In the past decade, particularly since the last summary of the subject (Calvet, 2004), the Quaternary glaciation of the Pyrenees has been the focus of new research. Unequal progress has been achieved on three aspects: mapping the extent of the Pyrenean ice field, quantifying the geomorphological impact of glaciation on the preglacial landscape and refining the chronology of the glacial fluctuations.
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Provided for non-commercial research and educational use only. Not for reproduction, distribution or commercial use.
This chapter was originally published in the book Developments in Quaternary Science, Vol.15, published by Elsevier, and the attached copy is provided by Elsevier for the author's benefit and for the benefit of the author's institution, for non-commercial research and educational use including without limitation use in instruction at your institution, sending it to specific colleagues who know you, and providing a copy to your institution’s administrator.
All other uses, reproduction and distribution, including without limitation commercial reprints, selling or licensing copies or access, or posting on open internet sites, your personal or institution’s website or repository, are prohibited. For exceptions, permission may be sought for such use through Elsevier's permissions site at:
From: Marc Calvet, Magali Delmas, Yanni Gunnell, Régis Braucher and Didier Bourlès, Recent Advances in Research on Quaternary Glaciations in the Pyrenees. In J. Ehlers, P.L. Gibbard and P.D. Hughes, editors: Developments in Quaternary Science, Vol. 15, Amsterdam, The Netherlands, 2011, pp. 127-139. ISBN: 978-0-444-53447-7
the valley slopes high above the poorly weathered and well-
preserved lateral moraines of Wurmian age (Delmas, 2009;
Delmas et al., 2011).As in theCarol valley, the boulders also
show signs of post-depositionalweathering, of displacement
along fractures and even of taffoni development. Following
the methodological recommendations of Putkonen and
Swanson (2003), out of the sampled population, the oldest10Be age was considered to be the most accurate, providing
a resulting age of 122 ka for the deposits near Caraybat
(Fig. 11.4).
Further insight into the Middle Pleistocene glaciations
has been gained by examining the relative altitudes of glacial
FIGURE 11.3 Age-bracketed terminal sequence
of glacier deposits in the Gallego catchment. 1:
High-relief landforms, razorback and hogback
ridges. 2: aretes. 3: Tread of highest alluvial
deposits (coronas). 4: Proglacial alluvial sheets
and mantled wash pediments (weathered and rubi-
fied: Qt 5, alias T2). 5: Proglacial alluvial sheets
(poorlyweathered:Qt6-7-8, aliasT1). 6:Wurmian
ice-contact sedimentary units (lacustrine or flu-
vial). 7:Moraineridgesandtill deposits.8:Scoured
rock surfaces on bedrock steps. 9: Successive
Wurmian ice front positions. 10: Middle Pleisto-
cene (MIS 6) ice fronts. 11: OSL and 14C-dated
sites. After the geomorphological map by Barrere
(1966) and data by Pena et al. (2004), Lewis et al.
(2009) and Jalut et al. (1992).
Quaternary Glaciations-Extent and Chronology130
Author's personal copy
deposits, the advanced weathering state of which indicates
that they are older thanWurmian. This approach has allowed
differentiation between generations of deposits that are leg-
acies of different stadial events within MIS 6, or perhaps
even of different glacial cycles: (i) in the Tarascon basin
and in the higher Ariege catchment, glacial evidence from
MIS 6 occurs as trails of weathered till deposits and erratic
boulders situated ca. 50–100 m higher than Wurmian
FIGURE 11.4 Pleistocene glaciation in the Ariege catchment. 1: Razorback and hogback scarps in Mesozoic or Cenozoic limestone, sandstone or con-
glomerates of the outer Pyrenean fold belt. 2: Relict Pliocene to early Pleistocene alluvial fan deposits (formation du Lannemezan). 3: Higher alluvial
and rubified (Basse Boulbonne, Vernajoul and ice-marginal terraces of Antras and Foix–Cadirac). 6: Lowermost terrace T1, ofWurmian age, unweathered
and capped by brown soils. 7: Middle Pleistocene ice boundaries; a: probable MIS 6 ice limits, grading to the T2 terraces; b: erratic boulders or residual
morainic deposits dating to MIS 6 or older. 8: Wurmian maximum ice extent; the corresponding population of unweathered moraines grades to the T1
alluvial sheets. 9: MIS 2 ice boundaries, that is, Garrabet stadial. 10: Recessional moraines at Bompas–Arignac and Berniere (19–20 ka). 11: Recessional
moraines (Petches and Freychinede) dating to the Oldest Dryas. 12: Position of dated sites. Distribution of glaciofluvial deposits in the piedmont zone,
based on Hubschman (1975, 1984) and on the geological map of the Pyrenees Quaternary by Barrere et al. (2009). The ice limits and 10Be data are after
Delmas (2009) and the 14C ages after Jalut et al. (1982, 1992).
Author's personal copy
Quaternary Glaciations-Extent and Chronology132
Author's personal copy
maximum glacial moraines. In the Foix–Montgaillard basin,
the Antras ice-contact depositional sequence and the ice-
marginal sediment body at Becq-en-Barguillere (Taillefer,
1973) represent two additional sites to which theMIS 6max-
imum ice extent can be tethered. Given these new interpre-
tations, the ice-marginal meltwater terrace sequence at Foix–
Cadirac was ascribed byHubschman (1975, 1984) to the Ris-
sian. This was because of the advanced state of weathering of
its debris. However, it occurs at a lower elevation than the
Becq and other MIS 6 benchmarks and could instead repre-
sent the legacy of a deglacial event that occurred toward the
end of the MIS 6 cycle. Further down the Ariege valley, the
Foix–Cadirac terrace grades to the Basse Boulbonne (T2)
alluvial sheet. T2 (mapped by Barrere et al., 2009) is capped
by rubified soils displaying a depleted upper horizon. Char-
acteristically evolved soils, such as this, suggest an episode
of weathering and eluviation during the Eemian Stage
(MIS 5e) climatic optimum. (ii) Erratic boulders sourced
by the granitic and gneissic Aston massif have been found
resting on the Mesozoic limestone and shale outcrops at Car-
aybat, down-valley from the Antras delta and 50 m above it.
This boulder trail was produced by one or several ancient gla-
ciers of corresponding thickness and has been tentatively cor-
related with elevated moraine deposits on the outer flanks of
the Arize massif and in the Tarascon basin at Quemenailles
and Menties. In their lower reaches, these ice masses may
have extended as far as Pech deVarilles, at the proximal edge
of the Pamiers plain. At Pech de Varilles, Faucher (1937)
reported erratic boulders of gneiss exceeding 6 m in diame-
ter, sourced from the Astonmassif. He hypothesised that they
were of glacial origin. However, other geologists later either
argued that the boulders had been supplied from the under-
lying Eocene conglomerate bedrock or else ascribed them to
unconformable Pliocene fan deposits that excavated the
Eocene conglomerate and locally contain greater than 1 m-
sized quartzite boulders. Nevertheless, the boulders
described by Faucher are conspicuous by their petrography
and provenance, by their large size, and by their extremely
limited weathering state. They occur ca. 460 m a.s.l., that
is, 90 m above the tread of the deeply weathered Haute Boul-
bonne (T3) terrace, which occupies the opposite bank of the
Ariege river in the same area (Calvet, unpublished). These
indications imply that this elevated erratic boulder trail is
of early Middle Pleistocene age and correlates, as in the
Carol valley, with the older alluvial terrace T4 deposits. This
topographical position suggests that the erratics at Pech de
Varilles are comparable to the ancient moraine deposits that
occur much further west at Arudy. The latter also occur at
elevations similar to those of the relict Pliocene to early
Pleistocene piedmont deposits of the Gave d’Oloron catch-
ment (Hetu and Gangloff, 1989). CRN dating was not
attempted on the Pech de Varilles erratics because field evi-
dence indicates that they have been exhumed from a finer
matrix and have been displaced by slope movements.
11.3. THE LAST PLEISTOCENE GLACIALCYCLE (WURMIAN STAGE)
The chronology of the Pyrenean last glacial cycle, corre-
lated with theWurmian Stage of the Alps by most Pyrenean
authors, has been mostly based on a large set of 14C ages
produced over the past 30 years, supplemented in the last
five mostly by new 10Be and OSL ages. The image of the
Wurmian chronology provided by this body of evidence
is still incomplete, with contradictions between the methods
employed giving sometimes puzzling contrasts between
adjacent valleys. Radiocarbon dating is usually indirect in
that it provides an age for proglacial or ice-marginal organic
sediments whose relation to adjacent moraines is not always
straightforward. Even though a few terminal moraines have
been dated directly, most OSL dating raises similar prob-
lems. This is because sampling strategies target alluvial
sheets that are sometimes situated several tens of kilometres
from the nearest ice front deposits. CRN dating, in contrast,
systematically allows direct dating of glacial landforms,
whether moraines, isolated erratic boulders or ice-scoured
bedrock steps. However, CRN exposure ages are, by defi-
nition, minimum ages, with the precautions this limitation
inevitably brings upon interpretation of the data.
11.3.1. The Age of Wurmian Maximum IceExtent: In or Out of Phase with the Global LastGlacial Maximum?
The mid-latitude Pyrenean mountain range is under direct
influence of the North Atlantic weather systems. This
makes the Pyrenees an ideal setting for testing whether or
not mid-latitude mountain glaciation and global climatic
fluctuations were synchronous (Gillespie and Molnar,
1995; Hughes and Woodward, 2008; Thackray et al.,
2008). The now classic view that the Wurmian Pyrenean
maximum ice extent and the global last glacial maximum
(LGM-defined as the time period between 23 and 19 ka
cal. BP, Mix et al., 2001) did not occur simultaneously
was presented in the 1980s on the basis of radiocarbon ages
obtained on ice-contact lake sediments on the north side of
the range (Andrieu et al., 1988; Jalut et al., 1988, 1992;
Andrieu, 1991; detailed synthesis in Delmas, 2009,
pp. 96–109). This radiocarbon-based perspective on
Pyrenean glaciation advocated a Wurmian maximum ice
extent between 70 and 50 ka. After this time, the glaciers
stagnated slightly up-valley from their maximal advance
positions before retreating to the uppermost reaches ca.
29–25 ka,dependingonsites.Toexplain therecessional state
of the Pyrenean ice cap by the time of the globalLGM(which
is broadly correlable to MIS 2), it has been argued that cli-
matic conditions were too dry to sustain extensive ice fields
on the Pyrenees. Although data are comparatively scarcer, a
Chapter 11 Recent Advances in Research on Quaternary Glaciations in the Pyrenees 133
Author's personal copy
similar model has been proposed for the southern side of
the Pyrenees. In the Noguera Ribagorcana catchment
(Fig. 11.5), the ice-marginal site of Llestui, interpreted
as marking a Wurmian post-maximum ice position, was
initially radiocarbon dated at 31–34 14C ka BP (Vilaplana,
1983). However, the samples were found to have been con-
taminated by inorganic carbon and later redated at 18–2114C ka BP (Bordonau et al., 1993). For reasons of sus-
pected sample contamination, the 30 14C ka BP radiocar-
bon age obtained from the Bassots palaeolake, which is
ponded behind the terminal lobe of the Ribagorcana
moraine, has also been discounted as unreliable (Bordonau
i Ibern, 1992).
FIGURE 11.5 Terminal sequence of glacier deposits in the Noguera
Ribagorcana catchment, age of the Wurmian moraines. 1: Mountain ridge.
2: Unweathered Wurmian till or moraine. 3: Wurmian fluvio-lacustrine
ice-contact deposit. 4: Wurmian ice front limits and typical U-shaped val-
ley cross-profiles; 14C- and 10Be-dated sites. The map and chronological
data are after Bordonau et al. (1993), Pallas et al. (2006) and Rodes (2008).
Amidst mounting uncertainty over radiocarbon ages of
ice-marginal materials, critics have begun to question the
asynchroneity of the Pyrenean glacial model. Basing their
reassessment on biostratigraphical and geochemical evi-
dence, palynologists in the late 1980s were first to challenge
the view that Pyrenean glacier fluctuations were out of
phase with the global LGM (Turner and Hannon, 1988;
Reille and Andrieu, 1995). The first CRN ages produced
for glacial landforms in the Noguera Ribagorcana (Pallas
et al., 2006) prompted a more systematic critique of radio-
carbon ages from the Pyrenees, most of them being deemed
unreliable and were rejected. The competing hypothesis of
synchroneity between the Wurmian maximum ice extent
and the global LGM was asserted. This was based on10Be ages of 18.1�1.9 ka obtained from a lateral moraine
of the Noguera Ribagorcana trunk glacier and 21.3�4.4 ka
obtained from an erratic boulder close to the terminal lobe
(Fig. 11.5). These 10Be ages were revised to 19.2�2 of the
Noguera Ribagorcana trunk glacier and 22.6�4.7 ka,
respectively, by Rodes (2008). Pallas et al. (2006) have nev-
ertheless not rejected the hypothesis that a relatively pro-
tracted ‘Pleniglacial’ might have prevailed between
before 30 ka and ca. 20 ka. This means that glacier snouts
may have been stationed far down the valleys not only dur-
ing the global LGM but also for quite some time before it.
The body of geochronological information available
today, however, provides evidence of a somewhat less static
history of glacier fluctuation, and consequently, a less seri-
ous antagonism between the in-phase and out-phase schools
of Pyrenean glaciation. The asynchroneity of maximum
advances of mountain and continental glaciers needs no
longer to be challenged. The Wurmian maximum ice extent
now appears to have occurred everywhere quite early dur-
ing the Wurmian Stage, irrespective of the dating methods
used to test this hypothesis. More controversial is the
relative position of ice fronts during the MIS 2 or more
accurately during the global LGM, either because of insuf-
ficient data or because of genuine variability between
valleys along the mountain range.
11.3.2. Dating of Pyrenean WurmianTerminal Moraine Occurrences
On the Spanish side, the most abundant evidence has been
obtained from the Gallego catchment (Fig. 11.3). The ter-
minal moraine at Senegue has yielded two OSL ages of
36�3 and 36�2 ka, respectively (Pena et al., 2004; Lewis
et al., 2009). In the western part of the catchment, elevations
do not exceed 2.7 km. As the radiocarbon ages of 30–25 14C
ka BP obtained from proglacial lake sediments around Tra-
macastilla and Pourtalet demonstrate (Garcıa-Ruiz et al.,
2003; Gonzalez-Samperiz et al., 2006) this area was already
partially deglaciated in MIS 2 and cut off from transfluent
Quaternary Glaciations-Extent and Chronology134
Author's personal copy
sources of northern ice spilling over the Col du Pourtalet. A
core obtained from a cirque located 2 km to the east of the
Col du Pourtalet indicates a possible readvance between
24.17 and 19.25 14C ka BP. This advance would have
occurred during MIS 2 even though ice had vacated the
topography at elevations situated above that site since
before 19.25 14C ka BP. The position of the Gallego ice
front in MIS 2 is difficult to establish, but the 20.8 14C
ka BP (i.e. 24.217 ka cal. BP) radiocarbon age obtained
from the base of the Bubal lake deposits (Jalut et al.,
1992) suggests that the LGM glacier front was stationed
as much as 15 km up-valley from Senegue.
In the Cinca catchment, the maximum extent of the Ara
glacier is constrained by ice-contact lake deposits at Linas
de Broto. A 50-m drillhole here produced a mid-core AMS14Cageof 30.38 14CkaBP (Martı-Bonoet al., 2002), indicat-
ing an early age ofWurmian maximum ice extent and possi-
ble maintenance of this advanced position until MIS 2 and
global LGM time. In the Cinca valley, till at Salinas de Sin
is associated with glaciofluvial deposits and has yielded
OSLages of 46�4, 63�6and71�15 ka.TheSalinas expo-
sure has been subjected to glaciotectonic deformation, proof
that it has also been overridden by the glacier since 46 ka
(Lewis et al., 2009).
On the Noguera Ribagorcana (Fig. 11.5), MIS 2 ages
have been obtained by 10Be dating of deposits situated in
the vicinity of the presumed glacier terminus (Pallas
et al., 2006). The terminal lobes, however, have been signif-
icantly eroded, making it difficult to draw any firm conclu-
sions from this valley about the age of the Wurmian
maximum ice extent. In the Valira catchment, most of the
data are still unpublished but they clearly confirm that
the Wurmian maximum ice extent occurred much earlier
than the global LGM. The ice-marginal lake sediment
sequence at La Massana, which formed where the Valira
del Nord and Valira d’Orient glaciers separated, has been
subjected to detailed sedimentological and stratigraphical
analysis of its progradational deltas. AMS radiocarbon
dates have returned ages of 25.63, 21.51 and 17.43 14C
ka BP (29031 to 27820, 24150 to 23310, 21097 to 20271
ka cal. BP) and indirectly confirm the presence of this lake
since �41 ka on the basis of sedimentation rate calcula-
tions. The lake sediments subsequently recorded a succes-
sion of three glacier advances of diminishing intensity, all
falling chronostratigraphically within MIS 2 (Turu i
Michels, 2002). The maximum ice extent may have been
captured by a 59�1.18 21Ne ka age obtained from an
ice-scoured bedrock surface that was sampled on the upper
flank of the glacial trough above the village of Canillo
(Turu i Michels et al., 2004).
In the eastern Pyrenees, 10Be exposure ages show that
the MIS 2 maximum ice advance reached positions within
a few hundred metres of the Wurmian maximum ice extent.
The frontal moraine of the Tet glacier at Mont Louis
(Fig. 11.6) has provided an age of 21.4�3.7 ka (Delmas
et al., 2008), since then recalibrated as 22�3.6 ka (Delmas,
2009). However, the dated boulders belong to the innermost
lobe crest of the Tet Wurmian terminal moraine sequence.
The two outermost lobes have so far not been dated but
could correspond to the sites of maximum ice extent
reached earlier during the Wurmian stage. In the Carol
catchment (Fig. 11.2), the right flank lateral moraine has
FIGURE 11.6 Terminal sequence of Wurmian
glacier deposits in the Tet catchment: age of fron-
tal and lateral moraines. 1: Quaternary incision of
valleys into the Cenozoic erosion surface of the
Col de La Perche. 2: Alluvial sheet T2, weathered
(MIS 6). 3: Proglacial alluvial sheet T1, unweath-
ered and connected to the Wurmian moraines. 4:
Wurmian unweathered till and moraines, with
the position 10Be exposure-dated sites. 5: Succes-
sive Wurmian ice front positions. 6: Wurmian ice-
contact sedimentary units (lacustrine or fluvial).
The mapping and geochronological data are after
Delmas et al. (2008) and Delmas (2009).
Chapter 11 Recent Advances in Research on Quaternary Glaciations in the Pyrenees 135
Author's personal copy
been dated with great precision by three samples:
22.9�2.7, 22.1�3.4 and 21.7�2.8 ka (Delmas, 2009).
However, as in the case of the Tet moraine sequence, the
dated lateral moraine connects with a frontal lobe that is sit-
uated ca. 0.6 km up-valley from the Vinyola terminal lobe.
The dated lateral moraine is therefore not the oldest in the
Wurmian sequence. Finally, the small Malniu glacier,
which remained disconnected from the Carol trunk valley
throughout the Wurmian stage, has yielded 10Be ages of
76.5, 42.3 and 40.6 ka. The post-maximum Malniu frontal
lobe, which is situated 0.5 km further up-valley, provided
an age of 21.3 ka, which accords with the 23.9 ka age of a
coeval lateral moraine preserved on the valley side (Pallas
et al., 2010).
Among the north-facing glacial catchments, only the
Ariege basin has recently delivered new data (Delmas,
2009; Delmas et al., 2011). Results, which largely rely on10Be exposure ages, confirm the classic view of asynchro-
neity between mountain glaciation and the continental ice-
sheet record and are broadly compatible with existing
results from the Spanish catchments. The Wurmian maxi-
mum ice extent (Fig. 11.4) is constrained by two CRN ages
that suggest two glacier advances of comparable extent: a
81.4�14.6 ka age obtained from an erratic boulder in the
vicinity of the terminal moraine and a younger age of
34.9�8.5 ka obtained on the mid-valley lateral moraine
at Larcat. The Ariege glacier advanced to within 2 km of
the town of Foix. The older age is in good agreement with
the 91.4�2.4 ka U–Th age obtained for the stalagmite floor
in the prehistoric cave at Niaux, which occurs 8 km up-val-
ley and is just slightly older than the Niaux glacial deposits
(Sorriaux, 1981, 1982; Bakalowicz et al., 1984). The MIS 2
maximum extent ice front has not been directly dated, but
its position almost certainly coincides with the Garrabet
frontal moraine, which is located 8 km up-valley from the
Wurmian terminal moraine. It has two recessional fronts
at Bompas–Arignac and Berniere, respectively. This chro-
nology was established indirectly by a surface exposure age
with a strong component of nuclide inheritance (31.2 ka) on
the bedrock step of the Tarascon basin and by the age of an
adjacent erratic boulder resting directly on the dated bed-
rock-step surface (16.7 ka). Such a conjunction suggests
so far unreported ice recessions and readvances, that is sta-
dials, within the Pyrenean last glacial cycle. Other bedrock
steps have yielded exposure ages of 19 ka, for example, on a
scoured rock exposure situated 2.5 km up-valley from Gar-
rabet. These ages equate to a �19 ka age for the Berniere
frontal lobe and provide the timing of the deglaciation
increment that followed the Garrabet stage.
So far, no new data have been produced that might help
to update the radiocarbon-dated terminal lobes of the
Garonne, Gave de Pau and Gave d’Ossau catchments
(Fig. 11.1). In these western tracts of the Pyrenean range,
the glaciers are thought to have receded to the upper valleys
as early as 29–25 14C ka BP, and no post-maximum ice front
position corresponding toMIS 2, for example, similar to the
Garrabet frontal system, has been clearly identified. The
interpretations of Jalut et al. (1992) and Reille and Andrieu
(1995), respectively, have diverged on these issues.
Accordingly, a discussion of their respective merits in the
context of this review seems timely. Assuming first that
the 38.4�2 14C ka BP age obtained at the Biscaye site
by Mardonnes and Jalut (1983) is invalid and should be
replaced by the AMS 14C ages obtained by Reille and
Andrieu (1995) from glaciolacustrine clays at the Lourdes
1 terminal moraine site, a Wurmian maximum ice extent
older than 20.025 14C ka BP (22486 to 21562 ka cal. BP)
appears correct, with vacation of the piedmont zone
between 16.67 and 14.46 14C ka BP. New 10Be exposure
data from the Gave de Pau catchment have been obtained
frommorainic boulders situated at the Ossen diffluence site,
immediately above the town of Lourdes. However, with
five Lateglacial and Holocene ages and another of
53.5�38 ka, results remain inconclusive (Rodes, 2008).
Likewise, the moraine sequence at Aucun has yielded four10Be ages ranging between 13 and 10 ka. All these ages,
however, are suspiciously young for this small glacier,
which at that time was disconnected from the Gave de
Pau trunk valley glacier. At the time, as shown by the posi-
tion of the lateral moraines at Argeles-Gazost (Barrere
et al., 1980), this larger glacier still descended to elevations
below 400 m a.s.l. and to positions less than 10 km from the
outermost Wurmian ice front at Lourdes. Assuming now,
instead, that the chronology advocated for the Wurmian
maximum ice extent by Mardonnes and Jalut (1983) is
valid, it is feasible that the Argeles-Gazost and Aucun
moraine ridges could define the post-maximum glacier
position during MIS 2 peak.
If these local insights are assembled into a more coher-
ent Pyrenean-scale picture, it appears that the Wurmian
maximum ice extent occurred somewhat earlier than the
global LGM, perhaps with glaciers advancing out of the
crest zone to similar positions repeatedly during MIS 5d,
5b and 4. By definition, all 10Be ages are minimal exposure
ages and thus cannot precisely date the moraine ridges that
have been sampled. Caution, therefore, must be exercised
when determining the exact number and position of glacier
advances and retreats. For example, the MIS 2 ice fronts
have only been positively identified in the eastern, more
Mediterranean part of the mountain range: for example,
in the Tet, Carol, Malniu, Ariege, Valira and Ribagorcana
valleys. It further appears that, in this region, MIS 2 ice
fronts also nearly reached the Wurmian maximum glacier
positions. The MIS 2 ice fronts reached those positions at
the time of, or just before, the global LGM (i.e. between
23–19 ka cal. BP, Mix et al., 2001). In the western catch-
ments (Garonne, gave de Pau, Gave d’Ossau, Gallego),
which are under greater direct influence from Atlantic
Quaternary Glaciations-Extent and Chronology136
Author's personal copy
weather conditions, MIS 2 appears to have coincided
instead with a recessional period. It is too soon to decide
whether such contrasts should be ascribed to methodologi-
cal artefacts inherent in the dating techniques used, or
whether the observed differences in glacier behaviour along
the strike of the mountain range truly reflect palaeoclimatic
differences. Far from being implausible, such differences
could indeed have been controlled over the western Medi-
terranean by the Balearic low atmospheric pressure centre,
which may have been more active than Atlantic weather
systems during MIS 2 because of the southerly position
of the Polar Front at the time (Florineth and Schluchter,
2000; Hughes and Woodward, 2008).
11.3.3. Evidence of Rapid Deglaciation fromthe end of the global LGM
The stages of ice recession are only known for certain val-
leys, and it is not yet possible to produce a coherent synthe-
sis for the entire Pyrenean range. Evidence of a first
recessional stage is provided by frontal moraines situated
only a short distance behind the MIS 2 maximum ice
advance fronts, with additional evidence of significantly
thinner glaciers merely lining the valley floors (Delmas,
2009). In the Ariege catchment (Fig. 11.4), the ice field
was dismembered into disjunct trunk and tributary valley
glaciers, with, for example, the Courbiere, Vicdessos and