Comparison of Climate and Carbon Cycle Dynamics During Late Quaternary Interglacials - Integrated Analysis of Interglacial Climate Dynamics (INTERDYNAMIC) - page 8

EPICA Dome C and Talos Dome ice cores, Antarctica, over the time interval 1
thousand years (ka) before present (BP) - 25 ka BP and 105 - 155 ka BP (Schneider

et al. 2013 ).

2.4 Palaeodata Assemblage

We reviewed a total of 162 publications af liated with MIS 11.3 climate (Kleinen

et al. 2014 ) and a set of 234 records for MIS 5.5. However, this extensive search
revealed that the majority of the publications provide only qualitative, discontin-
uous and/or poorly dated information about the past climate. Therefore we focused
on more recently published continuous records of interglacial climate and vegeta-
tion with well constrained chronologies, most suitable for a robust data-model
comparison (e.g., Kleinen et al. 2011 ).

3 Key Findings

3.1 Trends in Interglacial Carbon Cycle Dynamics

During Termination I, atmospheric CO 2 rose as the ocean released carbon to the
atmosphere (Schmitt et al. 2012 ). The δ

13 CO 2 measurements over Termination II
point at the same processes being responsible for the CO 2 increase as in Termi-
nation I, however, with different phasing and magnitude (Schneider et al. 2013 ).
After an initial CO 2 peak, CO 2 decreases during most interglacials, while the
Holocene and MIS 11.3 reveal CO 2 increases by about 20 and 10 ppmv, respec-
tively. Ruddiman ( 2003 ) interprets the rising CO 2 during the Holocene as the onset
of the Anthropocene. However, simulations with CLIMBER2-LPJ suggest that the
ocean mostly operates as a source of CO 2 to the atmosphere during interglacials
(Kleinen et al. 2010 ). Carbonate compensation and the excessive accumulation of
CaCO 3 in coral reefs lead to a slow CO 2 release to the atmosphere (Kleinen et al.
2010 ). The role of land carbon is more complex. During the Holocene, the land
mainly serves as a sink of carbon since soil and peat storages on land are slowly
growing (Kleinen et al. 2012 ). Reconstructed

13 CO 2 data from ice cores (Elsig
et al. 2009 ; Schmitt et al. 2012 ; Schneider et al. 2013 ) suggests a continuous
increase in the terrestrial carbon storage from 12 to 6 ka BP and a mostly neutral
role of the land from 6 to 2 ka (Fig. 1 a, b), invalidating the early Anthropocene
hypothesis. The CLIMBER2-LPJ simulation for MIS 11.3 using the same forcing
setup as for the Holocene shows CO 2 dynamics close to observations (Fig. 1 d). Our
new ice core measurements also show constant CO 2 concentrations over most of
MIS 5.5. However, the CLIMBER2-LPJ results are in disagreement with CO 2
reconstructions (Fig. 1 c), suggesting that the model still misses some important

δ