EGU Journal Logos (RGB) Open Access Open Access Open Access Advances in Annales Nonlinear Processes Geosciences Geophysicae in Geophysics Open Access Open Access Natural Hazards Natural Hazards and Earth System and Earth System Sciences Sciences Discussions Open Access Open Access Atmospheric Atmospheric Chemistry Chemistry and Physics and Physics Discussions Open Access Open Access Atmospheric Atmospheric Measurement Measurement Techniques Techniques Discussions Open Access Open Access Biogeosciences Biogeosciences Discussions Open Access Open Access Clim. Past, 9, 1749–1759, 2013 Climate www.clim-past.net/9/1749/2013/ Climate doi:10.5194/cp-9-1749-2013 of the Past of the Past © Author(s) 2013. CC Attribution 3.0 License. Discussions Open Access Open Access Earth System Earth System Dynamics Dynamics Discussions On the effect of orbital forcing on mid-Pliocene climate, vegetation Open Access Open Access and ice sheets Geoscientific Geoscientific Instrumentation Instrumentation M. Willeit, A. Ganopolski, and G. Feulner Methods and Methods and Potsdam Institute for Climate Impact Research, Potsdam, Germany Data Systems Data Systems Discussions Open Access Correspondence to: M. Willeit ([email protected]) Open Access Geoscientific Geoscientific Received: 24 March 2013 – Published in Clim. Past Discuss.: 3 April 2013 Model Development Revised: 17 June 2013 – Accepted: 1 July 2013 – Published: 1 August 2013 Model Development Discussions Open Access Abstract. We present results from modelling of the mid- orbital forcing. Therefore, the climate (Earth system)Open Access sensi- Pliocene warm period (3.3–3 million years ago) using the tivity estimates from PlioceneHydrology reconstructions and which do not Hydrology and Earth system model of intermediate complexity CLIMBER- account for the effect of orbitalEarth forcing System can be biased toward Earth System 2 analysing the effect of changes in boundary conditions as high values. well as of orbital forcing on climate. Sciences Sciences First we performed equilibrium experiments following the Discussions Open Access PlioMIP (Pliocene Model Intercomparison Project) protocol Open Access 1 Introduction with a CO2 concentration of 405 ppm, reconstructed mid- Ocean Science Pliocene orography and vegetation and a present-day or- Ocean Science bital configuration. Simulated global Pliocene warming is The time period between 3.3 and 3 million years ago (Ma), Discussions about 2.5 ◦C, fully consistent with results of atmosphere– traditionally named the mid-Pliocene warm period (MPWP), ocean general circulation model simulations performed for is the most recent interval in Earth’s history which is believed Open Access the same modelling setup. A factor separation analysis at- to have been significantly warmer than the pre-industrialOpen Access cli- ◦ ◦ ◦ mate. Most palaeoproxies suggest that atmospheric CO con- tributes 1.5 C warming to CO2, 0.3 C to orography, 0.2 C 2 Solid Earth to ice sheets and 0.4 ◦C to vegetation. centration during the MPWPSolid was in theEarth range from 300 to Transient simulations for the entire mid-Pliocene warm 400 ppm (Raymo et al., 1996; Pagani et al., 2009; Seki et al., Discussions period with time-dependent orbital forcing as well as interac- 2010; Bartoli et al., 2011), the global surface air tempera- ◦ tive ice sheets and vegetation give a global warming varying ture was higher by 2–3 C (estimate based on model simula- within the range 1.9–2.8 ◦C. Ice sheet and vegetation feed- tions) (Haywood and Valdes, 2004; Jansen et al.Open Access , 2007; Lunt Open Access backs in synergy act as amplifiers of the orbital forcing, trans- et al., 2012b) and sea level was up to 25 m above present The Cryosphere forming seasonal insolation variations into an annual mean level during the warmestThe periods Cryosphere of the MPWP (Miller et al., Discussions temperature signal. The effect of orbital forcing is more sig- 2012; Foster and Rohling, 2013). If these estimates are suffi- nificant at high latitudes, especially during boreal summer, ciently accurate, this makes the MPWP a useful (although by when the warming over land varies in the wide range from 0 no means direct) analogue for the future greenhouse world to 10 ◦C. The modelled ice-sheet extent and vegetation dis- (Salzmann et al., 2009; Haywood et al., 2011b). tribution also show significant temporal variations. Modelled A comprehensive set of reconstructions of climate and and reconstructed data for Northern Hemisphere sea-surface boundary conditions for the MPWP is available from the temperatures and vegetation distribution show the best agree- PRISM3D (Pliocene Research, Interpretation and Synoptic ment if the reconstructions are assumed to be representa- Mapping) dataset (Dowsett et al., 2010). PRISM3D includes tive for the warmest periods during the orbital cycles. This global reconstructions for sea-surface temperature (SST), suggests that low-resolution Pliocene palaeoclimate recon- vegetation distribution, ice-sheet extent and orography. Most of the data are the outcome of a combination of reconstructed structions can reflect not only the impact of increased CO2 concentrations and topography changes but also the effect of data and modelling results. SST reconstructions show little change in the tropics relative to present day, while significant Published by Copernicus Publications on behalf of the European Geosciences Union. 1750 M. Willeit et al.: Effect of orbital forcing on mid-Pliocene climate warming is observed at mid- to high latitudes, up to 3–4 ◦C to the CLIMBER grid and then add them to present-day in the Northern Hemisphere (NH) with major warming oc- model orography. curring in the North Atlantic (Dowsett et al., 2009; Dowsett Some attempts were made to use the MPWP to derive the and Robinson, 2009). so-called Earth system sensitivity (ESS), i.e. the long-term There are numerous pieces of evidence from palaeodata quasi-equilibrium response of global temperature to radia- indicating that the spatial distribution of vegetation during tive forcing. In contrast to the climate sensitivity (CS, also the MPWP was substantially different from today. The main called Charney sensitivity) which involves only fast feed- difference was a northward expansion of taiga replacing tun- backs, ESS includes also the slow feedbacks related to ice dra in the NH. The coast of Antarctica was at least locally sheets, vegetation and atmospheric aerosols. The ESS esti- covered by some vegetation. The vegetation distribution in mates from MPWP range from modest 4.5 ◦C(Lunt et al., mid and low latitudes was essentially similar to present day, 2009) to very high 7.1–8.7 ◦C(Pagani et al., 2009) values. although the moister climate favoured forest and woodland This large range can be explained by uncertainties in recon- growth in some regions now covered by steppe or grassland structed CO2 concentrations and model-derived global tem- (Thompson and Fleming, 1996). The PRISM3D vegetation peratures. Similar attempts have been made to estimate the reconstruction is based on a combined model–data approach dependence of global mean sea level on temperature or CO2 and is described in Salzmann et al. (2008). It combines an in- (e.g. Foster and Rohling, 2013). Obviously, any attempt to ternally consistent dataset of reconstructions at 202 palaeob- employ MPWP data to constrain climate models and cli- otanical sites with results of simulations of the BIOME4 veg- mate sensitivity depends not only on the accuracy of palaeo- etation model driven by modelled mid-Pliocene climatology. climatic reconstructions but also on the validity of the im- Where data coverage is poor, the vegetation reconstruction plicit assumption that reconstructed MPWP climate repre- is mostly based on the model output. Some important re- sents quasi-equilibrium response to elevated level of CO2 gions with low data availability include high northern Eura- and modified topography. Testing this assumption has im- sia and North American mid-latitudes. This should be kept portant implications for the interpretation of MPWP recon- in mind when comparing model output with reconstructions. structions. Even if climatic conditions during the MPWP More applications of the BIOME4 model to the MPWP can were relatively stable compared to the more recent past, sig- be found in e.g. Haywood et al. (2002b) and Salzmann et al. nificant variability is evident in the benthic oxygen isotope (2009). Simulations with the BIOME4 model in offline mode (δ18O) record (Lisiecki and Raymo, 2005), suggesting a non- do not account for the feedback of the vegetation on cli- negligible response of the climate system to orbital forcing. mate. MPWP simulations with a coupled climate–vegetation PRISM3D SSTs are well known to represent warmest condi- model generally show a reasonably good agreement with re- tions during the MPWP because of the warm peak averaging constructed vegetation (PRISM2), but regionally important procedure involved in the reconstruction (Dowsett, 2007). differences exist, with the model underestimating the north- The temporal resolution of vegetation proxies, mostly fos- ward expansion of coniferous forest and overestimating the sil pollen, is very coarse, and consequently vegetation recon- extension of arid areas (Haywood and Valdes, 2006). structions do not resolve orbital variations (Salzmann et al., MPWP sea level has been estimated to be up to 25 m 2008; Thompson and Fleming, 1996). Where information on higher than present during the warmest periods of the MPWP vegetation