Influence of regional parameters on the surface mass balance of the Eurasian ice sheet during the peak Saalian (140 kya)

https://doi.org/10.1016/j.gloplacha.2009.03.021Get rights and content

Abstract

Recent geologically-based reconstructions of the Eurasian ice sheet show that during the peak Saalian (≈ 140 kya) the ice sheet was larger over Eurasia than during the Last Glacial Maximum (LGM) at ≈ 21 kya. To address this problem we use the LMDZ4 atmospheric general circulation model to evaluate the impact on the Saalian ice sheet's surface mass balance (SMB) from proglacial lakes, dust deposition on snow, vegetation and sea surface temperatures (SST) since geological records suggest that these environmental parameters were different during the two glacial periods. Seven model simulations have been carried out. Dust deposition decreases the mean SMB by intensifying surface melt during summer while proglacial lakes cool the summer climate and reduce surface melt on the ice sheet. A simulation including both proglacial lakes and dust shows that the presence of the former parameter reduces the impact of the latter, in particular, during summer. A switch from needle-leaf to tundra vegetation affects the regional climate but not enough to significantly influence the SMB of the nearby ice margin. However, a steady-state vegetation in equilibrium with the climate should be computed to improve the boundary conditions for further evaluations of the vegetation impact on the ice sheet's SMB. Finally, changes of the SST broadly affect the regional climate with significant consequences for the SMB.

Introduction

During the Quaternary Period, Northern Eurasia has been affected by several major glaciations. The Quaternary Environment of the Eurasian North (QUEEN) project reconstructed the Eurasian maximum extension for the last four major glaciations (Svendsen et al., 2004): The Late Saalian (180–140 kya), the Early Weichselian (100–80 kya), the Middle Weichselian (60–50 kya), and the Late Weichselian (21–15 kya). The maximum extents of these ice sheets have been determined by mapping marine and terrestrial glacial morphology. From these studies, it appears that the Late Saalian ice sheet (maximum at c. 140 kya referred to as the Late Saalian over Eurasia1) was the largest Northern Eurasian ice sheet of the Late Quaternary.

The Late Saalian ice sheet extended further east and southeast over Eurasia than at the (LGM) and it appears to have been as much as approximately 56% larger during its maximum extent (Fig. 1). This suggests that the ice volume was also larger at 140 kya over Eurasia, which is confirmed by a large lithospheric depression due to continental ice loading inferred from mapped paleo-lakes levels and raised shorelines dated to this time (Astakhov, 2004). However, the eustatic sea level is not sufficiantly constrained to conclude that significant differences existed between the global ice volume during the time periods for the LGM and the Late Saalian ice sheet. The review by Rabineau et al. (2006) shows that eustatic sea level estimations range from ≈ 120 m (Peltier and Fairbanks, 2006) to 163 m (CLIMAP, 1984) below the present sea level during the LGM and from 120 m (Shackleton, 1987, Rohling et al., 1998) to more than 130 m (Ferland et al., 1995) during the Late Saalian.

The main global forcings determining climate and build-up of ice sheets, such as orbital insolation and concentration of greenhouse gases, appear to have been similar for the Late Saalian and the LGM (Fig. 1). But the peak Saalian ( 140 kya) is also the consequence of a different evolution of the global insolation prior to its glacial maximum than that before the LGM. Moreover, it is clear that the Milankovitch theory alone cannot explain the variations in ice sheet volume and extent during the peak of the Quaternary glacial periods and the initial insolation forcing has to be increased by global and regional feedbacks (e.g. Krinner et al. (2004)). We focus this study on the potential impact of regional parameters such as proglacial lakes, dust deposition on snow, vegetation cover and sea surface temperatures on the regional climate and on the surface mass balance (SMB) of the Late Saalian Eurasian ice sheet. More specifically:

  • (1)

    We test the impact of dust deposition on the ice sheet's SMB since Calov et al. (2005) showed that dust deposition on snow reduced the albedo value by 10% to 30% and Krinner et al. (2006) showed that dust deposition on snow during the LGM can help explain the absence of ice sheets in Eastern Siberia. Considering that the downward solar radiation is about 100 W·m 2 at Northern latitudes and ice and snow albedo are of about 60% to 80%, the dust content of the snow induces an increase of 6 W·m 2 to 8 W·m 2, which is significantly higher than the effect produced by the atmospheric dust content.

  • (2)

    Since the LGM ice sheet extent was smaller than during the Late Saalian, the major northern Russian and Western Siberian rivers could still flow to the Arctic Ocean and no significant proglacial lakes seem to have formed in Siberia (Mangerud et al., 2004). On the contrary, the Late Saalian ice sheet blocked these rivers and large proglacial lakes could have formed in a similar way as during the Early Weichselian (Mangerud et al., 2001, Mangerud et al., 2004). Proglacial lakes can have a different influence on the regional climate depending on the regional settings. Krinner et al. (2004) showed that the large Early Weichselian ice-dammed lakes had a strong climatic impact as they cooled the near-surface summer climate due to their high thermal inertia. On the contrary, Hostetler et al. (2000) showed that at 11 kya in North America, the cooling induced by Lake Agassiz reduced the amount of moisture spreading over the Laurentide ice sheet, contributing to its retreat. Therefore, the impact of the proglacial lakes south of the Saalian ice sheet margin cannot be easily predicted and must be evaluated in GCM simulations.

  • (3)

    Vegetation plays an important role in the climate system as it modifies heat exchange and water vapor content of the atmosphere (Crowley and Baum, 1997). Simulations by Crowley and Baum (1997) showed that the replacement of conifer forest by tundra in Eurasia during the LGM cooled the regional climate.

  • (4)

    Sea surface conditions (sea-ice extent, thickness and concentration, sea surface temperatures) are an obvious climatic forcing on the adjacent continents. However, SST has been suggested to play a key role in the climate system. For example Smith et al. (2003) prescribed various sea-ice cover extents (SIC) to test the impact on the LGM and Holocene climate inducing large precipitation changes over Eurasia. Ruddiman and McIntyre (1979) as well as Hebbeln et al. (1994) showed that by introducing seasonally open water conditions in the North Atlantic during the LGM, mass balance of the Northern Hemisphere ice sheets was largely affected. Because there are no global- or even regional-scale reconstructions of the Late Saalian sea surface conditions, we test two different LGM reconstructions as possible surrogates.

The relative importance of these regional-scale forcings for the SMB of the Late Saalian Eurasian ice sheet is assessed in this study using an atmospheric general circulation model.

The relative importance of these regional-scale forcings for the SMB of the Late Saalian Eurasian ice sheet is assessed in this study using an atmospheric general circulation model. Our model exercise complements the HOTRAX sediment core studies by addressing the paleoenvironment during the Marine Isotope Stage 6 (MIS 6; when the Late Saalian Eurasian ice sheet existed), which is manifested in sedimentary features indicative of a very large glaciation at the Arctic margins (Adler et al., 2009-this issue).

Section snippets

Methods

We use the LMDZ4 AGCM (Hourdin et al., 2006) which takes into account the climatic impact of open water surfaces and dust concentration in snow (Krinner, 2003, Krinner et al., 2006). The model has been run with 96 × 72 grid cells horizontally and with 19 vertical layers. The horizontal resolution is irregular, varying from the highest resolution of 100 km grid cells centered over Eurasia at 65°N/60°E to 550 km elsewhere. The impact of using a stretchable grid on climate as been discussed in

Present-day control run

The simulated present-day surface climate (REF0) is here compared with observations compiled by the Climatic Research Unit (CRU, New et al. (1999)) and by the European Centre for Medium-Range Weather Forecast (ECMWF, Uppala et al. (2005)). REF0 takes into account the effect of dust deposition on snow, but due to the low present-day dust deposition rates (Mahowald et al., 1999), this effect is small, as reported previously by Krinner et al. (2006). The simulated winter surface air temperature is

Discussion

The six Late Saalian simulations carried out here (REF140, LAKES140, DUST140, FULL140, VEG140, SST140) aim to evaluate the impact from dust deposition, proglacial lakes, vegetation and sea surface temperatures on the surface mass balance (SMB) of the large Late Saalian Eurasian ice sheet. The main goal is to understand if these parameters could have contributed significantly to the maintenance of this ice sheet during the glacial maximum.

Table 2 shows that the surface mass balance calculated

Conclusions

To summarize, the main results and conclusions of this work about the Late Saalian period are the following:

  • (1)

    In a “warm” glacial climate, the effect of proglacial lakes on climate and surface mass balance is enhanced and the effect of dust deposition on snow is reduced;

  • (2)

    Dust deposition on the ice sheet and on the surrounding continental ice-free regions leads to increased surface melt;

  • (3)

    Combined with dust deposition, that is, in a “warm” background climate with higher melt rates, proglacial lakes

Acknowledgments

The authors particularly thank Lev Tarasov and Masa Kageyama for their very constructive suggestions. We would like to thank Xavier Fettweis for his personal communications and G. Spada for his helpful comments. The authors wish to acknowledge the use of the Ferret program for analysis and graphics in this paper. Ferret is a product of NOAA's Pacific Marine Environmental Laboratory (Information is available at http://ferret.pmel.noaa.gov/Ferret/). The authors acknowledge support by the Agence

References (53)

  • Abe-OuchiA. et al.

    Climatic conditions for modelling the northern hemisphere ice sheets throughout the ice age cycle

    Clim. Past

    (2007)
  • CalovR. et al.

    Transient simulation of the last glacial inception. Part II: sensitivity and feedback analysis

    Clim. Dyn.

    (2005)
  • ClaquinT. et al.

    Radiative forcing of climate by ice-age atmospheric dust

    Clim. Dyn.

    (2003)
  • CLIMAP

    The last interglacial ocean

    Quat. Res.

    (1984)
  • CrowleyT.

    Ice age terrestrial carbon changes revisited

    Global Biogeochem. Cycles

    (1995)
  • CrowleyT. et al.

    Effect of vegetation on an ice-age climate model simulation

    J. Geophys. Res.

    (1997)
  • CrowleyT. et al.

    Transient nature of late Pleistocene climate variability

    Nature

    (2008)
  • DelmonteB. et al.

    Epica Dome C Ice Cores Insoluble Dust Data

    (2004)
  • FettweisX.

    Reconstruction of the 1979–2006 Greenland ice sheet surface mass balance using the regional climate model MAR

    Cryosphere

    (2007)
  • HaysJ. et al.

    Variations in the Earth's orbit: pacemaker of the ice ages

    Science

    (1976)
  • HebbelnD. et al.

    Moisture supply for northern ice-sheet growth during the last glacial maximum

    Nature

    (1994)
  • HostetlerS. et al.

    Simulated influence of the Lake Agassiz on the climate of the central North America 11,000 years ago

    Nature

    (2000)
  • HourdinF. et al.

    The LMDZ4 general circulation model: climate performance and sensitivity to parametrized physics with emphasis on tropical convection

    Clim. Dyn.

    (2006)
  • JollyW. et al.

    A generalized, bioclimatic index to predict foliar phenology in response to climate

    Glob. Chang. Biol.

    (2005)
  • KageyamaM. et al.

    Impact of the North American ice-sheet orography on the last glacial maximum eddies and snowfall

    Geophys. Res. Lett.

    (2000)
  • KrinnerG. et al.

    The Antarctic surface mass balance in a stretched grid general circulation model

    Ann. Glaciol.

    (1997)
  • Cited by (34)

    • Late Quaternary sedimentary processes in the central Arctic Ocean inferred from geophysical mapping

      2020, Geomorphology
      Citation Excerpt :

      While traces of sea-ice are difficult to identify in the seafloor morphology, ridges and bathymetric highs in the central Arctic Ocean with crests shallower than ~1000 m have consistently been scoured by deep-draft glacier ice from icebergs calved from ice sheets and thick ice shelves (Jakobsson, 1999; Kristoffersen et al., 2007; Polyak et al., 2001; Jakobsson et al., 2014). The extension and thickness of ice sheets and ice shelves in the Arctic Ocean during past glacial periods has been a widely discussed topic over several decades (Mercer, 1970; Hughes et al., 1977; Grosswald and Hughes, 2008; Hebbeln et al., 1994; Sher, 1995; Polyak et al., 2001; De Vernal et al., 2005; Colleoni et al., 2009; Niessen et al., 2013; Jakobsson et al., 2014). Results from geophysical mapping and sediment coring show that the seafloor morphology of the Lomonosov Ridge, Chukchi Plateau, Arlis Plateau, Yermak Plateau and Morris Jesup Rise have been imprinted by grounding ice.

    • Major hydrological shifts in the Black Sea “Lake” in response to ice sheet collapses during MIS 6 (130–184 ka BP)

      2019, Quaternary Science Reviews
      Citation Excerpt :

      During this period, δ18Oostracod varied in concert with δ18OSofular Cave due to the water vapour source effect identifying the Black Sea as the main source for the Sofular Cave speleothem growth (Lachniet, 2009; Badertscher et al., 2011; Wegwerth et al., 2014). Such similarity between the δ18Oostracod and δ18OSofular Cave records is also evident during the first half of MIS 6 (180-167 ka BP), which corroborates a long-term and even stronger period of meltwater discharge owing to the considerably larger Eurasian ice volume (Svendsen et al., 2004; Colleoni et al., 2009; Badertscher et al., 2011; Marks et al., 2018). A similar scenario with strong meltwater supply is also conceivable for the second period of oxygen isotope depletion in the Black Sea “Lake” (160-145 ka BP), where the record of Sofular Cave bears a gap (Badertscher et al., 2011).

    • Differences between the last two glacial maxima and implications for ice-sheet, δ<sup>18</sup>O, and sea-level reconstructions

      2017, Quaternary Science Reviews
      Citation Excerpt :

      For example, dust transportation is thought to have been more intense during the LGM than the PGM (e.g., Naafs et al., 2012). In addition, the PGM EIS was affected by large pro-glacial lakes, which (because of large heat capacities) can cool regional summer climates, and which also modify precipitation through meso-scale atmospheric feedbacks (e.g., Krinner et al., 2004; Colleoni et al., 2009). No geological evidence exists for such lakes during the LGM (Mangerud et al., 2004).

    • An East Siberian ice shelf during the Late Pleistocene glaciations: Numerical reconstructions

      2016, Quaternary Science Reviews
      Citation Excerpt :

      Simulated grounding is in agreement with the geomorphological evidence dated to MIS 6 found by Polyak et al. (2001); Jakobsson et al. (2001, 2005, 2008, 2010). Krinner et al. (2006) and Colleoni et al. (2009b) show the importance of the dust-albedo feedback on the perennial snow cover during the last two glaciations. In particular, Krinner et al. (2006) shows that during the LGM, the dust deposition on snow enables the melting of the perennial snow cover along the southern margin of the Eurasian ice sheet.

    View all citing articles on Scopus
    View full text