Search

Understanding the evolution of, and the interactions between, ice sheets and the global climate over geological time is important for being able to constrain earth system sensitivity. However, direct observational evidence of past CO2concentrations only exists for the past 800,000 years. Records of benthic d 10 18O date back millions of years, but contain signals from both land ice volume and ocean temperature. In recent years, inverse forward modelling has been developed as a method to disentangle these two signals, resulting in mutually consistent reconstructions of ice volume, temperature and CO2. We use this approach to force a hybrid ice-sheet – climate model with a benthic d18O stack, reconstructing the evolution of the ice sheets, global mean sea level and atmospheric CO2 during the late Pliocene and the Pleistocene, from 3.6 15 million years (Myr) ago to the present day. During the warmer-than-present climates of the Late Pliocene, reconstructed CO2varies widely, from 320 – 440 ppmv for warm periods such as Marine Isotope Stage (MIS) KM5c, to 235 – 250 ppmv for the MIS M2 glacial excursion. Sea level is relatively stable during this period, with a high stand of 6 – 14 m, and a drop of 12 –26 m during MIS M2. Both CO2 and sea level are within the wide ranges of values covered by available proxy data for this period. Our results for the Pleistocene agree well with the ice-core CO2 record, as well as with different available sea-level 20 proxy data. During the early Pleistocene, 2.6 – 1.2 Myr ago, we simulate 40 kyr glacial cycles, with interglacial CO2decreasing from 280 – 300 ppmv at the beginning of the Pleistocene, to 250 – 280 ppmv just before the Mid-Pleistocene Transition (MPT). Peak glacial CO2 decreases from 220 – 250 ppmv to 205 – 225 ppmv during this period. After the MPT, when the glacial cycles change from 40 kyr to 80/120 kyr cyclicity, the glacial-interglacial contrast increases, with interglacial CO2 varying between 250 – 320 ppmv, and peak glacial values decreasing to 170 – 210 ppmv

Polar amplification and paleoclimate sensitivity (S) have been the subject of many paleoclimate studies. While earlier studies inferred them as single constant parameters of the climate system, there are now indications that both are conditioned by the type of forcing. Moreover, they might be affected by fast feedback mechanisms that have different strengths depending on the background climate. Here we use the intermediate complexity climate model CLIMBER-2 to study the influence of land ice and CO2 on polar amplification and S. We perform transient 5-Myr simulations, forced by different combinations of insolation, land ice, and CO2. Our results provide evidence that land ice and CO2 changes have different effects on temperature, both on the global mean and the meridional distribution. Land ice changes are mainly manifested in the high latitudes of the Northern Hemisphere. They lead to higher northern polar amplification, lower southern polar amplification, and lower S than more homogeneously distributed CO2 forcing in CLIMBER-2. Furthermore, toward colder climates northern polar amplification increases and consequently southern polar amplification decreases, due to the albedo-temperature feedback. As an effect, a global average temperature change calculated from high-latitude temperatures by using a constant polar amplification would lead to substantial errors in our model setup. We conclude that to constrain feedback strengths and climate sensitivity by paleoclimate data, the underlying forcing mechanisms and background climate states have to be taken into consideration.

In order to investigate the relation between ice sheets and climate in a warmer-than-present world, recent research has focussed on the Late Pliocene, 3.6 to 2.58 million years ago. It is the most recent period in Earth history when such a climate state existed for a significant duration of time. Marine Isotope Stage (MIS) M2 (~ 3.3 Myr ago) is a strong positive excursion in benthic oxygen records in the middle of the otherwise warm and relatively stable Late Pliocene. However, the relative contributions to the benthic δ18O signal from deep-ocean cooling and growing ice sheets are still uncertain. Here, we present results from simulations of the late Pliocene with a hybrid ice-sheet–climate model, showing a reconstruction of ice sheet geometry, sea-level and atmospheric CO2. Initial experiments simulating the last four glacial cycles indicate that this model yields results which are in good agreement with proxy records in terms of global mean sea level, benthic oxygen isotope abundance, ice core-derived surface temperature and atmospheric CO2 concentration. For the Late Pliocene, our results show an atmospheric CO2 concentration during MIS M2 of 233–249 ppmv, and a drop in global mean sea level of 10 to 25 m. Uncertainties are larger during the warmer periods leading up to and following MIS M2. CO2 concentrations during the warm intervals in the Pliocene, with sea-level high stands of 8–14 m above present-day, varied between 320 and 400 ppmv, lower than indicated by some proxy records but in line with earlier model reconstructions.

Following Milankovitch's theory the incoming insolation or summer energy at 65°N is typically analysed to predict the waxing or waning of land ice. We here use a model-based deconvolution of the LR04 benthic-d18O stack into land ice distribution (de Boer et al., 2014, Köhler et al., 2015) to verify if the latitudinal focal point of land ice dynamics has changed over the last 2 Myr or whether this choice of 65°N in orbital data is indeed well justified. We find that the 5°-latitudinal band which contributes most to land ice albedo radiative forcing (ΔR_[LI]) is 70-75°N between 2.0-1.5 Myr, which is then until 1.0 Myr gradually substituted by 65-70°N. During the last 1 Myr both 60-65°N and 65-70°N dominate ΔR_[LI] and contribute approximately the same amount, while the relative importance of 70-75°N is shrinking. Our analyses illustrates that the choice of 65°N seems for the last 1 Myr to be well justified, while for earlier parts of the last 2 Myr the dominant land ice changes seems to happen up to 10° further to the north. Focusing on the last 800 kyr (the time for which precise data on atmospheric CO2 concentration exists) we furthermore find that the multi-millennial land ice growth and proxy-based reconstruction of global cooling (= the glaciation) appear synchronously to each other and to decreasing obliquity, but diverge from CO2. This suggests that the global cooling associated with Earth's way into an ice age as deduced in the reconstructions has to be mainly caused by the land ice albedo feedback, and is not dominated by the CO2 greenhouse forcing. One way of perceiving this CO2-glaciation divergence in reconstructions is that the reduced incoming insolation at high latitudes causes land ice growth and cooling, while there is a coexisting process that keeps CO2 at a relatively constant level. Solid Earth modeling experiments have indicated that falling sea level might lead to enhanced magma and CO2 production at mid-ocean ridges. Hasenclever et al. (2017) suggested that the combination of marine volcanism at mid-ocean ridges and at hot spot island volcanoes might react to decreasing sea level and be a potential cause for this CO2-glaciation divergence. This CO2-glaciation divergence needs to be considered, when using paleo data to quantify paleoclimate sensitivity: periods with diverging CO2 and global temperature change should be filtered out when approximating the relationship between global temperature rise and CO2 concentrations (Köhler et al., 2018). References: de Boer et al. (2014). https://doi.org/10.1038/ncomms3999. Köhler et al. (2015). https://doi.org/10.5194/cp-11-1801-2015. Hasenclever et al. (2017). https://doi.org/10.1038/ncomms15867. Köhler et al. (2018). https://doi.org/10.1029/2018GL077717.