorado, USA. Catena
120, 29–45.
Parker, A., 1970. An index of weathering for silicate rocks. Geol. Mag. 107, 501–504.
Pérez Alberti, A., 2001. Análisis geomorfológico y evolución paleoclimática de Galicia
durante el Terciario y el Cuaternario. Cienc. Sociais Humanid 13, 11–66.
Pérez Estaún, A., Martínez Catalán, J.R., Bastida, F., 1991. Crustal thickening and deforma-
tion sequence in the footwall to the suture of the Variscan belt of northwest Spain.
Tectonophysics 191, 243–253.
Power, E.T., Smith, B.J., 1994. A comparative study of deep weathering and weathering
products: case studies from Ireland, Corsica and Southeast Brazil. In: Robinson, D.A.,
Williams, R.B.G. (Eds.), Rock Weathering and Landform Evolution. Wiley, Chichester,
pp. 21–40.
Romero, V.S., Clayton, M., 2006. The role of engineering geology on rehabilitation of the
Likelike Highway Wilson Tunnel, Oahu, Hawaii. Proceedings of 57th Annual Highway
Geology Symposium. Breckenridge, Colorado, pp. 360–377.
Sequeira Braga, M.A., Lopes Nunes, J.E., Paquet, H., Millot, G., 1990. Climatic zonality of
coarse granitic saprolites (“arènes”) in Atlantic Europe from Scandinavia to
Portugal. Sci. Geol. Mem. 85, 99–108.
Sequeira Braga, M.A., Paquet, H., Begonha, A., 2002. Weathering of granites in a temperate
climate (NW Portugal): granitic saprolites and arenization. Catena 49, 41–56.
Sousa, R.L., Einstein, H.H., 2012. Risk analysis during tunnel construction using Bayesian
networks: Porto metro case study. Tunn. Undergr. Space Technol. 27, 86–100.
Taboada, T., García, C., 1999a. Smectite formation produced by weathering in a coarse
granite saprolite in Galicia (NW Spain). Catena 35, 281–290.
Taboada, T., García, C., 1999b. Pseudomorphic transformation of plagioclases during the
weathering of granitic rocks in Galicia (NW Spain). Catena 35, 291–302.
Tardy, Y., Bocquier, G., Paquet, H., Millot, G., 1973. Formation of clay from granite and its
distribution in relation to climate and topography. Geoderma 10, 271–284.
Taylor, G., Eggleton, R.A., 2001. Regolith Geology and Geomorphology. Wiley & Sons,
New York.
von Eynatten, H., Pawlowsky-Glahn, V., Egozcue, J.J., 2002. Understanding perturbation on
the simplex: a simple method to better visualize and interpret compositional data in
ternary diagrams. Math. Geol. 34, 249–257.
White, A.F., Bullen, T.D., Schulz, M.S., Blum, A.E., Huntington, T.G., Peters, N.E., 2001. Differ-
ential rates of feldspar weathering in granitic regoliths. Geochim. Cosmochim. Acta
65, 847–869.
Woronow, A., Love, K.M., 1990. Quantifying and testing differences among means of com-
positional data suites. Math. Geol. 22, 837–852.
Wyns, R., 1991. Use of continental paleosurfaces for thematic probability mapping
(in French). Géol. Fr. 1991 (3), 3–9.