Configuration of the Scandinavian Ice Sheet in southwestern Norway during the Younger Dryas

The configuration of the Scandinavian


Introduction
Ice sheets are part of important climate feedback mechanisms that lead to sharp latitudinal contrasts in temperature.Despite their significant role in the Earth's climate system, our understanding of their configuration -particularly their height -through time and space is still limited (Hughes et al., 2016, Dalton et al., 2020, Clark et al., 2022).Geological observations of past ice-sheet configuration remain critical for validating and improving climate modeling (e.g., Sommers et al., 2021), for elucidating glacio-isostatic adjustment of the crust that is critical for understanding the causes of relative sealevel change (Gowan et al., 2021;Fjeldskaar & Amantov, 2018), for constraining post-glacial immigration of flora and fauna (Alsos et al., 2022), and for predicting spatial patterns of future sea-level changes (e.g., Caron et al., 2018).
The history of Scandinavian Ice Sheet margin fluctuations during the last deglaciation is reasonably well known in coastal southwestern Norway (Fig. 1).The Bergen area was deglaciated at around 14 ka (Mangerud et al., 2017), and remained ice-free for 2000 years with an ice-sheet position at least 40 km east of Bergen (Mangerud et al., 2019).Subsequently, during the Younger Dryas, the ice sheet re-advanced and built the Herdla Moraine close to the outer coast ̴ 11.6 ka (Mangerud et al., 2016; Fig. 2).The Bergen area became ice free again ~11.5 ka as the ice front retreated rapidly inland in response to Holocene warming (Mangerud et al, 2019).The timing and lateral extent of the Younger Dryas re-advance is better documented in the Bergen-Hardanger area than any other place around the Scandinavian Ice Sheet.Lateral moraines also show the slope of the ice surface along Hardangerfjorden (Follestad, 1972;Mangerud et al., 2013).
However, the steepness of the ice margin, and the height of the ice sheet during this event, remains largely under-constrained in the topographically complex area around Bergen, i.e., between the two large fjords Hardangerfjorden and Sognefjorden.What was the Younger Dryas ice-sheet profile in the Bergen area and how might this guide knowledge of the shape of the Scandinavian Ice Sheet during the Younger Dryas?
The main aim of this study has been to resolve this question by 10 Be-exposure-dating of erratic boulders found on the highest mountains close to the Younger Dryas ice margin, i.e., mountains around the city of Bergen.We also include 10 Be-ages from erratics on higher mountain peaks farther inland to determine the height of the ice surface.In other projects, 10 Be ages had been obtained from lower elevations in the area overrun by the Younger Dryas re-advance.We include these results, as well as new radiocarbon ages from a key stratigraphic section, because they improve our understanding of inheritance of 10 Be in our samples and they also represent a first step in quantifying glacial erosion in such a landscape.

Background
The Scandinavian Ice Sheet covered most, if not all, coastal mountains in western Norway during the Last Glacial Maximum (Hughes et al., 2016;Regnéll et al., 2021).Ice flowed across the Bergen area from the higher elevations to its east and advanced onto the continental shelf in the North Sea.Once offshore, this inland ice was confluent with the north-flowing Norwegian Channel Ice Stream, which terminated at the continental shelf break on the southern flank of the Norwegian Sea ~250 km down flowline from Bergen (Fig. 1) (Sejrup et al., 2003).In this section, we synthesize previously published age constraints to generate two alternative exposure histories for the highest mountains in the Bergen area (Fig. 3) from which to compare our suite of new 10 Be ages.We outline this history in three stages: Stage 1 -pre-Younger Dryas deglaciation Sejrup & Hjelstuen (2022) depict the recession of Norwegian Channel Ice Stream west of Bergen between 19 and 18 ka.Despite this relatively early post-LGM recession, available data indicate that the ice margin did not recede inside the outermost islands in this area until between ~15 and ~14 ka (Mangerud et al., 2017).At the village of Blomvåg in Øygarden (Fig. 2), basal marine sediments have radiocarbon ages of 14.1 ± 0.2 cal ka BP [re-calibrated using Marine20 (Brendryen et al., 2020)] supported by 10 Be ages from nearby erratic boulders that average 14.0 ± 1.4 ka (Mangerud et al., 2016).
Following deglaciation of the western coastline, the ice front continued to retreat at least ~40 km inland (east) of the Bergen area.Near Bergen, the oldest radiocarbon ages from re-worked shells found in or below a till associated with the Younger Dryas re-advance cluster at ~14.0 cal ka BP (Mangerud et al., 2016).

Stage 2 -Younger Dryas re-advance
The youngest among about 90 radiocarbon ages on shells re-worked into a till deposited during the Younger Dryas re-advance the Bergen-Hardangerfjorden area date to ~12.5 cal ka BP (Mangerud et al., 2016).This indicates that the ice sheet advanced across the Bergen area shortly after 12.5 cal ka BP.Relative sea-level curves showing a 10 m rise that culminated at the end of Younger Dryas suggest that the re-expansion of the Ice Sheet started during the Allerød interstadial period (Lohne et al., 2007).The ice margin continued to expand westward to a position ~12 km down flowline from Bergen (Fig. 2).There are multiple lines of evidence indicating that the re-advance attained its maximum extent shortly before the transition to the Holocene; in some areas, Younger Dryas maximum phase sediments were deposited stratigraphically well above the Vedde Ash (12.1 cal ka) (Bondevik & Mangerud, 2002;Lohne et al., 2004), and in other areas, ice-marginal deposits rest on in situ shells dating to ~12 cal ka (Mangerud et al. 2016).
How thick the Younger Dryas ice was around the city of Bergen is unknown and allows for two different exposure histories.If adjacent summits were not re-occupied during the Younger Dryas, then exposure ages of perched boulders should be ~14 ka (and with inheritance, potentially older; Fig. 3).If, on the other hand, summits were re-occupied during the Younger Dryas, then boulders without inheritance should have exposure ages of ~11.5 ka (Fig. 3).

Stage 3 -Recession from the Younger Dryas
The timing of ice recession has been dated precisely from plotting the elevation of raised ice-marginal deltas in a well-dated "master" shore-line diagram (Mangerud et al., 2019), indicating that the Bergen area was deglaciated between 11.5 and 11.4 cal ka BP.

Be dating
We sampled bedrock surfaces and perched boulders on or near the highest summits in the Bergen area, which lies slightly within the Younger Dryas ice extent (Fig. 4).High topographic areas below former ice sheets are challenging landscapes for cosmogenic-nuclide exposure dating, because cosmogenicnuclide accumulation can survive multiple glacial cycles due to minimal glacial erosion (e.g., Brook et al., 1996;England, 1999;Briner et al., 2005;Fabel et al., 2006;Goehring et al., 2008).Thus, our sampling strategy includes sampling many perched erratic boulders.Our bedrock samples are from various elevations, allowing assessment of the relative magnitude of glacial erosion efficiency, and hence inheritance and its potential influence on boulder age distributions.Our dataset consists of 22 10 Be ages from boulders from 49 to 677 m a.s.l., and 11 10 Be ages from bedrock surfaces from 90 to 477 m a.s.l.(Fig. 5).At all sites, sampled boulders rested directly on bedrock that exhibited evidence of erosive subglacial conditions.To constrain the inland height of the ice sheet, we supplemented the Bergen dataset with five samples from erratic boulders from the mountain Tarven in a tributary valley to Sognefjorden, which is a high mountain near the center of the ice sheet that serves the purpose of constraining ice-sheet thickness.Tarven is a logistically feasible location with suitable boulders that are located 1234-1620 m a.s.l.~120 km up flowline and serve as a test of inland ice-sheet height during the Younger Dryas (Fig. 6).We collected samples with an angle grinder, hammer and chisel.Subsequent sample processing took place in two labs, one in Buffalo, USA and one in Glasgow, UK.Both labs follow similar quartz purification and beryllium isolation protocols (Kohl & Nishiizumi, 1992;Corbett et al., 2016).Samples processed in Buffalo had beryllium ratios measured at the Center for Accelerator Mass Spectrometry in Livermore, California, US.BER samples were processed and measured at the Scottish Universities Environmental Research Centre, East Kilbride, UK.Table 1 includes metadata about AMS standards and other sample-specific information.
We calculate all 10 Be ages using the 10 Be production rate of Goehring et al. (2011Goehring et al. ( , 2012) ) and version 3 of the online exposure age calculator (https://hess.ess.washington.edu/;Balco et al. 2008).This applies both to ages we report here for the first time and those previously published.We use the "Goehring rate" and the Lal/Stone scaling scheme (referred to hereafter as Lm; Lal, 1991;Stone, 2000) to be consistent with calculations used at other sites in western Norway.The "Goehring rate" is an average rate calibrated locally in western Norway at two sites, one of which is a Younger Dryas moraine with independent radiocarbon age control, and one is a rock fall deposit at 6.0 cal ka BP that killed a tree that was radiocarbon dated (Nesje, 2002).Other production rate choices include the "Scandinavian rate" of Stroeven et al. (2015) or the "Arctic rate" of Young et al. (2013) or the "global rate" from the CRONUS project (this is the default rate in the age calculator v3), among others.We favor the Goehring rate due to the proximity of the calibration sites, but note that it is higher than other rates, and thus yields the youngest ages among the production rate choices [e.g., ~6 % younger than the "global" rate of Borchers et al. (2016) and 2-3 % younger than the "Scandinavian" rate of Stroeven et al. (2015)].Sample density used is 2.65 grams/cc, surface erosion is zero, atmospheric model is "std."Ages calculated using production rate of Goehring et al (2012) and "Lm" scaling.See text for link to web calculator.The lithology of most boulder samples is gneiss; "not local" indicaters field evaluation that lithology does not appear to be locally derived.The bedrock lithologies can be obtained here: https://geo.ngu.no/kart/berggrunn_mobil/. None of the bedrock/ boulder surfaces sampled show striations, the degree of weathering/erosion is too advanced, i.e. quartz veins typically protrude 1-3 cm above the gneiss surface.

Radiocarbon dating
We exploited a temporary sediment section in the Bergen valley (informal site name Fjøsangerveien) at the entrance of a new tunnel for a light rail (60.358213°N, 5.340767° E) into the mountain Løvstakken.An excavation (Figs. 5 & 7) was created from the surface to bedrock exposing up to 8 m of sediments that we subdivide into three units (lower, middle, upper).The lower unit, which rests directly on bedrock, is a compact diamicton dominated by subangular pebbles and boulders.This is draped by a middle unit (22-25 m a.s.l.) that is a 2-3 m-thick clay-rich, ice-drop diamicton that can be followed laterally throughout the exposure.It is somewhat deformed and tectonized.In places, it contains high concentrations of in situ marine mollusc shells, mainly Chlamys islandica.Other species include Mytilus edulis, which requires influx of temperate Atlantic water and a seasonal ice-free zone to thrive (Mangerud & Svendsen, 2017).The shell-bearing sediments are covered by a meter-thick dark gray layer that mostly consists of silt and clay and is void of shells.The section is capped by the upper unit, which is a massive and compact diamicton up to 4 meters thick and densely packed with subangular gravel particles, stones, and boulders.Rooted in this diamicton are large granitic boulders that bear evidence of glacial transportation.
Figure 7.The sediment section temporarily exposed at Fjøsangerveien.Inset shows eight calibrated radiocarbon ages on marine material collected from this outcrop (circles) and two calibrated radiocarbon ages from marine shells re-worked into till exposed in a sediment section nearby (diamonds; Mangerud et al., 2016).Ages reported in Table 2; photo of Fjøsangerveien section in Fig. 5.
We obtained eight radiocarbon ages from the shell-bearing unit, including five mollusc shells and three samples from different parts of a growth layer of a calcareous algae (Lithohamnion sp) enclosing a pebble.The samples were from different parts of the exposed strata, but not in stratigraphic order.
Radiocarbon ages were measured at the Radiocarbon Dating Laboratory, Lund University, Sweden. Results

Be ages
The 22 new 10 Be ages of boulders in the Bergen area range from 10.4 ± 0.4 to 16.2 ± 0.3 ka with an average age of 12.4 ± 1.4 ka (Table 1; Fig. 4).On the Ulriken summit, the highest mountain in the Bergen area, where the majority (15) of our boulder samples are from, the ages range from 10.9 ± 0.2 to 16.2 ± 0.3 ka and average 12.8 ± 1.6 ka.The elevations of the Ulriken samples range from 401 to 677 m a.s.l.Eliminating six boulders (five too old, one too young) following Chi-squared analysis (using version 3 of the online calculator; https://hess.ess.washington.edu/)results in a mean of 11.8 ± 0.9 ka of the remaining nine Ulriken boulders.There is no relationship between sample age and elevation (r 2 = 0.17).
The 11 bedrock samples have apparent 10 Be ages that range from 12.70.4 to 23.4 ± 0.9 ka, of which seven samples pre-date 15 ka.All surfaces exhibit evidence of glacial erosion from sliding ice.
The elevations of the bedrock samples range from 90 to 477 m a.s.l.Samples with apparent ages older than 15 ka lie between 90 and 200 m a.s.l.
Five boulders from our inland site are from elevations spanning 1234 to 1620 m a.s.l., and 10 Be ages range between 9.4 ± 0.2 and 14.6 ± 0.3 ka (Fig. 6).Eliminating two boulders (one too old, one too young) following Chi-squared analysis reported on version 3 of the online calculator (https://hess.ess. washington.edu/) results in a mean of 11.0 ± 0.2 ka for the remaining three boulders.

Radiocarbon ages
The sediment section Fjøsangerveien is interpreted as containing a lower and upper till with a glaciomarine unit between (Fig. 7).Four shell dates from the glaciomarine unit yield ages of around 14,000 cal kyr BP whereas one yields a younger age of around 13,200 cal kyr BP (Table 2).
The three radiocarbon ages from the clast coated with lithohamnium algae yield identical ages of around 12,800 yrs BP.

Beryllium-10 ages, inheritance and glacial erosion
Prior results demonstrate that valleys in the Bergen area were finally deglaciated at ~11.5 ka, following the ice sheet re-advance to the Herdla end moraine system during the Younger Dryas (Mangerud et al., 2019).Thus, 10 Be ages from bedrock surfaces in the lower elevations near Bergen should date to ~11.5 ka if they have experienced sufficient glacial erosion, on the order of 2-3 m (Briner et al., 2016), during the preceding glacial overriding.Alternatively, if surfaces were minimally eroded during the Younger Dryas advance, they should date to ~14.5 ka, their first deglaciation following the longer overriding during the Last Glacial Maximum.In fact, most 10 Be ages from bedrock surfaces, all of them glacially sculpted, pre-date not only 11.5 ka, but also 14.5 ka, by thousands of years (Fig. 4).This stands in clear contradiction to our new radiocarbon ages from the Fjøsangerveien site, which lies close to our nearby bedrock site with ages of ~22 ka (Fig. 4).We interpret these anomalously high bedrock ages as being influenced by varying amounts of isotopic inheritance, and thus consider them as apparent ages only.
We find it plausible that the glacier erosion was less than 2-3 m during this last ice-sheet advance that probably covered Bergen for only a few hundred years during the late Younger Dryas.
Many of the apparent 10 Be ages pre-date ~14 ka, implying inefficient erosion also during the Last Glacial Maximum, which is unexpected.Five samples in valley bottom locations range between ~16 and ~23 ka, indicating relatively inefficient subglacial erosion during the Last Glacial Maximum even in the valley bottom.This valley bottom inheritance may relate to the protected nature of these valleys as regional ice flow for long periods was mostly across, not along, the investigated valleys (Wirsig et al., 2017).This is supported by the preserved last interglacial sediments at Fjøsanger (Mangerud et al., 1981;Fig. 4), proving that at least parts of these valley bottoms were not eroded during the last glaciation.
The remaining bedrock samples, from surfaces on summit ridges west of Bergen, have apparent 10 Be ages that are equally old (~23 ka) and others that range between ~13 and ~15 ka.
Were the summits near Bergen ice-covered during the Younger Dryas?
The mean (~11.8 ± 0.9 ka) of the 9 accepted 10 Be ages from boulders perched on the bedrock surfaces on Ulriken's mountain plateau is statistically inconsistent with an age of ~14 ka for the deglaciation of the summit.Our new radiocarbon ages confirm previous dating results indicating that the Bergen valley became ice free in the late Bølling, a little before 14,000 yrs ago and remained ice free until sometime after 12,500 yrs BP when the ice sheet advanced towards the Herdla Moraine (Mangerud et al., 2016).
Taken together, these results show that the highest summits adjacent to Bergen were ice covered for some period during the Younger Dryas.We selected large boulders resting in stable positions perched directly on bedrock surfaces, and there is little to no sediment cover on the uplands surrounding Bergen.Shielding by snow would require snowpack thicknesses and snow densities that are inconsistent with meteorological data and with observations of large boulders being windswept even during in the snowiest months of the year.We therefore find no plausible explanation for boulder ages that are ~2000 years younger than the deglaciation age if the mountain peaks remained ice free during the Younger Dryas.The seven additional boulders from adjacent summits in the Bergen area similarly post-date ~14 ka.We thus interpret the distribution of 10 Be ages as supporting the view that all summits around Bergen were ice covered during the Younger Dryas.We note that most boulders that failed the Chi-squared test pre-date the mean age, which is expected given the levels of inheritance found in bedrock samples.
Bolstering our interpretation of summits being occupied by ice during the Younger Dryas is the comparison of boulder age distributions beyond vs. within the mapped Younger Dryas margin.
Twelve 10 Be ages from boulders west of the Younger Dryas ice margin average 14.0 ± 1.4 ka (Mangerud et al., 2017).Using the same Chi-squared criteria as above to exclude four outliers (three too old, one too young), the remaining eight boulder samples have a mean age of 13.9 ± 0.1 ka (Fig. 8).This mean age is statistically distinct from the mean age 12.0 ± 0.6 ka of the boulders within the Younger Dryas margin, all from summit positions (Fig. 8).We note that the boulders west of the Younger Dryas margin are statistically overlapping with the independent 14 C age control on the deglaciation of that area of 14.1 ± 0.2 cal ka BP, indicating that the 10 Be chronometer (and production rate that we employ) is working well in that landscape.

Ice sheet or local ice?
These dating results alone cannot distinguish whether the summits were covered by the Scandinavian Ice Sheet or simply by local glaciers.The Younger Dryas equilibrium-line altitude has been estimated in middle Hardangerfjorden to be at ~1300 m a.s.l.(Mangerud et al., 2013;Regnéll et al., 2021).
This elevation is well above the Ulriken plateau (~600 m a.s.l.).However, the Younger Dryas snowline lowered considerably seaward.At the island Stord, ~50 km south of Bergen and along a similar uplift isobase (Mangerud et al., 2016), the Younger Dryas equilibrium-line altitude is estimated to be only ~500 m a.s.l., which is below the elevation of the Ulriken plateau (~600 m a.s.l.).Thus, based on an assessment of the glaciation limit during the Younger Dryas it is not unreasonable to assume that the Ulriken plateau hosted a local ice cap that merged with the adjacent ice sheet, as was the case for the ~1200 m a.s.l.Ulvanosa massif in outer Hardangerfjorden (Regnéll et al., 2021) and in other coastal uplands in coastal western Norway (Mangerud, 2023).A compilation of hundreds of glacial striae in the Bergen area ( Mangerud et al., 2019), on the other hand, shows a pattern of mostly uniform ice flow from inland to coast, and does not display a radial pattern expected to be produced by a local ice source on Ulriken or other summits.Nor are there any obvious local moraines that could indicate local ice cap glaciation.
We are nevertheless open to the possibility that a local ice cap formed on top of Ulriken at some time, but the consistent west-oriented striae shows that the entire mountain plateau was overrun by the Scandinavian ice Sheet during its most recent episode of ice cover.

Ice-sheet geometry during the Younger Dryas
We next discuss the ice-sheet profile geometry of the Younger Dryas ice margin and farther inland.
Previous studies that reconstruct ice-sheet height in this area have focused mainly on Hardangerfjorden (Mangerud et al., 2013;Aakesson et al., 2018Aakesson et al., , 2020;;Regnéll et al., 2021).Here, ice-surface slopes are comparable to the contemporary ice-sheet outlet glacier Sermeq Kajulleq (Jakobshavn Isbrae), Greenland.We reconstructed a surface slope of the Younger Dryas ice-marginal area near Bergen, consistent with our dating results, to rise above the studied summits (Fig. 9).This profile is only slightly steeper than the ice-sheet profile along Hardangerfjorden based on mapping of lateral moraines.
This implies that despite ice being channeled through that ~5 km-wide fjord system, the ice sheet still had a basal shear stress comparable to what it had while flowing across the more rugged terrain near Bergen.We also note that our ice-sheet profile is steeper and higher than that of Fjeldskaar & Amantov (2018), which also depicts ice that is too thin at the Tarven site.
We leverage the two-dimensional ice-margin profile reconstruction with previously published information to reconstruct a surface contour map of the westernmost part of the Scandinavian Ice Sheet during the Younger Dryas (Fig. 9).The map is a slight modification and an extension of the maps by Hamborg & Mangerud (1981) and Fareth (1972) and it is based on the following types of observations: The pattern of the Younger Dryas end moraines, the positions and slopes of lateral moraines and the directions of glacial striae considered to represent Younger Dryas ice flow, with glaciologically plausible interpretations, are in line with the general glacial geology of the area.
Much of the map's contours are interpolated between sites with height information, and we see the reconstruction as the best available field-based map that can be compared with ice-sheet model simulations.Along Hardangerfjorden, lateral moraines are mapped from sea level and up to 1000 m a.s.l.(Follestad, 1972;Regnéll et al., 2021), constraining an ice-profile that is similar to the surface profile of Jakobshavn Isbrae (Mangerud et al. 2013).Along Samnangerfjorden (Fig. 9), lateral moraines are traced up to 600 m a.s.l. and along Fanafjorden (Fig. 9) lateral moraines are traced up to 300 m a.s.l.(Hamborg & Mangerud, 1981;Mangerud et al., 2019;Aarseth & Mangerud, 1974).We interpolated and extrapolated these profiles and altitudes along the ice margin.Between Samnangerfjorden and Fensfjorden (Fig. 9), the Younger Dryas end moraines show a wide and quite regular lobe form, which is reflected in the directions of the glacial striae, including striae on Ulriken and other mountains, pointing at right angles to the moraines (Mangerud et al., 2019;Aarseth and Mangerud, 1974).
Well inland, our new 10 Be ages from high elevations post-date the Younger Dryas.Previously published 10 Be ages from boulders on adjacent mountain uplands are 11.2 ka (1354 m a.s.l.; 60.9042°, 7.2764°) and 10.5 ka (1569 m a.s.l.; 60.9851°, 7.2817°) (Andersen et al. 2018); these ages have been re-calculated to be consistent with the calculation method used here.Collectively, these data also confirm a Younger Dryas ice height above these elevations.More broadly, we know of no robust evidence that would demand the height of the ice sheet during the Younger Dryas to be lower than southern Norway's topography.Rather, if one uses the modern-day Greenland Ice Sheet as a guide, we suggest that the height of the ice-sheet surface during the Younger Dryas could likely have been well above southern Norway's topography.To investigate this further, we selected a transect in southern Greenland that has many commonalities with southern Norway during the Younger Dryas:   (bottom).Each ice-sheet surface is drawn in blue and underlying topography in brown.Note that the age of the Younger Dryas maximum phase may be slightly different in the west and east ends of this transect, but this does not significantly impact the overall shape of this ice-dome reconstruction.Greenland ice surface and bed from layers within the QGreenland QGIS package (Moon et al., 2022;Morlighem et al. 2017).Topography data from Norway are from norgeskart.no.The thick line portion of the Scandinavian Ice Sheet profile is our reconstruction shown in Figure 9.

Figure 1 .
Figure 1.Map of southern Norway and the North Sea showing the extent of the Scandinavian Ice Sheet during the Last Glacial Maximum (LGM) and Younger Dryas (YD).The dashed white line shows the assumed Younger Dryas ice divide.

Figure 2 .
Figure 2. Map showing the extent of the Younger Dryas ice limit in the Bergen area of southwestern Norway (fromMangerud et al., 2016), along with previously published 10Be ages (re-calculated here) fromMangerud et al. (2017).13 ka limit fromMangerud et al. (2019).The topography of the A-A' line is shown in Fig.8.

Figure 3 .
Figure 3. Ice-sheet cover scenarios for mountain summits in the Bergen area, depicting locally derived glacially transported boulders (gray, higher possibility of having 10 Be inheritance) and long-transported glacial boulders (black, lower possibility of having 10 Be inheritance).(A) Ice free at 14 ka.(B) Ice free during the Younger Dryas.(C) Ice covered during the Younger Dryas.

Figure 4 .
Figure 4. Map of Bergen region showing 10 Be ages of bedrock (black diamonds) and perched boulders (red circles).Fv = Fjøsangerveien site.

Figure 5 .
Figure 5. Selected photos of field locations and samples.

Figure 6 .
Figure 6.Map showing location of Tarven and 10 Be ages (in ka).Map location shown in Figure 1.

Figure 8 .
Figure 8. (A) Topographic cross-section across the Younger Dryas ice limit and a simplified depiction of the Younger Dryas ice-sheet profile (profile location shown in Fig. 2); sample localities within and beyond the Younger Dryas ice limit are shown.(B) 10 Be age distributions within (right) and beyond (left) the Younger Dryas ice extent; thin lines are statistical outliers (see text).

Figure 9 .
Figure 9. Ice-surface geometry of the Younger Dryas ice configuration in southwestern Norway.Surface contour labels in meters; Fe = Fensfjorden; Sa = Samnangerfjorden; Fn = Fanafjorden; St = Stord.Small circles show dated samples near the head of Sognefjorden on the mountain Tarven.Bottom panel shows relevant ice-sheet profiles; Solid blue line is reconstructed from Hardangerfjorden(Mangerud et al., 2013), dotted is modeled fromFjeldskaar & Amantov (2018); Dashed is measured from the Greenland Ice Sheet at Sermeq Kujalleq.Bergen ice profile shown in black is the reconstruction from this work.

Figure 10 .
Figure 10.Comparison of southernGreenland (top)  with southern Norway (bottom).Each ice-sheet surface is drawn in blue and underlying topography in brown.Note that the age of the Younger Dryas maximum phase may be slightly different in the west and east ends of this transect, but this does not significantly impact the overall shape of this ice-dome reconstruction.Greenland ice surface and bed from layers within the QGreenland QGIS package(Moon et  al., 2022;Morlighem et al. 2017).Topography data from Norway are from norgeskart.no.The thick line portion of the Scandinavian Ice Sheet profile is our reconstruction shown in Figure9.

Table 2 .
Radiocarbon ages of marine fossils in the Bergen valley