Impact of solar forcing on the surface mass balance of northern ice sheets for glacial conditions

https://doi.org/10.1016/j.epsl.2012.04.043Get rights and content

Abstract

The climate of the last glacial period has been punctuated by abrupt changes, termed the Dansgaard–Oeschger (DO) events, occurring every 1500–4500 yr. So far, the cause of these events, which involve changes in the thermohaline circulation, remains an open issue. It has been proposed that small changes in the freshwater flux in the North Atlantic, possibly coming from cyclic variations in solar activity, could act as a pacemaker and synchronize the events. Here we use the general circulation model IPSL_CM4 to investigate the impact of changes in the total solar irradiance (TSI) on the freshwater flux coming from ablation of the Northern hemisphere ice sheets. We test four different TSI values between 1360 and 1375 W/m2, and in this range establish a linear relationship between TSI and ablation rates over different sectors of the ice sheets. Our results show that a change in TSI smaller than 1%, that would be undetectable in paleo-records, can trigger changes in the freshwater flux in the North Atlantic at an amplitude similar to the one required to synchronize abrupt events in the climate model of intermediate complexity CLIMBER. Given the uncertainties on the past solar activity, we conclude that the hypothesis of a solar origin of the periodicity of D/O events cannot be ruled out and that the relationship between ice ablation and TSI variations is worth being further investigated.

Highlights

► We test the impact of four values of total solar irradiance with IPSL_CM5 for the LGM. ► We compute the surface mass balance of the ice sheets for these TSI values. ► We establish a relationship between ice ablation and solar activity. ► The ablation rate is found to respond to a weak change in solar activity.

Introduction

The last glacial period has been punctuated by many abrupt warmings first found in Greenland ice, the so-called Dansgaard–Oeschger (DO) events (Dansgaard et al., 1993). During these events, the paleo temperature reconstructions from ice cores from Greenland show an initial abrupt warming of 8–16 °C in a few decades (Wolff et al., 2010), followed by a gradual cooling over several hundred to several thousand years, and finally an abrupt cooling back to cold conditions. Pollen records from marine and terrestrial sediment cores show that these abrupt warmings are also found around the whole North Atlantic region, leading to major vegetation changes (e.g. Fletcher et al., 2010, Harrison and Sánchez-Goñi, 2010). The records show a pacing of around 1500 yr, at least during part of MIS3 (Alley et al., 2001, Schulz, 2002). The statistical analysis of the record shows that the waiting times between warmings occur more often around 1500 yr or, with smaller probability, around 3000 and 4500 yr (Alley et al., 2001). The statistical significance of the periodicity of DO events has recently been questioned, and some authors suggest it to be only a noise-induced phenomenon (Ditlevsen et al., 2007). However a hidden periodic driver cannot be ruled out (Ditlevsen and Ditlevsen, 2009) and so far the issue of the periodicity is unresolved (Wolff et al., 2010). As changes in Atlantic Meridional Overturning Circulation (AMOC) seem to be involved in this millennial-scale variability during the last glacial (Keigwin et al., 1994, Ganopolski and Rahmstorf, 2001) used the intermediate complexity climate model CLIMBER-2 to investigate the stability of the glacial climate and the response of the AMOC to freshwater forcing. They found that for the last glacial, two modes exist: a “cold” mode where the North Atlantic Deep Water (NADW) forms South of Iceland and a “warm” unstable mode where it forms in the Nordic Seas. In CLIMBER-2, the shift from the cold mode to the warm mode can be easily triggered with a small negative anomaly in North Atlantic freshwater flux. When a freshwater forcing with an amplitude of 30 mSv is imposed in the Atlantic (50–80°N) with a periodicity of 1500 yr, the model simulates abrupt warm events similar to the observed DO events, at the same periodicity. In another study with the same model, Ganopolski and Rahmstorf (2002) tested the hypothesis from Alley et al. (2001) of stochastic resonance, and showed that abrupt warmings could also be simulated when a white-noise component is added to the freshwater flux North of 50°N. In this case, the duration between the abrupt warming events depends on the noise amplitude, and for a strong noise converges to 1000–2500 yr. If a weak sinusoidal forcing of periodicity 1500 yr and only 10 mSv in amplitude is added to the noise, DO events triggered by the noise are synchronized with a preferred interspike of 1500 or 3000 yr. The model shows stochastic resonance at noise amplitudes near 50 mSv. However, the possible origin of a periodic freshwater forcing is not investigated in these papers. The authors suggest that it could for instance arise from changes in the sea-ice export, the runoff or the mass balance of the ice sheets, which represent a large amount of freshwater stored near the North Atlantic. Such changes could be triggered by changes in the solar forcing, but at present no 1470 yr solar cycle has been detected in the records of cosmogenic nuclides used to reconstruct past solar activity (Marchal, 2005, Muscheler and Beer, 2006). However, two centennial-scale solar cycles of 210 and 87 yr may exist (Wagner et al., 2001, Peristykh and Damon, 2003). Braun et al. (2005) proposed that the combination of the two cycles could have synchronized a 1470 yr climate cycle. They forced CLIMBER-2 with two freshwater sinusoidal cycles of periodicity 210 and 86.5 yr superimposed on a background freshwater flux. They tested different amplitudes for the sinusoidal signals and the background freshwater flux and showed that there is a range of values for which the model simulates abrupt warmings over Greenland with spacing 1470 yr.

Here we use the Atmosphere-Ocean General Circulation Model (AOGCM) IPSL_CM4_v1 (Marti et al., 2010) to test the hypothesis of a solar impact on the freshwater budget in the North Atlantic via changes in the surface mass balance of the adjacent ice sheets (North America, Greenland and Fennoscandia). The configuration used here is for the Last Glacial Maximum (LGM), for which a reference experiment existed at the beginning of our study (Kageyama et al., 2009). This was the experiment with boundary conditions closest to MIS3 that was available and stabilized, the others being for pre-industrial, mid-Holocene or the Eemian, which could not be used for our study.

We test different values for the total solar irradiance (TSI) to simulate a possible solar variability. The exact value of TSI is not well known, even for the present day (Kopp and Lean, 2011); but what is relevant for our study are the slight variations around the reference value (1365 W/m2). The amplitude of past TSI variations is controversial. For the Maunder Minimum for instance, Lean et al. (1995) estimated the decrease in TSI compared to its present-day level to −0.25% and (Reid, 1997) to −0.65%. The reconstruction from Bard et al. (2000) over the last 1200 yr, based on cosmogenic nuclides and different scaling factors to match different previous reconstructions for the Maunder Minimum, gives a maximum amplitude of TSI changes of about 1% (13 W/m2) between the Medieval Warm Period and the Spörer Minimum. The last IPCC report however gives estimates of a TSI decrease ranging from 0 to −0.27% (Forster et al., 2007, Table 2.10, p. 191). The latest estimates support rather small changes, not stronger than current minima of a solar cycle (e.g. Krivova et al., 2007, Steinhilber et al., 2009), and solar variability over the past 9 kyr may only equal to 0.9±0.4 W/m2 (Steinhilber et al., 2009). Given these controversies and the fact that we do not have any reliable reconstruction of solar activity during glacial periods, we test the impact of four different values of TSI: 1365 W/m2, the value usually used as the reference in IPSL_CM4 runs; 1360 W/m2, 1367 and 1375 W/m2 (i.e. variations of −0.36%, +0.15% and +0.73% respectively compared to the reference value of 1365 W/m2). We establish the sensitivity of the freshwater flux from the northern ice sheets to TSI changes within this range of values. This sensitivity could be used in the future when TSI variability is better constrained.

Section snippets

Material and methods

The IPSL_CM4_v1 AOGCM includes LMDz.3.3, the atmospheric component, with resolution of 96×72×19 in longitude×latitude×altitude and a regular horizontal grid, and ORCA2, the ocean module, with an irregular horizontal grid of 182×149 points and 31 depth levels. Sea ice is dynamically simulated using the LIM2 model. Since interactive dynamic vegetation is not available with this version of the IPSL model, vegetation is fixed to its present-day distribution, including agriculture, and the leaf area

Results: climatic response and induced changes on the surface mass balance of the ice sheets and freshwater fluxes

The sensitivity of the global annual temperature to TSI changes, defined as Δ(temperature)/Δ(TSI), is about 0.1 °C/W m−2 in our simulations. This is in line with the sensitivity of the model to TSI changes under present conditions (IPSL_CM4, Servonnat et al., 2010). Thus, the slightly higher global albedo at the LGM (0.35 in our LGM control simulation) compared to present (0.32 in the 20th century simulation with IPSL_CM4) due to the presence of the northern ice sheets, the change in snow cover

Discussion and conclusion

The North American ice sheet on the Pacific side appears to be rather sensitive to TSI changes. This result is in line with the work on Alaskan glaciers (Wiles et al., 2004), suggesting a cyclicity of glacial activity of about 200 yr over the last millennium, which might be linked to the variations in solar activity. Roche et al. (2010) showed that an increase of freshwater in that region in the LOVECLIM model increased sea ice and led to a cooling. However, they tested high values of

Acknowledgements

The simulations have been run on the computer of the Centre de Calcul Recherche et Technologie (CCRT, France). We thank three anonymous reviewers for their comments to improve this paper.

References (46)

  • M. Bougamont et al.

    Impact of model physics on estimating the surface mass balance of the Greenland ice sheet

    Geophys. Res. Lett.

    (2007)
  • P. Braconnot et al.

    Results of PMIP2 coupled simulations of the Mid-Holocene and Last Glacial Maximum. Part 1: experiments and large-scale features

    Clim. Past

    (2007)
  • H. Braun et al.

    Possible solar origin of the 1470-year glacial climate cycle demonstrated in a coupled model

    Nature

    (2005)
  • P. Clark et al.

    Freshwater forcing of abrupt climate change during the last glaciation

    Science

    (2001)
  • W. Dansgaard et al.

    Evidence for general instability of past climate from a 250-kyr ice-core record

    Nature

    (1993)
  • P.D. Ditlevsen et al.

    The do-climate events are probably noise induced: statistical investigation of the claimed 1470 years cycle

    Clim. Past

    (2007)
  • P.D. Ditlevsen et al.

    On the stochastic nature of the rapid climate shifts during the last ice age

    J. Clim.

    (2009)
  • A. Fanning et al.

    Temporal-geographical meltwater influences on the North Atlantic conveyor: implications for the Younger Dryas

    Paleoceanography

    (1997)
  • J. Flückiger et al.

    Variations in atmospheric N2O concentration during abrupt climatic changes

    Science

    (1999)
  • Forster, P., Ramaswamy, V., Artaxo, P., Bernsten, T., Betts, R., Fahey, D., Haywood, J., Lean, J., Lowe, D., Myhre, G.,...
  • A. Ganopolski et al.

    Rapid changes of glacial climate simulated in a coupled climate model

    Nature

    (2001)
  • A. Ganopolski et al.

    Abrupt glacial climate changes due to stochastic resonance

    Phys. Rev. Lett.

    (2002)
  • M. Kageyama et al.

    Glacial climate sensitivity to different states of the Atlantic Meridional Overturning Circulation: results from the IPSL model

    Clim. Past

    (2009)
  • View full text