1The sediments of many mountain areas covered by Pleistocene glaciers are over‑consolidated by the ice load. Palaeoglacial conditions are in this Thesis deduced from site investigations and geomorphological features of the glacial extent. The last and the latter are used as a framework for hydro‑mechanical flow simulations in the valley glacier of Andorra and the lower Isère glaciated valley (Fig. 1a). From the effects observed within the sedimentary record in these glaciated valleys, a reliable reconstruction of palaeo‑glaciers thickness has been possible to do where former ice‑tonges overlayed porous aquifers. Subglacial water pressure limited subglacial till formation and subglacial erosion. The in-situ measurements using pressurometer tests recorded minimum over-consolidation values from former high pore-water pressures. The last is interpreted as a consequence of buoyant glaciers overlaying its bed. The study of a modern analogue, the Hansbreen sub‑polar glacier in SW Spitzbergen (Fig. 1a), permits to set up of a conceptual model allowing a better understanding of the studied valley glaciers' former glaciology, being these assimilable to a polythermal glacier type.

2The glaciolacustrine deposits from an ice‑dammed lake in Andorra (La Massana) allow for a high resolution framework (Fig. 2) of the local deglaciation patterns between GS‑9 (40 ka b2k) and H1 (17.5 ka b2k). The final deglaciation is characterised by relict cirque glaciers disappearing during GS‑1, linked to a general rise in local river base levels until the Holocene Optimum. Using geochemical data (Al, Ti, Ca, K, P), a basic limnological study and palaeotemperatures permit us to provide a palaeoenvironmental interpretation in Andorra between GS‑7 to GS‑2a, in addition to the last four unreported inland δ13C cycles linked to low water levels in the La Massana palaeolake result from AMS dates on bulk carbon samples from the Andorran valleys (Fig. 1b). At the beginning of each cycle, enhanced δ¹³C bulk carbon values are found (> ‑23‰), a proxy (among others) of abrupt shifting from Type‑C3 to Type‑C4 vegetation. The beginning of the LGM and Heinrich events H3 and H2 were marked by enhanced δ¹³C values interpreted as a product of strong climate change that boosted aridity in the area. The retrieval period towards δ¹³C depleted values (< ‑23‰) spans 4,500±500 years (Fig. 1b). However, from the palaeoenvironmental data of the Pyrenees, a deglaciation phase started sooner than in The Last Termination, and aridity affected the size of the glaciers from GS‑10 resetting the valley glacier conditions ahead of GI‑7. The last affected mainly prominent glacier-ice transfluence pass in the Pyrenees, although the extension of the ice tongues at both sides of the mountain belt between GI-9 and GI-7, also at GS‑5.2 and GI‑5.1. During and posteriorly Heinrich event H4, aridity did not favour the development of valley glaciers in Andorra until stadial GS‑7; however, glaciers progressed during the following Heinrich event H3 (Fig. 2). The beginnings of H3 and H2 events were relatively dry, but moister during their second half, allowing glaciers to advance in Andorra (Fig. 2). The first occurrence of sediments coming from the motion of a temperate‑base glacier type in Andorra is from stadial GI‑3 (Fig. 3). Moist conditions suddenly stopped at the beginning of Heinrich event two (H2) and returned back at its end when local glaciers advance (Fig. 2). The cited moisture behaviour during H2 on the southern slopes of the Pyrenees were opposite to the wet‑to‑dry conditions described in NW Iberia for H1. Conversely, the H1 event had a wet‑to‑dry structure coast to coast of Iberia, including the Valira valleys (Andorra).

3Abrupt glacier advances and quick glacier recession in the Valira valleys are interpreted as surges from a mass‑imbalanced glacier (Fig. 3). Two kinds of surge events are distinguished from available data in the Pyrenees, those surges produced by overfed ice tongues fed by short‑lived cold spells (in GS‑2b and H1) and those surges produced by buoyant melting ice tongues within the glaciated valley (between stadials GS‑5/GI‑4, stadials GI‑3/GS‑3 and stadials GS‑2.1b/GS‑2.1a) during a warmer and moister climate. A concurrence of winter solar rate increase and the polythermal‑type glaciers occur during GI‑3 & GS‑5.1 and interdigitated with wet‑ice type oversaturated surging glaciers until GS‑2a matching with the increase of summer insolation; however, the snow‑overfed glacier surges correlates with the decreasing of winter insolation during GS‑2b.

4In Iberia, the LGM would be placed between 23‑17.5 ka and the spread of glaciers (Fig. 3), linked to both wet westerlies from the North Atlantic and moisture supply of Mediterranean influence, reaching almost the eastern side of the Central Pyrenees, leading us to suppose an NW‑NE seesaw climatic relationship across Iberia almost since GS‑5.1 influencing the occurrence of glacier (Fig. 3).

5Like most southern‑faced Pyrenees, the Last Maximum Ice Extent and the Global LGM did not concur in places experiencing multiphase glacier advances. Indeed, both extreme behaviours do not correspond to the same type of glaciers. Polar‑type glaciers in NW Iberia are invoked in this research, while tempered or polythermal were frequent in NE Iberia (Fig. 3). The northern fringe of the Iberia Peninsula shows a transition from Multiphase glacier advances (SE Pyrenees) to Single‑like glacier advances (NW Iberia). The last probably was due to a change over from the presence of temperate‑based glaciers to cold‑based glaciers. However, in NW Iberia glacier stability decreased through time while sensibility to global changes increase (Fig. 3). By classifying the glacial phases of the northern Iberian Peninsula fringe, four glacial phases arise for the last glacial cycle (LGC):

6An early LGC starting at MIS 5d having a recessional period during MIS 5c. Cold‑type glaciers are expected to have existed in some of the extreme NW of the Iberian mountains until Termination‑I (monoglacial behaviour).

7The Last Maximum Ice Extent occurred mainly during MIS 5a – MIS 4. An asymmetrical glacier recession during MIS 3 was related to an eastward aridity increase.

8Significant glacier fluctuations during the MIS 3 – MIS 2 hinge, the appraisal of temperated‑ polythermal type of glaciers accompanied by a generalised moisture increase entailing valley glaciers to surge.

9Side‑to‑side mountain range‑scale deglaciation dissymmetry during MIS 2. The widespread expansion of tempered‑polythermal type glaciers during the LGM period and a proportional expansion to the available moisture during Termination‑I.

10Unravelling the aforecited glacial phases of the research allows for an unexpected mapping of the SW continental Europe concerning part or all of the LGC outlined above (Fig. 4).

• Type 1 – Areas where glaciers were prevalent during MIS 2, like the Iberian Central System, the NW and S French Massif Central, the NW Jura and the maritime Alps (MIS 2 Type glaciation).

• Type 2 – Areas having a far-flung end moraine produced in a previous glacial phase (MIS 6) or at the beginning of the LGC (MIS 5e) showing stability until the MIS 2, as for ice caps/fields from the southern half of the Galicia mountains (Monoglacial Type).

• Type 3 – Areas of pseudo-Pleniglacial or apparent-Pleniglacial condition, despite previous glacier recessions phases (albeit challenging to identify), as in most of the northern slope of the Pyrenees except the Ariège valley.

• Type 4 – Areas of multiphase glacier advances, like most of the southern slope of the Pyrenees and Ariège, most of the Cantabrian Mountains, the half north of the Galicia mountains, the High Atlas, Sierra Nevada and the SW French Massif Central, the western Alps and the Vosges experiencing a multifold glaciation..

• Type 5 – Areas having experienced glaciers from the LGM until Termination-I, like the northern Iberian range and Sanabria in Iberia (Termination-I Type glaciation). Nevertheless, other mountain ranges have a Type 5 glaciation scenario, like the southern Black Forest in Germany.

11The boundary between these glacial characteristics will change in the next future when the number of works on this topic increase, especially in understudied areas. Nevertheless, the figures 3 and 4 represent the first attempt to correlate the asymmetrical glacial facies of the westernmost European massifs during the LGC.

12Fig. 1: Studied area and δ¹³C cycles from Andorra during MIS 2

13(A) Extended studied areas away from Andorra, like SW Spitzbergen as a modern analogue, are referred to as distant glaciated valleys like the SW Alps and the Cantabrian Mountains until the Pyrenees. See Turu et al. (2007, 2017, 2023), Jalut et al. (2010), Serrano et al. (2013), and Ménard (2014). (B) The δ¹³C cycles from Andorra during MIS 2. Isotope fractionation data came from bulk carbon content in Andorra sediments. Four decreasing δ¹³C cycles of 4,500 ± 500 years each. Age intervals of these cycles are: 30,668-27,303 b2k; 25,895-21,650 b2k; 20,988-16,989 b2k and 16,308-13,155 b2k.

14Fig. 2: The most representative dates related to glacier advances or retreats with their two- sigma confidence intervals. Ages in cal BP.

15High probability match when a couple of one-sigma distributions coincide. Short horizontal lines below the X-axis represent only one sigma bar. In colour, summation distribution area for H3, H2, LGM and H1 in Andorra. Three dates related to glacial activity cover the period between GS4 – GS3. Vertical red shadow bars show the Low Stand System Tracts (LST) timespan identified in La Massana palaeolake. At the same time, the thin blue lines represent stratigraphical unconformities related to glacier advances. These advances concern the main valley of Andorra and its tributary (Till 1 to Till 7). Depositional Sequences (DS) from La Massana palaeolake are highlighted, as stratigraphic boundaries (SB) and subdivisions (a, b, c) between LST periods, following Turu et al. (2017) and Turu (2018).

16Fig. 3: Proposed correlation for the last glacial cycle between the principal mountain ranges of southwestern Europe (sites in Figure 4).

17In blue glaciers and cold climate (light blue, expected cold climate). In brown glacier recession and arid climate (light brown, expected arid climate). In red glacier recession and temperate climate (light red, expected temperate climate). In green, expected glacier surges and moist climate. Yellow lines, indicate climatic trends across glaciated massifs (discontinuous yellow lines, expected). Note a pivotal climatic trend at 32 ka in the Pyrenees, starting at 45 ka in west Iberia. The climatic trend may also progress through the continent (to the east) until the Alps end at the LGM period. The aforecited climatic trend starting at 32 ka may correspond to the generañ changeover of the glacier type from cold-ice base to tempered-ice base glaciers.

18Fig. 4: LGC extension and glaciation Types

19MIS 2 glaciation (Type 1), Monoglacial (Type 2), Pseudo-Pleniglacial (Type 3), Multiphase glaciation (Type 4), Termination-I (T I.) glaciation (Type 5). Green arrows in the small window may indicate mobile boundaries, observed in La Massana palaeolake (Turu, 2018). Green arrows in the large window indicate singular climatic points identified from the WRF model (Ludwig et al., 2018), evidence of Termination I glaciation in the Black Forest (Hofmann et al., 2022) and Type 4 – Type 1 boundary close to the Maritime Alps (Rosique, 2004). Blue lines represent Termination I (T I) or Type 5 glaciation. Sites: 1‑Camaaño, 2‑Serra do Xistral, 3‑Sextas, 4‑Serra da Estrela, 5‑ Queixa‑Invernadoiro, 6a‑Oribio Mounts, 6b‑O Courel, 7‑Sanabria, 8a‑Castro Lake (Villaseca de Laciana), 8b‑Laguna‑A‑Lucenza, 8c‑Laguna Grande de Neila, 9‑Bejar massif, 10‑Gredos massif, 11‑Brañagallones, 12a‑, Porma/Lillo, 12b‑Redipollos, 13a‑Comeyas' polje, 13b‑Hayéu l'Osu cave, 14‑Campo Mayor, 15‑Bibei, 16‑Guadarrama, 17‑Hoya Pelada, 18a‑Ansón, 18b‑Trueba, 19‑Laguna Grande (Neila), 20‑Sierra Cebollera, 21‑Villanúa(Castiello de Jaca, 22‑ Serra Faro de Avión, 23‑Gavin, 24a‑Llinàs de Broto, 24b‑Viu, 25a‑Soum d'Ech, 25b‑Lourdes and Monge, 26‑Garbarnie, 27‑Pineta (Lago), 28‑Larri hanged valley, 29‑Salinas de Sin, 30‑Cotiella, 31‑Turbon, 32‑Barbazán, Garonne paleolake, 33‑Joèu, 34‑Têt – La Borde, 35‑Segre‑ TQ4 (Organyà), 36‑Tournac, 37‑Niaux cave, 38‑Roc del Quer, 39‑Arànser/La Llosa/Duran, 40‑Malniu, 41‑Querol/Puigcerda, 42‑Tamboreurets, 43‑Cantal, 44‑Lugarde (Cantal), 45‑Mont‑Dore, 46‑ Couze Chambon (Auvernat), 47‑Aubrac, 48a‑Isère‑Grenoble, 48b‑Trieves/Avignonet, 49‑ Montagne de Bange, 50‑Genève, 51‑Ramble de Chablais, 52‑Biolet‑Orjulaz, 53‑Vosges massif, 54‑Finsterhennen, 55‑Maritime Alps, 56‑Unterangerberg. Lambert Azimutal Equal Area Projection ERTS89.