Climatic patterns over the European Alps during the LGM derived from inversion of the paleo-ice extent

https://doi.org/10.1016/j.epsl.2020.116185Get rights and content

Highlights

  • A range of different climatic scenarios can reproduce the observed ice extent.

  • Results for all scenarios show an increase in the ELA across the Alps from west to east.

  • The N-S gradient in the ELA is not pronounced over the entire Alps.

Abstract

Quaternary climate has been dominated by alternating glacial and interglacial periods. While the timing and extent of past ice caps are well documented, local variations in temperature and precipitation as a response to cyclic glaciations are not resolved. Resolving these issues is necessary for understanding regional and global climate circulation. In particular, the impact of the cold high-pressure zone above the Fennoscandian ice cap on the position of the jet stream in Europe, and a possible change in the direction and the source of moisture flow are still debated. Here we reconstruct climate conditions that led to the observed ice extent in the European Alps during the last glacial maximum (LGM). Using a new inverse method to reconstruct the spatially variable position of the equilibrium line altitude (ELA), we investigate whether the Alpine LGM ice cap dominantly received precipitation in the south due to a strong southward shift of the westerlies in midlatitudes. We report inversion results that enable us to estimate the role of climate, inversion method parameters, ice dynamics, flexure and topography in modulating the inferred equilibrium line altitude. Our main finding is the increase of the position of the ELA from west to east and from north to south of the mountain range, which suggest dominant moisture delivery over the Alps during the LGM relatively similar to today.

Introduction

The last glacial maximum (LGM), spanning from 26.5 to 20 ka (kilo annum) (Clark et al., 2009), represents the last large dominating cold period with its legacy still imprinted on present-day topography. The increasing computational capabilities and the development of complex large scale climate (Braconnot et al., 2012), ocean (Mikolajewicz, 2011), ice flow (Seguinot et al., 2018) and vegetation (Janská et al., 2017) models, accompanied by an increase in available proxy data (e.g., pollen, SST (see surface temperature), speleothems, GDGTs (glycerol dialkyl glycerol tetraethers)), are opening opportunities to improve our understanding of the climate conditions during the LGM and the interplay between glacial and climate processes that led to these conditions.

A significant body of work on LGM climate in Europe provides estimates on precipitation and temperature patterns, especially through advances in coupled climate modeling through the iterations of the Paleoclimate Modeling Intercomparison Project (PMIP) (Kageyama et al., 2006; Ramstein et al., 2007; Rivière et al., 2010; Braconnot et al., 2012; Abe-Ouchi et al., 2015; Beghin et al., 2016; Kageyama et al., 2017). Yet, uncertainties remain, highlighted by the discrepancies that exist between proxy data and models (Jost et al., 2005; Ramstein et al., 2007). Pollen-based reconstructions for Europe (Peyron et al., 1998) suggest a mean annual temperature depression north of the Pyrenees-Alps line, compared to today, of 12 ± 3 C, a mean coldest month depression of 30 ± 10 C, and a decrease in mean annual precipitation of 800±100 mm (60 ± 20% lower than today). In contrast, modeling studies suggest warmer conditions compared to the estimates inferred from pollen (Jost et al., 2005; Kageyama et al., 2006; Ramstein et al., 2007; Braconnot et al., 2012), especially in winter where the discrepancy is up to 10 C for western Europe (Ramstein et al., 2007; Braconnot et al., 2012). This discrepancy could arise from the relatively coarse resolution of current global climate models, as they are designed to simulate large scale features. For instance, the ice cap over the Alps are generally not represented in paleoclimate simulations (Abe-Ouchi et al., 2015; Kageyama et al., 2017). However, both pollen analysis (Wu et al., 2007; Bartlein et al., 2011) and glacier modeling (Seguinot et al., 2018) indicate a decrease in precipitation and a west-east temperature gradient across the Alps during the LGM.

Another unresolved question is the position and intensity of storm tracks during the LGM over Europe (Florineth and Schlüchter, 2000; Kageyama et al., 2006; Kuhlemann et al., 2008; Hofer et al., 2012; Luetscher et al., 2015; Beghin et al., 2016; Löfverström et al., 2016; Löfverström and Lora, 2017; Löfverström, 2020). Different scenarios have been proposed. First, the LGM may have been dominated by westerly winds like today, but under drier conditions (Peyron et al., 1998; Kageyama et al., 2006). Second, there may have been an overall southward deflection of the jet stream (Florineth and Schlüchter, 2000), with precipitation arriving to central Europe from the Mediterranean (Luetscher et al., 2015; Monegato et al., 2017). Such a shift may be explained by the breaking of the Rossby waves west of the Alps and the influence of the Fennoscandian ice sheet (Florineth and Schlüchter, 2000; Luetscher et al., 2015) and would have resulted in a meridional flow of moisture over the Alps from the Mediterranean (Florineth and Schlüchter, 2000; Kuhlemann et al., 2008; Luetscher et al., 2015; Monegato et al., 2017). Third, Kuhlemann et al. (2008) proposed that the precipitation pattern was dominated by increased cyclogenesis, both in frequency and intensity, in the western Mediterranean. This would have led to increased precipitation in the southern Alps. Finally, climate models suggest the possibility of a more zonal jet stream/stormtrack during the LGM due to the reflection of the stationary waves caused by the high elevation of the Laurentide Ice Sheet (Löfverström et al., 2016; Löfverström and Lora, 2017). This scenario suggests a moderate southward shift of high-precipitation zones compared to geomorphological and geochemical studies (Florineth and Schlüchter, 2000; Luetscher et al., 2015; Monegato et al., 2017). While the scenarios proposing a small or no shift suggest precipitation patterns over the Alps comparable to today, scenarios proposing a significant southward shift or increased cyclogenesis in the western Mediterranean claim that moisture was dominantly delivered to the southern Alps. Note that climate modeling supports a southward shift of the Westerlies, but not as drastic as suggested by geomorphological reconstructions (Beghin et al., 2016). The characteristics of the Alpine climate during the LGM should be reflected in the position of the equilibrium line altitude (ELA) across the Alps and thus in LGM ice thickness and extent, which we explore here.

Ice extent and ice thickness over the Alps during the LGM are constrained by field observations from trimlines, moraines and erratic boulders, as well as from dating of glacial deposits (Bini et al., 2009; Ehlers et al., 2011). Such data serve as both input and validation basis for glacial modeling of the region (Jouvet et al., 2008; Becker et al., 2016; Cohen et al., 2017; Jouvet et al., 2017; Seguinot et al., 2018). Recently, several in-depth modeling studies have been conducted in the Alps, focusing on the LGM. Becker et al. (2016) and Seguinot et al. (2018) used a positive day-degree (PDD) model to parametrize the mass balance, and PISM (the Parallel Ice Sheet Model), a hybrid model combining the shallow ice approximation (SIA; Hutter, 1983) with the shallow shelf approximation (SSA; Morland, 1987), to model ice flow. Cohen et al. (2017) focused their analysis on the Rhine glacier, modeled the full set of momentum equations for viscous fluids using Elmer/Ice (Gagliardini et al., 2013) and parametrized the mass balance as a linear function of surface elevation relative to the equilibrium line altitude. Although modeling studies used a variety of models and parametrizations, they all report a mismatch ranging from 500 m to over 800 m between the modeled ice thickness and geomorphological reconstructions. The reason could be difficulties in assessing the elevation of the ice surface increasing with distance from the mapped trimlines, therefore underestimating the ice thickness in the geomorphological reconstructions. Uncertainties in the ice flow models, and why they differ from observations, could arise from the use of flux limiters and the choice of spatial resolution (Imhof et al., 2019), as well as the use of present-day bedrock with or without removal of ice cover and post-glacial sediments, the difficulty to correctly simulate climate patterns in mass balance, the parametrization of ice flow, and spatial variations in timing of the LGM.

Here, our objective is to reconstruct the spatial variation of the ELA across the Alps during the LGM using a newly proposed inverse method (Višnjević et al., 2018) and the mapped ice extent during the LGM (Ehlers et al., 2011). The idea of using ELAs to gain insight into past climate patterns is not new (Callendar, 1950). The ELA depends primarily on two climate variables, precipitation and temperature, which represent the effects of accumulation and ablation, respectively, and, to a lesser extent, radiation (e.g., Ohmura et al., 1992; Oerlemans, 1992). The method introduced by Višnjević et al. (2018) enables us to invert ice extent data for the reconstruction of spatial ELA variations at the scale of a mountain range like the Alps. The model assumes the ice cap to be in a steady state, in near-equilibrium with climate, a scenario that has been proposed for the LGM (Clark et al., 2009; Monegato et al., 2017). We input the mapped ice extent as the only climatic constraint because of the above mentioned uncertainties in reconstructed ice thickness. In order to reduce the calculation time, computation is accelerated using graphic processing units (GPUs).

Below, we start by explaining the forward ice flow model and the inversion of ice extent for the ELA reconstruction. We then review the topographic data and the ice extent used. We perform a series of inverse model runs with different parameter settings to investigate the sensitivity of the results to the mass balance parameters (mass balance gradient and maximum accumulation rate), the inverse model parameters (initial guess, smoothing, update parameters), ice dynamics (deformation, sliding parameter and constriction factor), flexure (elastic thickness) and bedrock topography. We continue with reporting and interpreting the results and eventually exploit the reconstructed ELA pattern to further constrain the Alpine climate during the LGM.

Section snippets

Methods

In this section, we summarize the implementation of the forward and the inverse models. Further detail can be found in Višnjević et al. (2018). For this study, we added flexural isostasy, a constriction factor for ice flow (Braun et al., 1999), spatially variable sliding and a damping factor to the forward model.

Data

Bedrock topography, ice extent and mass balance rate parameters, maximum accumulation rate c and mass balance gradient β (Eq. (2)) are needed as input for our calculations. We use a digital elevation model (DEM) as bedrock topography taken from NASA's Shuttle Radar Topography Mission (SRTM) (Jarvis et al., 2008), where present day ice cover and post-glacial sediments are included. To test how changes in bedrock topography influence the result (Fig. S4C Supplement), we run our model on the

Results

In this section, we report inversion results for 25 different scenarios. A common feature in our results is an increase in the position of the ELA from west to east (Fig. 2C); with values ranging from 1000 m to 1550 m about Lyon and from 1650 m to 2000 m east of the Traun glacier and north of the Drau glacier. The N-S change in the ELA is less clear. It is more pronounced in the west and the center of the Alps. In the west (Solothurn lobe - Durance glacier, Fig. 2A), ELA values go from 1350 m

Discussion

In this study, we report a series of inversion results that highlight spatial variations in the LGM ELA across the Alps, based on the inverse method described in Višnjević et al. (2018). To evaluate the robustness of our results, we ran a series of experiments in which we test the influence of the mass balance parameters, the inversion parameters, the ice dynamics, the flexural rigidity and variations in bedrock topography on the inferred equilibrium line altitude.

The equilibrium line altitude

Conclusion

We reconstructed the spatially variable altitude of the equilibrium line over the Alps during the LGM, following the inversion approach described in Višnjević et al. (2018). Ice extent data from Ehlers et al. (2011) and a DEM of the Alps are used as input data for the model. We have conducted sensitivity tests for both the inverse and forward model parameters. Our model can reproduce the mapped LGM ice extent for a range of climatic scenarios, resulting in different ELA fields but keeping a

Declaration of Competing Interest

The authors have no conflict of interest in publishing these results.

Acknowledgments

VV was funded through SNF grant CRSII_2154434. The authors would like to thank J. Mey for providing us with the bedrock map with removed sediment fill. Thanks to N. Stalder, J. Elkadi, B. Lehmann and L. Malatesta for discussions.

References (61)

  • P. Beghin et al.

    Hat drives LGM precipitation over the western Mediterranean? A study focused on the Iberian Peninsula and northern Morocco

    Clim. Dyn.

    (2016)
  • A. Bini et al.

    Die Schweiz während des letzteiszeitlichen Maximums (LGM)

    (2009)
  • P. Braconnot et al.

    Evaluation of climate models using palaeoclimatic data

    Nat. Clim. Change

    (2012)
  • J. Braun et al.

    A new surface-processes model combining glacial and fluvial erosion

    Ann. Glaciol.

    (1999)
  • J.L. Brown et al.

    PaleoClim, high spatial resolution paleoclimate surfaces for global land areas

    Sci. Data

    (2018)
  • W.E. Budd et al.

    Empirical studies of ice sliding

    J. Glaciol.

    (1979)
  • G.S. Callendar

    Note on the relation between the height of the firn line and the dimensions of a glacier

    J. Glaciol.

    (1950)
  • J.D. Champagnac et al.

    Quaternary erosion-induced isostatic rebound in the western Alps

    Geology

    (2007)
  • P.U. Clark et al.

    The last glacial maximum

    Science

    (2009)
  • G.K. Clarke et al.

    Strain heating and creep instability in glaciers and ice sheets

    Rev. Geophys.

    (1977)
  • D. Cohen et al.

    Numerical reconstructions of the flow and basal conditions of the Rhine glacier, European Central Alps, at the Last Glacial Maximum

    Cryosphere

    (2017)
  • D.L. Egholm et al.

    Modeling the flow of glaciers in steep terrains: the integrated second-order shallow ice approximation (iSOSIA)

    J. Geophys. Res., Earth Surf.

    (2011)
  • D. Farinotti et al.

    A method to estimate the ice volume and ice-thickness distribution of alpine glaciers

    J. Glaciol.

    (2009)
  • S.P. Frankel

    Convergence rates of iterative treatments of partial differential equations

    Math. Tables Other Aids Comput.

    (1950)
  • O. Gagliardini et al.

    Capabilities and performance of Elmer/Ice, a new-generation ice sheet model

    Geosci. Model Dev.

    (2013)
  • F. Herman et al.

    Evolution of the glacial landscape of the Southern Alps of New Zealand: insights from a glacial erosion model

    J. Geophys. Res., Earth Surf.

    (2008)
  • F. Herman et al.

    The response time of glacial erosion

    J. Geophys. Res., Earth Surf.

    (2018)
  • D. Hofer et al.

    Simulated winter circulation types in the North Atlantic and European region for preindustrial and glacial conditions

    Geophys. Res. Lett.

    (2012)
  • M. Huss et al.

    Determination of the seasonal mass balance of four Alpine glaciers since 1865

    J. Geophys. Res., Earth Surf.

    (2008)
  • Cited by (0)

    View full text