Available online at www.sciencedirect.com
Geomorphology 97 (2008) 190 – 207
www.elsevier.com/locate/geomorph
Combined numerical and geomorphological reconstruction of the
Serra da Estrela plateau icefield, Portugal
Gonçalo Vieira ⁎
Centro de Estudos Geográficos, Universidade de Lisboa, Faculdade de Letras, Alameda da Universidade, 1600-214 Lisboa, Portugal
Received 3 October 2006; received in revised form 5 December 2006; accepted 26 February 2007
Available online 23 June 2007
Abstract
The paper focuses on reconstructing a plateau icefield surface from field geomorphological data and a physical-based glacier
model. The results allow the analysis of the patterns of glacial erosion and the estimation of the palaeo-Equilibrium Line Altitudes
(ELA). The study area is the Serra da Estrela, a plateau in Central Portugal rising to ∼ 2000 m ASL. The glaciated area during the
Last Maximum of the Serra da Estrela Glaciation (LMGSE) was ∼ 66 km2. The reconstruction of the topography of the icefield and
valley glaciers in the LMGSE is based on the Schilling and Hollin model. It iterates along the valley longitudinal profile and is
based on the gradient, on valley-shape indices and on yield basal shear stresses. These variables influence ice thickness and
therefore, also the slope of the ice surface allowing for its reconstruction. The key variable is basal yield shear stress and its value
was included in the model manually starting from the points of maximum extent of the valley glaciers and following a range known
to occur in contemporary conditions. The input values are validated by matching the resulting ice surface to geomorphological
features. Where these are absent, a constant value of 100 kPa was used. The icefield was reconstructed from a radiating set of longsections, along which ice thicknesses were calculated. A DEM of the ice surface was constructed, allowing the estimation of the
hypsometric curves of the distinct glacier catchments and the calculation of the palaeo-ELAs. The results show a regional ELA at
1650 m ASL with spatial variations across the icefield reflecting mainly the effect of eastward snow drift. The LMGSE glaciers
were very sensitive to minor climatic changes, especially due to the large area of the plateau icefield, and to the positioning of the
ELAs, close to, or in the, flat part of the hypsometric curve. The model of the ice surface is of significant value for the analysis of
the patterns of glacial erosion at the landscape level. In the Serra da Estrela most of the glacial erosion occurred near the plateau
margins and in valley heads, where glacier surface slope was steeper allowing for a faster ice flow and where ice flow concentrated.
Strong glacier erosion in the Zêzere valley is linked to the tectonic setting, but also to the confluence of glaciers and to the
overfeeding of snow from the plateau.
© 2007 Published by Elsevier B.V.
Keywords: Plateau icefield landsystem; Glacial erosion; Glacier model; ELA; Balance ratio; Portugal
1. Introduction
⁎ Tel.: +351 217940218, fax: +351 217938690.
E-mail address: [email protected].
0169-555X/$ - see front matter © 2007 Published by Elsevier B.V.
doi:10.1016/j.geomorph.2007.02.042
Equilibrium Line Altitudes (ELA) of glaciers are
particularly sensitive to changes in summer temperature
and winter precipitation (Ohmura et al., 1992; Paterson,
1994; Nesje and Dahl, 2000). The accurate estimation
of the ELA is therefore of major importance for
G. Vieira / Geomorphology 97 (2008) 190–207
palaeoenvironmental reconstruction. Several empirical
studies show a correlation between temperature and
precipitation at the ELA in present-day glaciers and
therefore knowing one of the variables it may be
possible to estimate the other (Lowe and Walker, 1997).
Plateau icefields are glaciers showing topographic
control on ice flow and develop on generally flat surfaces limited by steep slopes. If the former ice cover was
cold-based the geomorphic signature of palaeo-icefields
can be subtle or non-existent (Rea and Evans, 2005).
The lack of glacial geomorphic evidence in plateaux
areas has led to misinterpretations of glacier extension in
some regions and consequently to problems in the
estimation of palaeo-ELAs (e.g. Sissons, 1980). Plateau
icefields have been an issue of new interest in the
geomorphological and glaciological literature and are
treated in detail by Rea and Evans (2005). The development of physical models for ice surface reconstruction enables us to determine ice thicknesses in areas
where geomorphological information is lacking. This
allows analysis of the influence of glacier characteristics
on erosional landforms, offering new insight into the
issue of glacial landscape evolution.
Ice sheets and ice caps react slowly to climate change in
comparison to small mountain glaciers that tend to react
quickly, even to small oscillations. Small mountain glaciers
located near the glaciation limits are therefore good
indicators of regional climate change. In the case of plateau
icefields with the ELA lying on the plateau, the climate
sensitivity becomes even greater. This is due to the fact that
small altitudinal changes in the ELA may result in a
significant modification to the relative sizes of the
accumulation and ablation areas. The consequences may
result in rapid retreat or advance of valley glaciers fed by
the icefield (Sugden and John, 1976). Therefore, it is
important to accurately define the hypsometry of the
plateau icefields and to distinguish them from simple valley
glacier systems in order to model correctly the palaeoELAs and their climatic sensitivity (Rea and Evans, 2005).
Rivera and Casassa (1999) have studied the Pio XI
glacier in the South Patagonian Icefield comparing the
hypsometric curve of the glacier surface with the recent
changes in the ELA. They showed that these changes
occurred in the steepest part of the curve, but that in the
next years the increase in the ELA will start to affect the
plateau area of the glacier, with significant implications
for the mass balance. This type of analysis can only be
performed accurately when the glacier topography is
known in detail. In the case of plateau icefields that lack
geomorphic evidence for ice thickness, the application
of physical-based models for glacier surface reconstruction is the best approach for ELA estimation.
191
In this paper, the plateau icefield of the Serra da Estrela
is reconstructed using the 2D model developed by
Schilling and Hollin (1981) and data from field geomorphological surveying. The application of the model along
individual flow lines enabled the identification of multiple
ice surface profiles that were interpolated according to
topography and geomorphic evidence in the valleys.
Similar approaches were conducted with success by Locke
(1995) in the American Rockies, McDougall (1995) in
Scotland and by Evans et al. (2002) in northern Norway. In
this study a digital elevation model (DEM) of the glacier
surface was obtained allowing for the estimation of the
ELAs in the different catchments. Plotting the ELA over
the hypsometric curve allowed the evaluation of the
climatic sensitivity of the Estrela glaciers.
The modelling approach links geomorphological
observations to physical-based models and is very
robust since it enables result validation in the valley
sections where palaeo-ice thicknesses are known. The
model outputs arise from a linkage between geomorphological and glaciological techniques, with outcomes
useful for both these disciplines, but also for palaeoclimatological purposes. It is worth noting that the ice
surface reconstruction originates from a significant
amount of data arising from geomorphological observations at the landscape level, but that it also provides new
insight into the patterns of erosion inside the plateau
icefield. This is not circular reasoning, since inputs are
solely the external limits of the icefield, while the
outputs are areal, enabling the spatial analysis of the
distribution of landforms that were not inputs for the
model (i.e. comparing the modelled ice thicknesses and
basal shear stresses with the patterns of glacial erosion).
2. Regional setting
The Serra da Estrela is a granite mountain in Central
Portugal showing a plateau that rises to 1993 m ASL
(Fig. 1). The mountain range is part of the Iberian Central
Cordillera and is limited by two steep fault scarps with a
relative relief of over 1000 m. The Serra da Estrela is an
important condensation barrier to the moist air masses
from the Atlantic that enter the Iberian Peninsula from the
west. The upper area shows two plateaus, divided by the
NNE–SSW tectonic lineament of the Zêzere and Alforfa
valleys. The western plateau is the highest at 1400–
1993 m ASL, while on the eastern plateau altitudes stay
below 1750 m.
The general geomorphological characteristics of the
Weichselian glaciation of the Estrela are well-known but a
glacial chronology is lacking. Cabral (1884) was the first
to show the occurrence of glacigenic features in the
192
G. Vieira / Geomorphology 97 (2008) 190–207
Fig. 1. Topography of the Serra da Estrela and location of the sites mentioned in the text.
Estrela, but it was only with Lautensach (1929) that the
glaciation was studied in some detail. Daveau (1971)
revised Lautensach's ideas and presented a more
comprehensive account of the glacial geomorphology of
the mountain. More recently, Vieira (2004) presented a
detailed study on the glacial and periglacial geomorphology of the Estrela, including geomorphological mapping,
sedimentological analysis of glacial and fluvioglacial
deposits and modelling of the glacier surfaces. Unpublished thermoluminescence dates obtained by Vieira and
Woronko from massive fluvioglacial silts from an
intermoraine depression and a fluvioglacial terrace
provided ages of 30 ± 4.5 TL ka and 33.1 ± 5.0 TL ka BP
respectively, indicating that the Last Maximum of the
G. Vieira / Geomorphology 97 (2008) 190–207
Glaciation of the Serra da Estrela (LMGSE) occurred
before the LGM. The TL method itself shows several
limitations for fluvioglacial sediment dating as discussed
by several authors (Gemmel, 1997; Prescott and Robertson,
1997; Preusser, 1999). However, it is worth noting that all
the dates obtained are in chronological agreement with their
geomorphological and stratigraphical setting. These ages
should be considered preliminary and used with care.
In order to try to define accurately the glacial chronology of the area and to compare the results from
different methods, cosmogenic radionuclide dating of
polished outcrops and moraine boulders are under way,
but no results are available yet. Palynological analysis
and radiocarbon ages from organic infills of glacial
basins suggest that small glaciers still existed into the
Late Glacial (Van der Knaap and Van Leeuwen, 1997).
The preliminary and still scarce absolute ages obtained
for the Serra da Estrela agree with the results from other
193
authors for the Iberian Peninsula, that indicate that the
last maximum glacial extent pre-dated the LGM (e.g.
Vidal Romaní et al., 1999; Fernández Mosquera et al.,
2000; García Ruiz et al., 2001, 2003; Jiménez Sánchez
and Farias Arquer, 2002; Pérez Alberti et al., 2004;
Pallàs et al., 2006). Fernández Mosquera and colleagues
(Vidal Romani, oral information) used cosmogenic
dating in the Serra da Queixa (Galicia) and indicate
that there, the maximum glacial extension pre-dated the
Weichselian.
The plateaux of the Serra da Estrela above ∼ 1750 m
are marked by erosional landforms, with areas of glacial
scouring, knock-and-lochan morphology and roches
moutonnées (Fig. 2). Tors and the weathering mantle are
lacking in the highest areas and only occur in small
patches where glacial erosion was limited or where tor
exhumation is post-glacial. Cirques are present mainly
in the eastern side of the plateau. The areas where glacial
Fig. 2. A. Glacial erosion landscape at the Estrela plateau (Salgadeiras area). B. Lateral moraine on the Alforfa valley.
194
G. Vieira / Geomorphology 97 (2008) 190–207
erosion was strongest still show widespread bare rock
outcrops. These were mapped using very detailed digital
orthophotos (1 m cell) followed by a GIS classification
of the percent coverage of bedrock per unit area (Fig. 3).
In the western plateau, areas between 1500 and
1700 m show few traces of glacial erosion, moraines and,
in some areas, remnants of weathering mantle. The
valleys show typical geomorphological features of the
glaciated valley landsystem, with cirques in the upstream
sections, u-shaped cross-sections, overdeepenings
(above ∼1200 m ASL), different types of moraines
and kame terraces (Fig. 2). In the Estrela valleys, features
classified as kame terraces are bouldery debris-flow
deposits that were transported along gullies or small
valleys and that accumulated against the former glacier
margin, and that are currently hanging on the valley side.
The glaciated area during the LMGSE was ∼66 km2
and is well-defined by latero-frontal moraines in the valleys
and ice-marginal moraine complexes (Fig. 3). These
features were mapped at a scale 1:15 000. Most of the
glaciation occurred in the western plateau, between Alto da
Torre and Fraga das Penas, where a plateau icefield with
several radiating valley glaciers was present. The largest
glaciers were: Zêzere, Alforfa, Loriga, Covão Grande and
Covão do Urso (Fig. 3). On the eastern plateau, only small
glaciers occurred and geomorphological evidence of
glaciation is lacking. Nevertheless, recent micromorphological analysis of a diamicton in an area without typical
glacial landforms, suggests that the diamicton is glacigenic
indicating the occurrence of an icefield on the eastern
plateau during an older glacial episode, probably of pre- or
early-Weichselian age (Vieira, 2004).
The glacier limits used in this paper correspond to the
LMGSE limits as reconstructed from geomorphological
observations. These limits were defined using the best
preserved external and highest moraines, which still
show a good degree of preservation. Drift and moraine
boulders occur in some sites outside the LMGSE limits,
but these are interpreted as belonging to older glacial
events. Along the valleys, smaller moraines showing a
better degree of conservation occur, but these are
interpreted as dating from stadial conditions during the
deglaciation. Due to the lack of absolute ages it is still
impossible to know for sure if the limits of the LMGSE
icefield are synchronous.
3. Methods
3.1. Plateau icefield reconstruction
The widespread glacial features in the valleys allow a
relatively straightforward reconstruction of the valley
glacier topography, especially associated with the highest
lateral moraines (Fig. 3). However, above this limit
(maximum elevation of lateral moraines — MELM) and
especially on the plateau, due to the absence of nunataks
(they only occur near the plateau edges), there is no
geomorphological evidence allowing the estimation of ice
thickness and glacier topography. The only way to
estimate the ice thickness on the plateau is through the
application of physically based models.
The method developed by Schilling and Hollin
(1981) adapts Nye's (1952) equation to valley glaciers
and calculates the altitude of the glacier surface (hi) by
including a valley-shape index (ci) that influences basal
yield shear stress (τav) and is applied in an iterative
procedure starting from the glacier front and moving upvalley at constant distance steps (▵x) (Eq. (1)). The
equation also includes ice thickness at the central flowline (ti), ice-density (ρ) and gravity (g).
hiþ1 ¼ hi þ sav =ci qg Dx=ti
ð1Þ
This equation is based on the assumption of perfect
plastic deformation of ice, while in reality the rheology
of glaciers is dominated by non-linear viscous flow.
Basal sliding is also not accounted for. However, this
simple model provides very good results for reconstructed glacier surfaces. It is especially useful where
palaeoclimatic conditions at the time of the glaciation
are unknown, since it contributes for palaeo-ELA
estimations and, as shown by Locke (1995), it has
been used successfully for the reconstruction of several
mountain glacier surfaces.
For the plateau icefield the model can be applied by
changing the value of the valley-shape index as
explained below. The model iterates up longitudinal
valley-floor profiles (Fig. 4A) and the glacier surface is
obtained from inputs of the bedrock gradient along the
profile, of valley-shape indices and of yield basal shear
stresses.
In valley glaciers the basal shear stress acting on
the valley floor is lower than for wide glaciers, because
the valley sides support some of the mass reducing
normal stress (Paterson, 1994). The valley-shape index
(c) accounts for this effect and is calculated from crosssections along the valley taken from DEM data (Eq. (2)).
The DEM was derived from a digital map at the base scale
of 1:25 000 with contours at 10 m intervals. In Eq. (2) A is
the area of the cross-section of the glacier, P is the
perimeter of the bed section and t is the ice thickness in the
valley centre. A spreadsheet was programmed to measure
A and P from data taken from the DEM at 10 m horizontal
intervals along the cross-sections. The glacier surface was
G. Vieira / Geomorphology 97 (2008) 190–207
Fig. 3. Geomorphological survey of the Serra da Estrela. A. Bare rock outcrops per unit area inside the glacier limits at the LMGSE (cell size 10 000 m2, values in legend in percent area). B. Moraines.
A. Alforfa, AL. Alvoco, C. Caniça (C. Grande), L. Loriga, LS. Lagoa Seca, NSA. Nave Santo António, U. Covão do Urso, VC. Vale do Conde, Z. Zêzere. The scarce moraine cover mapped in the
Zêzere valley is due to the reworking of the boulders within the talus slope debris. Contour interval 50 m.
195
196
G. Vieira / Geomorphology 97 (2008) 190–207
Fig. 4. A. Location of the long-sections used for modelling the plateau icefield. B. Location of the cross-sections used for calculating the valley-shape indices.
G. Vieira / Geomorphology 97 (2008) 190–207
considered to be flat along the cross-section with an
altitude equal to the altitude of the slope in the ice margin
contact. Due to the absence of sediment thickness data,
sedimentary deposits were not accounted for and the
present-day topography was used.
Hooke (1998, p. 51) discusses in detail the formula
for valley-shape index calculation. The valley-shape
index is 1 for an infinitely wide glacier (i.e. an ice-cap),
0.5 for a semi-circular shaped channel and 0.4 for glacial
cirques (Schilling and Hollin, 1981). This index was
originally developed for describing parabolic crosssections with values between 0.5 and 0.6. V-shaped
valleys have lower values, while u-shaped valleys with
flat floors, have higher values (Menzies, 1995).
A=P ¼ ct
ð2Þ
The bed gradient along the valley centre and the valleyshape index are variables obtained directly from the
topographic map. ▵x is 10 m and the c along the valleys
were calculated by linear interpolation of c-values
obtained in representative cross-sections (Fig. 4B). This
option was followed because the valley shape varies in a
somewhat regular way along the valley. For the plateau,
values of c = 1 were chosen. Ice thickness, a variable
needed for the calculation of c in the valleys was obtained
from the altitude of lateral moraines, kame terraces,
glacial trimlines and erosional steps in the valley slopes
aligned with other glacial evidence. In the valley sections
where glacial evidence was absent, ice thicknesses were
estimated from geomorphological features located in upvalley and down-valley sites.
Basal yield shear stress is the critical variable in the
model and it needs to be chosen by the modeller for each
distance step. The values to be chosen are based on the
range of measurements known to occur in contemporary
valley glaciers. These vary between 50 and 150 kPa with
most valley glaciers showing values close to 100 kPa
(Schilling and Hollin, 1981). The later value is a good
approach when modelling ice following perfect plasticity conditions (Paterson, 1994). However, a value of
100 kPa is too high for ice sheets and the values that have
been obtained vary from 0 to 100 kPa, with an average of
50 kPa. Evans et al. (2002) have used 100 kPa for plateau
icefields in Scandinavia and Pierce (1979) used
successfully values from 60 to 150 kPa for the Pinedale
icefield in Yellowstone. The values to be used have to be
checked against geomorphological evidence in order to
evaluate model consistence and to provide validation
(Bennett and Glasser, 1996).
In this study the glacier profiles were calculated
using the spreadsheet developed by Locke (1995) with
197
adaptations allowing for a better iterative estimation
of the ice surface and choice of basal shear stress values.
A starting value of 100 kPa was used for the whole
glacier. This value was then changed along the valleys
wherever geomorphological evidence providing ice
thickness estimation was present (i.e. moraines, kame
terraces or erosional steps; the latter being used when
occurring at heights correlative to and at short distance
from LMGSE lateral moraines). The location of those
geomorphological features was plotted in the longsection facilitating the iterative choice of the best values
of τav. The values of τav providing the best-fit between
the modelled glacier surface and the glacigenic features
were chosen.
In most of the cases, values of τav between 50 and
150 kPa were used (Table 2). Where geomorphological
features indicating ice thickness were lacking, such as
the plateau sections, a constant value of 100 kPa was
chosen. A similar number was used by Evans et al.
(2002). McDougall (1995) used a simpler modelling
approach based on the distance to the plateau margin
using a basal shear stress value of 50 kPa.
The variability of the basal shear stress along the
valley glaciers depends on several factors (Schilling and
Hollin, 1981): (a) poor estimation of the post-glacial
sediment volume in the valleys; (b) changes in the rates
of accumulation and ice flow velocity along the glacier;
(c) heterogeneous glacier temperatures; and (d) the
occurrence of basal melting and glacial surging.
Research in the last decades has shown the importance
of deformational beds in glacier dynamics (e.g. Paterson,
1994; Benn and Evans, 1998) and that a high plasticity of
the bed can also be the cause very low basal shear
stresses. All these factors have to be accounted for when
interpreting the modelling results, but it is rarely possible
to identify safely the cause for an abnormal basal shear
stress value.
The calculations were conducted along longitudinal
profiles in the glaciated valleys representing the centreflow lines (Fig. 4A). Ice thickness reconstruction started
from the glacier snouts and the iteration proceeded upvalley at 10 m steps. The same was done along
secondary flow lines interpreted from topography and
glacier erosion features (e.g. striae and grooves). The
whole icefield was reconstructed in this way from a
radiating set of long-sections, along which ice thicknesses were calculated. The data were integrated in a
GIS package and a manual interpolation of the ice
thicknesses was conducted, in order to obtain the glacier
topography. From this, a DEM of the ice surface with a
cell size of 100 m2 was calculated and also an ice
thickness map.
198
G. Vieira / Geomorphology 97 (2008) 190–207
Table 1
Shape index measures for the Serra da Estrela valleys
Maximum
Minimum
Alforfa
Alvoco
Candeeira
Covais
CG Caniça
CG Nave Travessa
C. Urso
Estrela
Loriga
Zêzere
0.60
0.55
0.55
0.55
0.50
0.50
0.60
0.57
0.52
0.46
0.51
0.45
0.57
0.53
0.61
0.52
0.51
0.41
0.58
0.48
3.2. Equilibrium Line Altitudes
The DEM allowed the estimation of the hypsometric curves of the different glacier catchments and
the calculation of the palaeo-Equilibrium Line Altitudes (ELA). Nesje and Dahl (1992, 2000) distinguish
the ELA-TP and the ELA-TPW, where T is summer
temperature, P is winter precipitation and W is the
wind induced snow drift. The ELA-TP is mainly
controlled by winter precipitation and temperature and
ELA-TPW is also controlled by wind redistribution of
snow. These concepts are especially interesting for the
Portuguese mountains because during the Weichselian
snow drift has been implicated as one of the main
factors inducing glacier asymmetry (Lautensach,
1929; Daveau, 1971; Ferreira et al., 1999; Vieira,
2004). The approach that is followed in the present
paper allows the estimation of the ELA-TPWs when
individual glacier catchments are considered, thus
contributing to the identification of the local controls
on ELA. The ELA-TP, or regional ELA, is obtained
by averaging the ELA for all quadrants (Nesje and
Dahl, 1992).
Three methods of ELA calculation were used: (a) the
maximum altitude of lateral moraines (MELM-ELA);
(b) the accumulation area ratio (AAR-ELA); and (c) the
balance ratio (BR-ELA).
The MELM-ELA was used because it has been
applied previously by other authors in the Serra da
Estrela (Lautensach, 1929; Daveau, 1971) and is also a
method frequently used in glacial geomorphology
papers on the Iberian Peninsula glaciations. Therefore
it is useful for regional comparison, but also for
comparison with the results obtained from the AARELA and BR-ELA. This is a geomorphic-based method
and its theoretical basis is the directions of debris
movement above and below the ELA. Above the ELA
the ice flow is towards the centre and bottom of the
glacier, whilst below the ELA it is towards the ice
margin and upwards towards the surface (Hooke and Le,
1998). Therefore, lateral moraines start to form at the
ELA and that is the Maximum Elevation of the Lateral
Moraines (Nesje, 1992; Nesje and Dahl, 1992). The
main problems are to know if in fact the lateral moraine
formed immediately below the ELA (there are several
variables that affect this, e.g. debris availability) and if
the original shapes and positions of the moraines are
preserved. As a consequence, the MELM-ELA provides
only a minimum value for the ELA. Meierding (1982)
compared several methods for ELA estimation in the
Front Range (Colorado) and found out that MELMELA was the one providing poorest results.
The values for the AAR-ELA in steady-state glaciers
of the mid and high latitudes generally range from 0.5 to
0.8 (Benn and Evans, 1998), but the typical values lie
between 0.55 and 0.65, which is the range more
frequently used (i.e. Nesje, 1992; Nesje and Dahl,
1992; Porter, 2001; Munroe and Mickelson, 2002).
However, for plateau icefields and piedmont glaciers
AAR-ELA can be very variable due to the special
characteristics of the hypsometric curves of these
glaciers (Nesje and Dahl, 1992). McDougall (1995)
emphasises the problems of using AAR-ELA for ELA
estimation of plateau icefields. Debris-covered glaciers
show very low values, with 0.15 indicated for Sierra
Nevada (USA) as indicated by Clark et al. (1994). In this
study we used a value of 0.6 for the AAR-ELA.
The balance ratio method was defined by Osmaston
in 1965 in order to overcome the problems induced by
glaciers with asymmetric hypsometric curves (Osmaston, 2005) and later developed by Furbish and Andrews
(1984). The method accounts for the glacier hypsometry, but also for the shape of the mass balance curve
(Benn and Gemmell, 1997). This method assumes that
for steady-state conditions, the annual accumulation
above the ELA balances the ablation below it and is
explained in detail by Benn and Gemmell (1997). The
authors provide a spreadsheet allowing for the BR-ELA
calculation. According to Osmaston (2005) Benn and
Gemmel's spreadsheet tends to underestimate the BRELA. B. Rea (per. comm.) tested this assumption and
found that this is only significant for glaciers with very
skewed hypsometry, which is not the case in Serra da
Estrela. A balance ratio of 2.0 was used. This value is
taken as representative of maritime glaciers of the mid
latitudes, as the present-day glaciers of Alaska and of the
Cascades (Benn and Gemmel, 1997; Benn and Evans,
1998). Lukas (2005) used a similar ratio for the
reconstruction of the ELA of an ice-cap in the NW
Highlands (Scotland).
G. Vieira / Geomorphology 97 (2008) 190–207
4. Results
4.1. Reconstruction of the glacier surface
Shape indices calculated for 29 typical cross-sections
in the Serra da Estrela valleys showed relatively
199
homogeneous values between 0.41 and 0.61 (Table 1).
This may be related to all the glaciated sections of the
valleys being cut in the same lithology, which is granite.
In general, upstream cross-sections show highest values
(more u-shaped) and downstream sections show the
lowest values (more v-shaped). This reflects the higher
Fig. 5. Modelled longitudinal profiles of the glaciers as shown in the spreadsheet, with triangles indicating the position of moraines and diamonds
showing kame terraces and erosional steps. The input values used for the model are represented in the lower graphs. A. Zêzere glacier, B. Covão
Grande (Nave Travessa) glacier.
200
G. Vieira / Geomorphology 97 (2008) 190–207
glacial erosion and longer permanence of the glaciers in
the upstream parts of the valleys.
User defined values of basal yield shear stress (τav)
were used to fit the modelled ice thickness to geomorphological evidence along the valleys. Values between 50
and 150 kPa showed a good agreement with the
geomorphological evidence in the majority of the valleys
of the Estrela (Fig. 5). The Zêzere glacier shown in the
figure is not the one providing the best modelling results,
but is the longest glacier of the Serra da Estrela. The
modelled surface shows an error of several meters between
1 and 3 km from the snout, especially when comparing to
the height of the geomorphological evidence. This error
could be easily solved by introducing changes in basal
shear stresses, but the option was to use a regular value
along this section of the valley. The difficulty in the
modelling may be related to the poor positioning of the
glacier front, since there is a lack of geomorphic evidence,
being possible that at the stage of formation of the lateral
moraine, the glacier front was situated a couple of hundred
metres upstream from the site indicated in the profile. On
the other hand, the overestimation of the ice thickness
between ∼5.5 and 7 km could be related to the inflow of
ice from the tributary glacier from the Candeeira valley.
Values of basal shear stress exceeding 150 kPa were
only found for the Covão Grande (Nave Travessa)
glacier with 170 kPa along 680 m of the long-section
(Fig. 5B). This is an atypical value compared to the
other Estrela glaciers and the need for its use originates
from the strong longitudinal gradient of the lateral
moraine, especially when compared to the weak
gradient of the valley floor. This suggests that there
was an overflow from the south (Lagoa Comprida) as
supported by the orientation of the roches moutonnées.
It is therefore possible that the main flow direction in
this section of the icefield during the deposition of the
lateral moraine was not along the valley axis, but in a
more transverse direction, towards the northwest. Such a
flow is not possible to model accurately using the
present type of approach. At the ice-divide the overall
coincidence of the ice thickness values between
different catchments indicates that if there is an error,
it is of minor significance.
Values of basal shear stress as low as 50–60 kPa were
found in several of the Estrela glaciers (Table 2). The
lowest values tend to occur near the glacier snout, a fact
that may be related to warmer ice in the ablation area
and to a thicker sedimentary cover. The very low values
in the Covão Grande (Nave Travessa) were used for
fitting the ice surface with the moraines in the
confluence with the glacier of Covão Grande (Caniça)
and they may be related with 2D modelling limitations
originating from the flow conditions in the confluence
zone.
The reconstruction of the topography of the icefield
and valley glaciers of the Serra da Estrela is in very good
agreement with the geomorphological observations and
shows that the model is of good value to the Serra da
Estrela. The glaciated area was of 66.2 km2 with the
largest catchment being that of the Zêzere glacier with
23 km2 (Figs. 6, 7, Table 3). The maximum ice thickness
was of 344 m in the Zêzere valley (Fig. 6B), a value that
is confirmed by lateral moraines. Other valleys showed
significant ice thicknesses, with 239 m in the Alforfa
and 217 m in Covão do Urso. On the plateau, ice
thicknesses were of ∼80–100 m and the icefield summit
altitude was ∼2090 m.
Ice volumes emphasize the significance of the Zêzere
glacier with 2.75 × 109 m3, a value that corresponds to
43% of the total volume and more than double the volume
in the second largest catchment (Covão Grande). The
longest glacier was also the Zêzere glacier at 11.3 km.
4.2. Equilibrium Line Altitudes
Lautensach (1929) and Daveau (1971) estimated the
MELM-ELA for the Serra da Estrela and obtained
values of ∼ 1650 m. Lautensach indicated that in the
eastern part of the mountain the MELM-ELA was at
1620 m and that in western part it was of 1650 m and
attributes this difference to snow drift. Daveau
suggested that the MELM-ELA might have been some
tens of meters lower in the northern part of the mountain
due to insolation differences. The values obtained by
these authors are in good agreement with the ones we
calculated using the AAR-ELA (1650 m) and BR-ELA
Table 2
Basal shear stress values (kPa) used for the modelling of the glacier surfaces
Average
Median
Maximum
Minimum
a
Alforfa
Alvoco
Candeeira
Covais
CG Caniça
CG Nave Travessa
Covão Urso
Estrela
Loriga
Zêzere
123
130
150
90
91
100
100
70
100
100
100
100
90
80
120
60
109
100
150
100
101
100
170 a
50
96
100
120
50
92
100
100
80
110
100
150
100
103
100
120
80
For a distance of 680 m.
G. Vieira / Geomorphology 97 (2008) 190–207
Fig. 6. A. Reconstruction of the surface contours of the Serra da Estrela plateau icefield and valley glaciers. Ice divides and flow directions represented. Contour interval 50 m. (A—Alforfa glacier, AT —
Alto da Torre, CG — Covão Grande glacier, CU — Covão do Urso glacier, E — Estrela glacier, L — Loriga glacier, Z — Zêzere glacier). B. Ice thicknesses of the Serra da Estrela plateau icefield and
valley glaciers.
201
202
G. Vieira / Geomorphology 97 (2008) 190–207
Fig. 7. Perspective of the modelled Serra da Estrela plateau icefield and valley glaciers. View from the South, with 50 m interval contours represented
over the glacier surface.
(1643 m) (Table 4). The regional ELA was not estimated
here using the MELM-ELA, but if we do not account for
the glaciers with topographical constraints induced by a
wide plateau area and narrow valleys (e.g. Loriga
glacier), the values of the MELM-ELA vary between
1580 m (Zêzere glacier) and 1700 m (Estrela glacier).
As will be shown below, this method does not provide
good results for local estimation of ELA differences
within a plateau icefield glaciated area.
The results obtained with the AAR-ELA method
show a general spatial pattern similar to the one obtained
with the MELM-ELA (Table 4). A major difference is
present between the S and SW facing glaciers with
higher AAR-ELAs (N 1700 m ASL), the NW facing
glaciers (1625 to 1670 m ASL) and the NE and SSE
facing glaciers with the lowest AAR-ELAs (1570 to
1590 m ASL). The values are exceptionally high for the
Loriga glacier with an AAR-ELA at 1830 m ASL, but
this seems to be related with the peculiar hypsometric
curve of this glacier, with a very wide area lying in the
plateau and therefore inducing a large error in the AARELA method (Fig. 8).
Table 3
Spatial measures of the glacier catchments of the Serra da Estrela at the LMGSE
Área (km2)
Área (%)
Max. ice
thickness. (m)
Volume (m3)
Volume (%)
Length (km)
Max. altitude
(m ASL)
Min. altitude
(m ASL)
Alforfa
Alvoco
Covais
Covão
Grande
Covão
do Urso
Estrela
Loriga
Conde
valley
Zêzere
Others
Total
5.2
7.9
239
1.3
2.0
96
0.1
0.2
67
13.4
20.2
191
8.9
13.4
217
1
1.5
93
8.1
12.2
180
3.1
4.7
158
23
34.7
344
2.1
3.2
–
66.2
4.7 × 108
7.3
5.8
2090
6.78 × 107
1.1
2.5
2090
3.93 × 106
0.1
0.6
1360
1.21 × 109
18.9
6.2
1970
8.84 × 108
13.8
7.2
1970
3.86 × 107
0.6
2.5
2090
6.38 × 108
9.9
6.7
2090
2.48 × 108
3.9
1.9
1830
2.75 × 109
42.9
11.3
2090
1.04 × 108 6.42 × 109
1.6
–
–
880
1400
1110
980
1040
1290
800
1590
750
–
G. Vieira / Geomorphology 97 (2008) 190–207
203
Table 4
Modeled values of the Equilibrium Line Altitudes for the Serra da Estrela glaciers during the LMGSE in meters above sea-level
MELM
AAR
BR
Alforfa
Alvoco
Covais
C. Grande
C. Urso
Estrela
Loriga
V. Conde
Zêzere
Average
1565
1571
1588
–
1913
1872
1330
1278
1290
1640
1667
1668
1640
1625
1617
1700
1700
1704
1450
1831
1651
–
1701
1721
1580
1589
1586
–
1650
1643
MELM — maximum elevation of lateral moraines, AAR — accumulation area ratio, BR — balance ratio.
The application of the BR-ELA method shows the
more homogeneous values and seems to show the
problems related to the plateau topography that
appeared in the MELM-ELA and AAR-ELA methods
in the Loriga glacier. In the western side of the Estrela
BR-ELA values range from ∼ 1615 to 1670 m ASL and
in the eastern side, between ∼ 1585 to 1590 m ASL. The
Loriga glacier provides a BR-ELA of ∼ 1650 m that is in
agreement with the other glaciers showing a similar
aspect. The southern glaciers of Estrela and Alvoco
show higher BR-ELAs with, respectively, 1705 and
1870 m ASL.
The low values of the ELA estimated for the Covais
glacier (1270 to 1320 m ASL) are linked to local
controls on the ELA-TPW, since this glacier is in a
shadow area favouring ice maintenance. On the other
Fig. 8. Hypsometric curves of the modeled glaciers of the Serra da Estrela during the LMGSE. Note the location of the BR-ELAs.
204
G. Vieira / Geomorphology 97 (2008) 190–207
hand, ELA values for the Conde glacier (1700 to 1720 m
ASL) seem to be overestimated.
5. Discussion
The application of the Schilling and Hollin (1981)
model to the Serra da Estrela allowed the reconstruction
of the plateau icefield from quantitative data. The ice
thickness map (Fig. 6B) is an important tool for
analyzing the distribution of the bare rock surfaces that
coincide with areas of high glacial erosion (Fig. 3A).
There is a coincidence of these areas with the locations
where bedrock gradients and ice thicknesses were
greater. The former induces a steeper gradient of the
ice surface that together with higher values of ice
thickness, generate higher basal shear stresses, promoting increased erosion. The linear regression between the
percent of bare rock area and the ice thickness at each cell
measured in GIS environment results in a R2 of 0.63
(p b 0.00001). The calculation procedure was done by
excluding the cells where bare rock outcrops are absent.
This removed from the analysis all the sites covered by
deposits that would bias the results. The map of bare rock
surfaces shows that most of the glacial erosion occurred
near the plateau margins and in valley heads, with
smaller values near the ice-divides where basal sliding
was smaller or lacking (Fig. 3A).
A comparison of the present-day topography with the
ice thickness map indicates the importance of the valleys
for concentrating the ice flowing from the plateau
icefield. The best example occurs in the Zêzere valley,
where the confluence of the Zêzere and Candeeira
glaciers, together with a brecciated fault-zone setting,
gave origin to a well-developed u-shaped valley
(Figs. 1, 7).
The main application of the glacier surface DEM is to
the estimation of palaeo-ELAs, since it allows for the
reconstruction of the hypsometric curves of the glacier
catchments. From the three ELA estimation methods,
the one showing the more spatially homogeneous results
was the BR-ELA method. This is in agreement with
observations by Benn and Gemmell (1997) that
highlighted the value of this method, which seems
especially useful for plateau icefields. The method
proved of great value in the estimation of the ELA for
the Loriga glacier, whose hypsometric curve showed a
wide plateau icefield area flowing into a narrow and
steep valley glacier. In addition, the values obtained for
the regional ELA are in agreement with ones calculated
by other authors using the MELM method that lie at
∼ 1650 m ASL. This suggests that the lateral moraines
suffered little modification following deglaciation, a fact
that is largely controlled by the gently sloping character
of the plateau.
The BR-ELA method allowed for a more detailed
insight into the local controls on glacier development
and added new data to previous studies. The W–E
asymmetry with BR-ELAs lying some 80 m lower in the
eastern part of the mountain seems to be linked with
snow drift. However, this difference does not appear to
explain the much larger extent of the Zêzere valley
glacier when compared to the other glaciers of Estrela.
The wide area of the catchment lying above the BRELA and the fact that the Zêzere glacier was nourished
also by the Candeeira and Covões glaciers seem to be
the controlling factors on this asymmetry. This supports
the importance of pre-glacial topography for the glacier
development.
The N–S asymmetry suggested by Daveau (1971)
and based on the glacier size and minimum altitudes is
not obvious for the larger glaciers. There are no
differences between the BR-ELAs of the Zêzere and
Alforfa glaciers, and in the NW there is only a 50 m
difference between the Covão do Urso and Covão
Grande glaciers which may be explained by insolation
control. However, this is a very small difference when
considering the intrinsic errors of the methodological
approach (i.e. derived from the lack of absolute datings,
inducing limitations in the definition of a synchronous
ice margin, or from cumulative errors derived from the
iterative procedure of ice thickness estimation). Only on
the small glaciers of Alvoco and Estrela is an insolationdriven asymmetry apparent, since they show very high
BR-ELAs (1872 and 1704 m ASL). However, geomorphological research conducted in the Alvoco valley did
not provide clear evidence for the maximum extent of
the valley glacier, since moraines are lacking in most of
the valley (Vieira, 2004). A poor estimation of the
position of the glacier front may explain the excessively
high BR-ELA. With respect to the Estrela glacier, the
high BR-ELA may be controlled by local topography
related to the very small and wind exposed catchment
area.
The most important contribution from the approach
presented in this paper is the reconstruction of the
hypsometric curves of the plateau icefield and valley
glaciers during the LMGSE. The plot of the BR-ELAs
on the hypsometric curves (Fig. 8) reveals important
information about the climatic sensitivity of the Serra da
Estrela glaciers. Some of the hypsometric curves show a
wide low-gradient sector that corresponds to most of the
plateau icefield and to the highest part of the glacier.
That sector is delimited by a sharp change into a steep
slope at ∼ 1600–1800 m ASL. This type of curve
G. Vieira / Geomorphology 97 (2008) 190–207
illustrates the climatic sensitivity of the Estrela glaciers
and of plateau icefields in general. If the ELA is at (or
near) the flat area of the curve, then a minor climate
change inducing a small altitudinal modification in the
ELA, would induce a significant movement in the ELA
along the curve, due to marked changes in the
accumulation and ablation areas. In the Serra da Estrela,
this situation is especially remarkable with striking
examples in the Loriga or Vale do Conde glaciers.
During climate deterioration, glaciers would have
formed quickly on the plateaux once the firn line
reached the plateau edge, and in periods of warming (or
increasing aridity), altitudinal increases in the ELA
would have led to retreat of the valley glaciers in
response to a rising ELA. This has probably occurred
during the LGM, when conditions in southwest Europe
were drier than before, leading to glacier retreat in the
valleys (Peyron et al., 1998; Florineth and Schlüchter,
2000; Reille et al., 2000; Muñoz Sobrino et al., 2001,
2004). As shown above, preliminary datings point to the
same behaviour in the Serra da Estrela (Vieira, 2004).
6. Conclusions
The application of the Schilling and Hollin model to
the reconstruction of the Serra da Estrela icefield and
valley glaciers provided very interesting results. A
significant part of the data entering in the model derives
from a classic geomorphological field survey and
mapping. This linkage between field-based landsystem
analysis and physical modelling gives origin to robust
modelling results contributing to bridge-the-gap between pure terrain observations and computer-based
approaches. It also allows the computation of ice
thicknesses in areas where geomorphological information is lacking, contributing therefore to the genetic
interpretation of the landforms. Furthermore, the
methodological approach culminates in significant
results for the estimation of the palaeo-ELAs, strengthening the linkage between geomorphological, glaciological and palaeoclimatological methods.
The model is based on simple plastic deformation
and is a 2D model adapted for a 3D glacier surface
reconstruction. It has, therefore theoretical constraints,
but shows the advantages of: (a) being of simple and
straightforward to use; (b) being a transparent model
where changes can be easily introduced; (c) being
constrained by geomorphological evidence in the
valleys and therefore directly validated in those sections
of the profiles; and (d) providing good results, as shown
by other authors that modelled present-day glacier
surfaces.
205
The main objectives of the application of the model
are estimating the ice thickness in areas without
geomorphological information and obtaining the hypsometric curves of the glacier surfaces for ELA calculations. The data integration in a GIS allowed computation
of various size parameters from the different glacier
catchments, and also 3D visualization of the reconstructed glaciers, a tool showing strong pedagogical
value, both for teaching and outreach activities.
The computation of the ice thickness for the icefield
allowed the evaluation of its influence on the distribution of the areas of maximum glacial erosion, and it was
shown that these tend to occur where ice thicknesses
were larger, and in the valley heads and plateau margins,
where bedrock shows steeper gradients.
From the three methods of ELA calculation, the
balance ratio method is the one that provides the results
that are most compatible with the spatial distribution of
glaciers and seems to provide a better explanation of the
complex topographical control induced by the plateau
topography. This became clearer in the case of the
Loriga glacier that showed the following ELAs according to method of calculation: (a) 1450 m (MELM-ELA);
(b) 1830 m (AAR-ELA); and (c) 1650 m (BR-ELA).
The reconstructed regional ELA is of ∼ 1650 m ASL
at the LMGSE. Local ELA's (ELA-TPW) show the
influence of the mountain topography on glacier extent.
The values support a W–E asymmetry, with lower ELAs
in the eastern side of the mountain range, a fact that is
connected to prevailing westerly winds. A weak S–N
asymmetry seems also to occur. The Zêzere and Alforfa
valleys show no difference in the ELA's. However,
there is a small altitudinal difference in the ELAs of the
Covão Grande and Covão do Urso glaciers in the
northwest. The small south facing Estrela glacier may be
controlled by the high insolation, but this still remains to
be proven.
The hypsometric curves of the different catchments
showed that the glaciers of the Serra da Estrela were
very sensitive to minor climatic changes. This behaviour
is due to the large size of the plateau icefield in
comparison to the total glacier area, and also to the fact
that the ELAs were located on the plateau or the slopes
close to the plateau margin. A small change in the ELA
would have affected significantly the accumulation and
ablation areas. Thus, the plateaux of the Serra da Estrela,
with altitudes of 1600–1900 m ASL must have been
very sensitive to conditions of glacial inception and
deglaciation. The Serra da Estrela is therefore a key area
for the assessment of the regional palaeoenvironmental
conditions on the western margin of Iberia during the
Pleistocene.
206
G. Vieira / Geomorphology 97 (2008) 190–207
Acknowledgements
This work has been partly funded by the project
ESTRELA (POCTI/CTA/11153/1998). The author
thanks: the Natural Park of the Serra da Estrela for the
logistical support in field work; John Hollin and William
Locke for bibliographical support and discussing details
on running the model; and Sven Lukas, Darek McDougall, Atle Nesje and Brice Rea for kindly providing
bibliographical support. Models for ice surface reconstruction were based on an adaptation of a spreadsheet
model provided by W. Locke. Balance ratio ELA
calculations were done using the spreadsheet model by
Benn and Gemmel (1997). Brice Rea and Juan Ramón
Vidal-Romaní provided thorough reviews of the manuscript that contributed to improve its quality. I am thankful
for their comments.
References
Benn, D.I., Gemmell, A.M.D., 1997. Calculating equilibrium-line
altitudes of former glaciers by the balance ratio method: a new
computer spreadsheet. Glacial Geology and Geomorphology, tn01.
http://ggg.qub.ac.uk/ggg/papers/full/1997/tn011997/tn01.html.
Benn, D.I., Evans, D.J.A., 1998. Glaciers and Glaciation. Arnold,
London.
Bennett, M.R., Glasser, N.F., 1996. Glacial Geology: Ice Sheets and
Landforms. Wiley, Chichester.
Cabral, F.A.V.P., 1884. Vestígios glaciários na Serra da Estrela. Revista
de Obras Públicas e Minas 1, 435–459.
Clark, D.H., Clark, M.M., Gillespie, A.R., 1994. Debris-covered
glaciers in the Sierra Nevada, California, and their implications for
snowline reconstructions. Quaternary Research 41, 139–153.
Daveau, S., 1971. La glaciation de la Serra da Estrela. Finisterra 6,
5–40.
Evans, D.J.A., Rea, B.R., Hansom, J.D., Whalley, W.B., 2002.
Geomorphology and style of plateau icefield deglaciation in fjord
terrains: the example of Troms-Finnmark, north Norway. Journal
of Quaternary Science 17, 221–239.
Fernández Mosquera, D., Marti, K., Vidal Romaní, J.R., Weigel, D.,
2000. Late Pleistocene deglaciation chronology in the NW of the
Iberian Peninsula using cosmic-ray produced 21Ne in quartz.
Nuclear Instruments and Methods in Physics Research. Section B,
Beam Interactions with Materials and Atoms 1–6.
Ferreira, A.B., Vidal Romani, J.R., Zêzere, J.L., Rodrigues, M.L.,
1999. A glaciação plistocénica da Serra do Gerês. Vestígios
Geomorfológicos e Sedimentológicos. CEG, Lisboa.
Florineth, D., Schlüchter, C., 2000. Alpine evidence for atmospheric
circulation patterns in Europe during the Last Glacial Maximum.
Quaternary Research 54, 295–308.
Furbish, D.J., Andrews, J.T., 1984. The use of hypsometry to indicate
long-term stability and response of valley glaciers to changes in
mass transfer. Journal of Glaciology 30, 199–211.
García Ruiz, J.M., Martí Bono, C., Valero Garcés, B., González
Sampériz, P., 2001. La evolución de los glaciares del Pleistoceno
Superior en el Pirineo Central español. El ejemplo de los glaciares
de Escarra y Lana Mayor, Alto Valle del Gállego. Cuaternario y
Geomorfologia 15, 103–119.
García Ruiz, J.M., Valero Garcés, B., Martí Bono, C., González
Sampériz, P., 2003. Asynchroneity of maximum glacier advances
in the central Spanish Pyrenees. Journal of Quaternary Science 18,
61–72.
Gemmel, A.M.D., 1997. Fluctuations in the thermoluminescence
signal of suspended sediment in the Alpine glacial meltwater
stream. Quaternary Science Reviews (Quaternary Geochronology)
16, 281–290.
Hooke, R., Le, B., 1998. Principles of Glacier Mechanics. Prentice
Hall, New Jersey.
Jiménez Sánchez, M., Farias Arquer, P., 2002. New radiometric and
geomorphologic evidences of a last glacial maximum older than
18 ka in SW European mountains: the example of Redes Natural
Park (Cantabrian Mountains, NW Spain). Geodinamica Acta 15,
93–101.
Lautensach, H., 1929. Eiszeitstudien in der Serra da Estrela (Portugal).
Zeitschrift für Gletscherkunde 17, 324–369.
Locke, W., 1995. Modelling of icecap glaciation of the northern Rocky
Mountains of Montana. Geomorphology 14, 123–130.
Lowe, J.J., Walker, M.J.C., 1997. Reconstructing Quaternary
Environments. Longman, Harlow, UK.
Lukas, S., 2005. Younger Dryas moraines in the NW Highlands of
Scotland: genesis, significance and potential modern analogues.
Ph.D. Thesis, University of St. Andrews, UK.
McDougall, D.A., 1995. The identification of plateau glaciers in the
geomorphological record: a case study from the Lake District,
northwest England. In: McLelland, S.J., Skellern, A.R., Porter,
P.R. (Eds.), Postgraduate Research in Geomorphology: Selected
Papers from the 17th BGRG Postgraduate Symposium, BGRG,
Leeds, pp. 1–8.
Meierding, T.C., 1982. Late Pleistocene glacial equilibrium-line in the
Colorado Front range: a comparison of methods. Quaternary
Research 18, 289–310.
Menzies, J., 1995. The dynamics of ice flow. In: Menzies, J. (Ed.),
Modern Glacial Environments. Processes, Dynamics and Sediments 1. Butterworth-Heinemann, Oxford, pp. 139–196.
Muñoz Sobrino, C., Ramil-Rego, P., Rodriguez Guitián, M.A., 2001.
Vegetation in the mountains of northwest Iberia during the last
glacial–interglacial transition. Vegetation History and Archaeobotany 10, 7–21.
Muñoz Sobrino, C., Ramil-Rego, P., Gómez-Orellana, L., 2004.
Vegetation of the Lago de Sanabria area (NW Iberia) since the end
of the Pleistocene: a palaeoecological reconstruction on the basis
of two new pollen sequences. Vegetation History and Archaeobotany 13, 1–22.
Munroe, J.S., Mickelson, D.M., 2002. Last Glacial Maximum
equilibrium line altitudes and paleoclimate, northern Uinta
Mountains, Utah, U.S.A. Journal of Glaciology 48, 257–266.
Nesje, A., 1992. Topographical effects on the equilibrium-line altitude
on glaciers. Geojournal 27, 383–391.
Nesje, A., Dahl, O., 1992. Equilibrium-line altitude depressions of
reconstructed Younger Dryas and Holocene glaciers in Fosdalen,
inner Nordfjord, western Norway. Norsk Geologisk Tiddskrift 72,
209–216.
Nesje, A., Dahl, O., 2000. Glaciers and Environmental Change.
Arnold, London.
Nye, J.F., 1952. A method of calculating the thickness of ice-sheets.
Nature 169, 529–530.
Ohmura, A., Kasser, P., Funk, M., 1992. Climate at the equilibrium
line of glaciers. Journal of Glaciology 38, 397–411.
Osmaston, H., 2005. Estimates of glacier equilibrium line altitudes by
the Area × Altitude, the Area × Altitude Balance Ratio and the
G. Vieira / Geomorphology 97 (2008) 190–207
Area × Altitude Balance Index methods and their validation.
Quaternary International 138-139, 22–31.
Pallàs, R., Rodés, Á., Braucher, R., Carcaillet, J., Ortuño, M.,
Bordonau, J., Bourlès, D., Vilaplana, J.M., Masana, E., Santanach,
P., 2006. Late Pleistocene and Holocene glaciation in the Pyrenees:
a critical review and new evidence from 10Be exposure ages, southcentral Pyrenees. Quaternary Science Reviews 25, 2937–2963.
Paterson, W.S.B., 1994. The Physics of Glaciers. ButterworthHeinemann, Oxford.
Pérez Alberti, A., Valcárcel Díaz, M., Blanco Chao, R., 2004.
Pleistocene glaciation in Spain. In: Ehlers, J., Gibbard, P.L. (Eds.),
Quaternary Glaciations — Extent and Chronology. Elsevier,
Amsterdam, pp. 389–394.
Peyron, O., Guiot, J., Cheddadi, R., Tarasov, P., Reille, M., Beaulieu, J.L.,
Bottema, S., Andrieu, V., 1998. Climatic reconstruction in Europe for
18 000 YR B.P. from Pollen Data. Quaternary Research 49, 183–196.
Pierce, K.L., 1979. History and dynamics of glaciation in the northern
Yellowstone National Park area. U.S. Geological Survey Professional Paper 729-F.
Porter, S.C., 2001. Snowline depression in the tropics during the Last
Glaciation. Quaternary Science Reviews 20, 1067–1091.
Prescott, J.R., Robertson, G.B., 1997. Sediment dating by luminescence: a review. Radiation Measurements 27, 893–922.
Preusser, F., 1999. Luminescence dating of fluvial sediments and
overbank deposits from Gossau, Switzerland: fine grain dating.
Quaternary Geochronology 18, 217–222.
Rea, B., Evans, D., 2005. Plateau icefield landsystems. In: Evans, D.
(Ed.), Glacial Landsystems. Hodder Arnold, London, pp. 407–431.
207
Reille, M., Beaulieu, J.L., Svobodova, H., Andrieu-Ponel, V., Goeury, C.,
2000. Pollen analytical biostratigraphy of the last five climatic cycles
from a long continental sequence from the Velay region (Massif
Central, France). Journal of Quaternary Science 15, 665–685.
Rivera, A., Casassa, G., 1999. Volume changes on Pio XI glacier,
Patagonia: 1975–1995. Global and Planetary Change 22, 233–244.
Schilling, D.H., Hollin, J., 1981. Numerical reconstructions of valley
glaciers and small ice caps. In: Denton, G.H., Hughes (Eds.), The
Last Great Ice Sheets. Wiley, New York, pp. 207–220.
Sissons, J.B., 1980. Palaeoclimatic inferences from Loch Lomond
Advance glaciers. In: Lowe, J.J., Gray, J.M., Robinson, J.E. (Eds.),
Studies in the Lateglacial of North-West Europe. Pergamon,
Oxford, pp. 31–44.
Sugden, D.E., John, B.S., 1976. Glaciers and Landscape: a Geomorphological Approach. Edward Arnold, London.
Van der Knaap, W.O., Van Leeuwen, J.F.N., 1997. Late glacial and early
Holocene vegetation succession, altitudinal vegetation zonation, and
climate change in the Serra da Estrela, Portugal. Review of
Palaeobotany and Palynology 97, 239–285.
Vidal Romaní, J.R., Fernández Mosquera, D., Marti, K., Brum Ferreira,
A., 1999. Nuevos datos sobre la cronología glaciar pleistocena en el
NW de la Península Ibérica. Cadernos do Laboratorio Xeolóxico de
Laxe 24, 7–30.
Vieira, G., 2004. Geomorfologia dos planaltos e altos vales da Serra da
Estrela. Ambientes frios do Plistocénico Superior e dinâmica
actual. PhD. Thesis, University of Lisbon, Portugal.