GENERAL CHARACTERISTICS OF THE GLACIAL GEOMORPHOLOGY OF THE

SERRA DA ESTRELA

GONÇALO TELES VIEIRA & ANTÓNIO DE BRUM FERREIRA

Centro de Estudos Geográficos and Department of Geography, University of Lisbon

1. INTRODUCTION Serra da Estrela glaciation (fig. 2). According to

DAVEAU (1971: 6), Lautensach’s general ideas
Glacial studies in Portugal started in the 19th remain valid, especially the fact that only
Century and were first marked by controversial remains of one glaciation are observed in the
observations. During that period raña deposits mountain, and also the idea that the glacial
(Pliocene debris-flow deposits) were frequently morphology is to a large part controlled by the
and erroneously interpreted as of glacial origin pre-glacial morphology and weathering
by authors like RIBEIRO (1866-67), CABRAL conditions. DAVEAU interpretation points
(1881) and DELGADO (1895-98). These towards a somewhat smaller extent of the main
interpretations suggested a very wide glaciation glaciers than previously presented by
especially in the North of Portugal, which LAUTENSACH.
according to CABRAL (1881) had reached the Since 1995 the glacial and periglacial
coast near Oporto. From these authors, CABRAL geomorphology of the Serra da Estrela are being
gave especially important contributions, once he studied by VIEIRA and FERREIRA following the
was the first to suggest with conclusive main methodological framework of the studies
observations the Serra da Estrela glaciation in by VIDAL-ROMANI et al. (1990) and FERREIRA
1884. It was also in that same year that PENCK, et al. (1992) in the Serra do Gerês (NW
according to OBERMAIER (1884 in LAUTEN- Portugal). Research is based in a detailed
SACH, 1929, 1932), suggested a glaciation for geomorphological mapping, aerial photo-
the Serra da Estrela and Sierra de Guadarrama interpretation, location and recognition
(Spain). of the main deposits, laboratory analysis
Modern studies on the Serra da Estrela of the sediments and data integration,
glaciation can be attributed to the current handling and analysis using a geographical
century. FLEURY (1914) has described the information system and database management
general features of the glaciation, but it was with system.
LAUTENSACH, who visited the area for three The present paper synthesises the present-day
months in 1927 and 1928 (papers published in knowledge on the Serra da Estrela glaciation.
1929 and 1932) that the glacial geomorphology The main features of the glaciation are mainly
of the Serra da Estrela became more accurately based in the results of DAVEAU (1971). The
interpreted and defined (figure 1). However, research of the authors conducted up to now
Lautensach’s research was based in a poor focus in the NE sector of the Serra da Estrela
topographic map at a scale 1:100,000 which and especially in the Zêzere valley. The new
made it difficult to map the glacial features. observations correspond still to poorly studied
Forty years later, DAVEAU (1971) publishes evidence and a more accurate sedimentological
‘La glaciation de la Serra da Estrela’ and with description and laboratory analysis, in order to
new field observations, a 1:25,000 topographical interpret their geomorphological significance is
map and aerial photographs, she presents a more going on. Therefore, many of the new
accurate study and overview of the glaciation of observations presented here, correspond to yet
the Serra da Estrela. A map of the glaciation at unsolved problems and most of the interpretation
an approximate scale of 1:90,000 was published is preliminary.
and still constitutes the best synthesis of the

Vieira, G. T. & Ferreira, A. B. (1998) - General characteristics of the glacial geomorphology of the Serra da Estrela, in Vieira, G.
T. (ed), Glacial and Periglacial Geomorphology of the Serra da Estrela. Guidebook for the field-trip, IGU Commission on Climate
Change and Periglacial Environments, 26-28 August 1998, CEG and Department of Geography, University of Lisbon: 37-48.
Figure 1 - The Serra da Estrela glaciation according to Hermann Lautensach (1929, 1932). Legend following the
order in the figure: 1. Contours of the glacier surface; 2. nunataks; 3. lakes; 4. Post-glacial drainage; 5. Glacial
maximum moraines; 6. Moraine of the retreat stages 1, 2 and 3; 7. probable glacier limit.

2. OVERVIEW OF THE SERRA DA ESTRELA high elevation (ca. 1,700-1750 m asl), did not
GLACIATION present active glaciers, but a relatively thin and
static snow (or ice) cover. This is supported by
2.1 The glaciers the conservation of tors, weathering mantle and
absence of glacial deposits.
The Serra da Estrela glaciation is attributed to Daveau (1971) identified three types of
the Last Glacial Maximum, but absolute datings valley glaciers:
does not exist. This interpretation is supported in - South and Southeast facing glaciers, which
the good conservation of the moraine deposits include the Estrela and Alvoco glaciers. These
and in the correlation with the western European were small glaciers with 2.5 and 2 km length
glacial episodes. which stood hanging from the ice-cap in the
The pre-existent morphology of the mountain southern part of the plateau at about 1,300 m asl;
largely controlled the glacier development - North and Northwest facing glaciers, including
(LAUTENSACH, 1932; DAVEAU, 1971). An ice- the Covão do Urso and Covão Grande glaciers,
cap formed in the upper plateaus to the west of both ca. 6.5 km long, finishing at 1,050 and 970
the Zêzere-Alforfa valleys between 1,993 and m asl;
ca. 1,700 m asl and from it several valley - the glaciers of the Zêzere, Alforfa and Loriga,
glaciers diverged towards the lower parts of the with respectively, 13 km, 5.5 km and 7 km
Serra da Estrela (fig. 2). DAVEAU (1971) length, that terminated at much lower altitudes,
estimated a glaciated area of ca. 70 km2 and 80 between 680 m and 850 m asl, a fact mostly
m thickness of ice in the Torre plateau. related to their lee position, which allowed a
Lautensach and Daveau suggest that the Poios significant accumulation of snow.
Brancos - Alto da Pedrice area, despite of its

38
Figure 2 - The Serra da Estrela glaciation according to DAVEAU (1971). 1. Scarps in the glaciated area; 2. lake; 3.
infilled lake; 4. moraine ridge; 5. moraine debris; 6. glacier boundary; 7. proglacial river; 8. kame terrace; 9.
proglacial deposit; 10. creek and cone.

The west-east asymmetry that characterise reflects important Summer insolation, may be
the Serra da Estrela glaciation was attributed by related to a Mediterranean climatic regime.
LAUTENSACH (1932) to west wind action that During Winter the role of wind would be
caused snowdrift towards the Zêzere and Alforfa prevailing.
glaciers. DAVEAU (1971) supports this idea but The low altitudes attained by the Zêzere,
also emphasises a north-south asymmetry related Alforfa and Loriga glaciers seems mostly related
to the more important snowmelt in the southern to the pre-glacial morphology that conduced a
slopes that must have occurred during Summer. rapid glacier flow from the ice-cap towards
Daveau also suggests that this asymmetry lower areas.

39
 Figure 3 - The ‘U-shaped’ sector of the Zêzere valley.

2.2 The erosion landscape - third, the belt of ‘bell-shape rocks and the
boulder and pebble fields’, described as
The distinction between glacial and non- equivalent to the transitional fluvio-glacial fans;
glacial landscape is relatively easy in the Serra - fourth, the domain of the tors and castle-
da Estrela. The areas affected by the glaciers are koppjes and ‘weathered slopes’.
mostly characterised by bare-rock outcrops or The glacial erosion forms appear especially
only by a veneer of recently weathered granite. above 1,300 m asl. Although, most of the glacial
The area outside the glacier limits usually show landscape results in fact from the glacial erosion
a granite weathering mantle inherited from pre- on the existing pre-glacial relief, which still
Quaternary morphoclimatic conditions and a prevails. This is especially true in what respects
typical tor and Bornhardt granite morphology. to the plateaus that are pre-Quaternary, where
This very obvious difference was used by the glaciers only caused minor erosion. In the
Lautensach and Daveau for the definition of the plateaus, where slope is higher and the glaciers
maximum extent of the glaciers and has proved were more active, a knob and basin morphology
to be a very valuable method in the area, once appears. Small lakes and peat-bogs develop and
the obtained results are usually confirmed by the in some areas a ‘roche-mouttonée’ morphology
more recently studied sedimentological evidence also appear (e.g. the Lagoa Comprida area).
(DAVEAU et al., 1997). It was along the valleys and steeper slopes
LAUTENSACH (1929, 1932) presented an that glacier flow was more active. Good
interesting simplification of the Serra da Estrela examples of typical ‘U-shaped’ valleys exist
landscape, which is influenced by this overall along parts of the Alforfa, Loriga, Alva and
symmetry. Roughly four concentric areas with Ribeira da Caniça valleys. The Zêzere valley has
characteristic morphology were to be found: an almost perfect ‘U-shape’ and is considered an
- first, the ‘roche moutonées and erratics area’ in excellent example of a glacially sculpted valley
the upper part of the mountain; (fig. 3). Its perfect form is due to an active and
- second, the thin belt of lateral and end- thick (more than 300 m) glacier, but moreover,
moraines;

40
to a fragile and tectonised substratum related to prepared the bedrock for glacial erosion and
the major Unhais da Serra - Bragança fault. explains the importance of the morainic
The typical glacial treads and risers are accumulations.
especially present in the upper parts of the The moraines appear in distinct locations in
valleys, with the best examples in the ‘Covões’ the Serra da Estrela, but concentrate in the areas
of the Loriga valley in the western slope of the away from the dispersion centre in the Torre
Serra da Estrela, where a succession of glacial plateau. The maximum altitude of the lateral
steps exist (fig. 2). Many of the treads are over- moraines was used by Lautensach (1929, 1932)
excavated and constitute basins occupied by and Daveau (1971) as an indicator of the snow-
lakes, especially where a well-developed riegel line during the glacial maximum, which was set
appears.. In the Serra da Estrela these basins are to ca. 1,650 m asl.
in many cases buried by the Holocene sediment The main moraine ridges are lateral moraines
deposition. Many of these sites were used for the and they are still present in all the main glaciated
construction of hydro-electrical dams (e.g. valleys. The Covão do Urso lateral moraine with
Covão do Ferro, Lagoa Comprida and Covão da ca. 4,500 m length is the longest moraine in the
Nave). Serra da Estrela and marks the northern limit of
Cirques appear around the plateau and the ice-cap. Similar, but smaller is the Covão
can be divided in two main types. The Zêzere Grande moraine, with ca. 2,500 m. In the Lagoa
valley head cirque, named locally Covão Seca several moraine arcs (fig. 4) show that this
Cimeiro, is the best example of a simple cirque col was a diffluence site of the Zêzere glacier
with a 850 m wide amphitheatre form and very and that the glacier had at least 300 metres
steep and long walls (ca. 300 m). The Covão do thickness in this section of the valley. But the
Ferro cirque is is also of this type, the same more impressive site in what respects to the
happening with other smaller cirques. The other morainic accumulations is the Nave de Santo
type of cirques are more complex features. They António (fig. 5), a col where the Zêzere and
are usually wider, not so steep and their Alforfa glaciers probably joined during the
complexity is related to the existence of smaller glacial maximum. There, it is possible to identify
cirques, steps, knobs and basins in the interior of two moraine ridges from both glaciers and a
amphitheatres up to 2,300 metres wide. These complex boulder field, which still raises many
cirques are generally in a transition position unsolved problems.
between the plateau and the valleys. Good DAVEAU (1971) was the first to identify
examples of complex cirques appear in the the existence of kame terraces along the main
Candieira, Covão do Boieiro and Covões. valleys. These consist of hanging accumulations
In what concerns to small-scale features, they of non-sorted debris-flow deposits that
appear all over the erosion areas. Very nice accumulated against the main valley glaciers.
examples of asymmetrical roches moutonnées Their hanging position and size make them
appear in the Lagoa Comprida area. Nye- easily observed both in the 1:25,000
channels were found in the Covões cirque. topographical maps and aerial photos and allow,
Friction cracks, striations, glacial polished quartz in the absence of lateral moraines, to identify the
veins and rock surfaces can be found especially position of the valley glaciers. Very good
in the treads and riegels in the Covão do Boieiro, examples of kame terraces are to be found in the
Covão do Meio, Covões and Candieira areas. Zêzere (fig. 6 and 7) and Alforfa valleys.
According to DAVEAU (1971), proglacial
accumulations appear in the Zêzere, Alforfa and
2.3 The accumulation features Alvoco valleys. They consist of accumulations
of poorly sorted, boulders, cobbles and pebbles
The accumulation forms of glacial origin, of rounded granite and are related with the
especially the moraines are well-developed melting of the glaciers. The best example are the
features in the Serra da Estrela. Daveau (1971) Unhais da Serra terraces which following the
considers that their development is large when author has deposits from the glacial maximum
compared to the extent and probable duration of and from a later recession stage. The Zêzere
the glaciation. This is attributed to the valley near Manteigas also presents significant
significance of pre-glacial weathering, which proglacial accumulations (fig. 8 and 7). Daveau

41
Figure 4 - The Lagoa Seca col with moraine accumulations that resulted from the Zêzere glacier diffluence. It is
possible to see a small moraine-dammed lake, now almost completelly infilled by post-glacial sediments.

Figure 5 - The Nave de Santo António col with the Poio do Judeu moraine and a complex boulder accumulation.

42
Figure 6 - The Covais kame terrace, first identified by DAVEAU (1971). The material accumulated against the
Zêzere valley glacier and now is hanging in the slope above the valley floor.

Figure 7 - Location of the main deposits mentioned in the text (Zêzere valley).

43
 Figure 8 - Glaciofluvial terrace of the Zêzere glacier in Manteigas.

suggests that these are not related to the identification of new evidence concerning the
maximum of the Zêzere glacier, but to a stage glacial history of the valley.
when the glacier was stable near Caldas de The problem of the maximum extent of the
Manteigas at an altitude of 850 m. Zêzere glacier, which has been a subject of
debate for a long time is still unsolved. The high
gradient of the slopes in the Manteigas area and
3. NEW OBSERVATIONS ON THE SERRA DA the importance of fluvial erosion in the valley
ESTRELA GLACIATION make the moraine and sediment conservation
very difficult. The scarcity of geomorphological
Research that is being conducted in the Serra elements is in fact the main reason for the debate
da Estrela since 1995 is in an early stage and that is going on for several decades.
several steps are still needed, especially in what
concerns to the conclusion of the systematic 3.1.1. The Manteigas football field deposit
geomorphological mapping and laboratory
analysis of the sediments. Several new findings Hermann Lautensach (1932) defined the
are worth mentioning. maximum glacier extent to be located near São
Gabriel, 2 km downvalley from Manteigas at an
altitude of 650 m. This was also what Cabral
3.1. The Zêzere valley (1884) had thought, based on deposits he
interpreted as a subglacial moraine. DAVEAU
The Zêzere valley is the best example of the (1971, 1973) has shown that those deposits were
effects of glacial morphogenesis in the Serra da periglacial and indicated the glacier limit some 1
Estrela. It had the longest valley glacier with ca. km upvalley at 680 m asl. This assumption was
13 km length, down to the Manteigas area (fig. based on the existence of a coarse and poorly-
2). The preliminary field survey, followed by the sorted deposit (fig. 9 and 7) probably related to
aerial photo interpretation, systematic detailed debris-flow phenomena in the left slope of the
geomorphological mapping at a scale 1:10,000 valley, near the Manteigas football field, at an
and study of the deposits, is allowing the altitude of 900 m (fig. 9). The slope position of

44
 Figure 9 - Manteigas football field deposit (Zêzere valley).

the deposit, hanging over the valley bottom, led deposits, non-sorted, silty-sandy, with gravel and
Daveau to interpreted it as a kame terrace of very compacted. They are probably sub-glacial
genesis similar to others that appear upvalley. lodgement tills (DAVEAU et al., 1997). Several
DAVEAU et al. (1997) rise some questions diamicts of this kind were found in exposures all
concerning this interpretation. Without a detailed over the Zêzere valley upvalley from Manteigas
sedimentological study and geomorphological in variable slope positions and altitudes (fig. 7).
mapping, it is difficult to accurately define the The exposures found between Lameiro do Grilo
existence or not of a relationship between the and Caldas de Manteigas between 1,050 and
‘football field deposit’ and the glacier itself. 1,085 m asl indicate that the glacier in the
Field observations suggest that the deposit is part Lameiras area was at least 200 m thick (DAVEAU
of a more complex fan, probably of debris-flow et al., 1997). Furthermore, these deposits support
origin. The accumulation of the material in a the interpretation proposed by DAVEAU (1971)
slope position may be possible without a glacier for the Covais accumulation - a kame terrace.
blocking the sediment transport, once in this part
of the slope there exists an erosion level, which
may have favoured the accumulation. 3.1.3 The lateral moraine of the Zêzere glacier
in Fonte Santa

3.1.2. The subglacial tills of the Zêzere valley Another very recent finding in the Zêzere
valley that is important for the definition of the
Recent investigations have also stressed on end of the Zêzere glacier, is a well-preserved
the existence of several diamicton outcrops with lateral moraine up-slope of Fonte Santa, ca. 1
a facies very similar to the ones of the Serra da km upvalley from Manteigas. The moraine has
Gerês, for which a glacial origin has been proved an height of 3-5 m where it is better preserved at
through laboratory analysis (VIDAL ROMANI et appears inside a small forest area in the right
al., 1990). These diamictons are granitic origin slope of the valley. Its location inside the forest

45
area may be the cause for the lateness of its took place near the Espinhaço do Cão moraine,
finding. Its existence was suggested by the just below the Candieira hanging valley (fig. 7).
1:15,000 stereoscopic aerial photo interpretation The Espinhaço do Cão is a very well preserved
and later confirmed in the field. It appears in the lateral moraine presumably of the Zêzere glacier
upvalley continuation of a small kame-terrace and corresponding to a late glacial stabilisation
indicated by DAVEAU (1971) and therefore also phase.
confirms its genesis. The deposits that are worth mentioning
The finding of the moraine is important appear in the opposite slope from the Espinhaço
because it is located in a low slope position, de Cão, some 10 m above the valley floor. It is a
indicating that the Zêzere glacier reached a sedimentary complex probably related to a
location about 1 km upvalley from Manteigas. recessional stage of the Zêzere glacier after the
Furthermore, the moraine geometry seems to be Espinhaço de Cão stage. The deposits are
in relation with the location of the Covais kame characterized by horizontal or sub-horizontal
terrace, indicating a stability position of the layers of poorly sorted coarse sand and granules,
glacier front. The presence of slope tors in the intercalating with silty and sandy layers. These
right slope downvalley from this site makes it seem to correspond to a valley train environment
difficult to point for a longer glacier. Yet another marked by recessions and small advances of the
possibility is that a weathering mantle could glacier. The sediments are glacio-fluvial, with
have been preserved subglacially from glacier layers of flow-till and even compact sub-glacial
erosion and that the slope tors are post-glacial till. But these observations are still preliminary
forms. and a systematic study of the complex is to be
done. Furthermore, the presence of a peat layer
3.1.4. The glaciofluvial deposit of Manteigas in the lower part of the complex gives us some
(Zêzere valley) hope on obtaining absolute datings for the
complex. It would be the first dating on the
An important type of glacial deposit found in glaciation of the Portuguese mountains.
the Serra da Estrela are the proglacial com-
plexes - figure 8 (DAVEAU, 1971). During field-
work in the area of Manteigas several exposures 3.2 The Cerro Rebolado - Fraga das Penhas area
of the Zêzere proglacial complex of glaciofluvial (Penhas Douradas plateau)
origin were found (fig. 7). These exposures
evidence a coarse deposit of rounded cobbles The Cerro Rebolado - Fraga das Penhas area
and boulders with a coarse sandy matrix, with is very interesting in a geomorphological
imbrication fabric and a typical glacio-fluvial perspective. Besides its significance in what
structure. The complex appears in several sites, concerns to present-day processes it presents
in erosive terraces about 2 meters higher than the also important evidence on the Pleistocene
present valley floor (this corresponding to a glacial dynamics of the Serra da Estrela. The
minimum of the post-glacial fluvial erosion in area marks a very clear geomorphological limit
the site). between the glaciated plateaus to the south and
The deposit must have been continuous the glacier free areas, to the north. This
forming a valley train from the Santo André area difference is sharply defined in the landscape,
(Manteigas) at least down to the Leandres with a typical granite morphology related to
valley. In the latter site a periglacial scree covers differential weathering with tors and castle-
the glaciofluvial terrace. This and other recent koppjes to the north, and a tor free area to the
observations may give us some information on south. The limit of the plateau glacier is
the relationships between glacial and periglacial coincident with the Fraga das Penhas - Cerro
evolution. Rebolado ridge, where a huge linear boulder
field corresponding to a moraine exists. This
3.1.5 The Espinhaço de Cão complex moraine is in fact the eastern extension of the
(Zêzere valley) Covão do Urso lateral moraine (DAVEAU et al.,
1997) and marks the flow towards the west of
Apart from several recently discovered the glacier that was channelled along the Conde
moraine ridges of small size, significant findings valley.

46
 Figure 10 - The Cerro Rebolado till.

However, the most striking geomorphological accumulation, with some stratification. Its
feature is an exposure ca. 150 m long showing a genesis is probably related to the snowmelt and
regressive sequence of the Conde valley glacier runoff dynamics. It is a typical Holocene deposit
(fig. 10). The exposure appears in a south-north in the Portuguese granitic mountains. Above this
oriented artificial gully that transports the unit a humic soil develops, corresponding to a
outflow water from a small dam in the Conde recent mountain ranker. Towards the south of
valley. It is the major and more complex till the exposure, the upper horizon becomes more
found up to date in the Serra da Estrela. The humic related to the low energy environment of
deposit presents roughly three sedimentary units. the Conde valley where some peaty layers
The lower unit, visible in the northern part of appear. The small plain of the Conde valley,
the gully presents the characteristics of a sub- with abundant herbaceous vegetation and
glacial lodgement till. It is a very compact seasonal flooding corresponds to a post-glacial
diamict, in some places sandy-silty and in others sediment filling.
more coarse-grained. In some sectors stratified The discovery of the Cerro Rebolado
sand lenses related to subglacial melting appear. complex is very recent and a detailed
In places, the sand lenses are deformed may be sedimentological study is to be carried in the
due to ice-margin glaciotectonic processes. Silt near future. The main objectives are to fully map
and clay injections in fractures in the granite and describe the exposure, determine its
substratum are present in parts of the gully. implications in what concerns to the glacial
The intermediate unit is a coarser and poorly history of this part of the mountain and to obtain
consolidated diamicton. It presents no organised absolute datings.
structure and is constituted by cobbles and
boulders with a sandy matrix. It is interpreted
has a supra-glacial till deposited during glacier References
recession.
The upper unit present all over the exposure DAVEAU, S. (1971) - La glaciation de la Serra da Estrela,
is a gyttja-type deposit, an organic rich sandy Finisterra, 11, Lisbon: 5-40.

47
DAVEAU, S. (1973) - Quelques examples d’évolution LAUTENSACH, H. (1929) - Eiszeitstudien in der Serra da
quaternaire des versants au Portugal, Finisterra, 15, Estrela (Portugal), Zeitschrift fur Gletscherkunde,
Lisboa: 5-45. Berlin, XVII: 324-369.
DAVEAU, S.; FERREIRA, A. B.; FERREIRA, N.; VIEIRA, G. T. LAUTENSACH, H. (1932) - Estudo dos glaciares da Serra da
(1997) - Novas observações acerca da glaciação da Estrela, Memórias e Notícias, Coimbra, VI: 1-60
Serra da Estrela, Estudos do Quaternário, 1, Lisboa: (transl. Lautensach 1929).
41-51. VIDAL-ROMANI, J. R.; VILAPLANA, J. M.; FERREIRA, A. B.;
FERREIRA, A. B.; VIDAL-ROMANI, J. R.; VILAPLANA, J. M.; ZÊZERE, J. L.; RODRIGUES, M. L.; MONGE, C. (1990) -
RODRIGUES, M. L.; ZÊZERE, J. L.; MONGE, C. (1992) - Los tills de la Serra de Gerês - Xurés y la glaciación
Formas e depósitos glaciários e periglaciários da Serra pleistocena (Minho - Portugal - Ourense, Galicia),
do Gerês - Xurés (Portugal; Galiza). Levantamento Cuaternario y Geomorfologia, 4, Logroño: 13-25
cartográfico, Cuadernos do Laboratorio Xeoloxico de
Laxe, 17: 121-135.

48