New estimates of Earth system sensitivity from the Cenozoic Introduction: Earth system sensitivity Case study 1: The mid Pliocene Case study 2: The PETM and the early Eocene Conclusions Dan Lunt, Alan Haywood, Paul Valdes, Harry Dowsett EoMIP: M. Heinemann, M. Huber, A. Legrande, A. Winguth, C. Loptson PlioMIP: W. Le-Chan, N. Rosenbloom, M. Chandler, F. Bragg Common theme – model uncertainty and intercomparisons

the change in global mean near-surface air temperature that would result from a sustained doubling of the atmospheric CO 2 concentration likely to be in the range 2 to 4.5°C with a best estimate of about 3°C, and is very unlikely to be less than 1.5°C. Climate sensitivity IPCC, AR4, 2007 Given a CO 2 change, can estimate global mean temperature response Given a temperature change, can estimate CO 2 change

But….. (1) ….official definition is century-scale (short-term feedback processes only). (2) Even models that do include dynamic oceans neglect many other processes…. Therefore define EARTH SYSTEM SENSITIVITY – long-term response to sustained elevated CO 2 concentrations, including all Earth system feedbacks. Knutti and Hegerl, Nature Geoscience, 2008

….combined model - data approach…. (has advantages vs. pure-data or pure-modelling approaches) Find a palaeo time period when (a)CO 2 was higher than modern (i.e. not the LGM!) (b) we have an idea of ice sheets and palaeogeographies and other boundary conditions…. Run a climate model with the elevated CO 2 and prescribed (from data) ice and vegetation….compare with proxy data where available. Modelled global mean temp will give the long-term climate sensitivity (Earth System Sensitivity) to this CO 2 forcing. Model is an interpolator of paleo data.

(1) CASE STUDY 1: The mid-Pliocene…. Boundary conditions from the PRISM2 project, USGS CO 2 prescribed as 400 ppmv

ΔT=3.3 o C CO2 = 400ppmv Mid-Pliocene simulated temperatures compared with pre-industrial:

Mid-Pliocene simulated temperatures compared with data Global-annual mean response is good, but…. Dowsett et al, PPP, 2011

Total mid-Pliocene temp change, relative to pre- industrial Direct contribition from CO2 (climate sensitivity) Vegetation feedbacks Ice feedbacks EARTH SYSTEM SENSITIVTY Feedbacks due to CO2 alone Orography forcing Implications for Earth system sensitivity:

dT CO2 = 1.6 o C dT ESS = 2.3 o C Lunt et al, Nature Geoscience, 2010 CO2 = 400ppmv…

Model dependence?? PlioMIP Haywood et al., GMD, (2011)

Vegetation: Salzmann et al., GPC, (2008) New boundary conditions (PRISM3)… Hill et al (2007) Ice: Orography: Based on Markwick (2007) : derived from lithologic, tectonic and fossil information, the lithologic databases of the Chicago Paleogeographic Atlas Project, and DSDP/ODP data.

Hadley 3.3 o C GISS 2.1 o C MIROC 3.5 o CCCSM4 1.8 o C PlioMIP ensemble: Haywood et al, in prep NOR-ESM 3.2 o CECHAM 3.3 o C CO2 ~ 400ppmv…

The mid-Pliocene: summary Initial work suggested ESS ~ 4.6 o C, ESS/CS ~ 1.5 PlioMIP with new boundary conditions suggests ESS ~ 3.6 o C – 7 o C, (ESS/CS ~ 2 for HadCM3) Future – more Earth system components into the models, assessment of orbital contribution.

(1) CASE STUDY 2: The early Eocene …. Eocene: boundary conditions Palaeogeography (~ Ma)

Eocene: boundary conditions (2) CO ppmv = 6* 1120ppmv = 4* 560ppmv = 2* Beerling and Royer, Nature Geoscience, (2011)

Runs at 1,2,4,6 * pre-industrial CO 2. ~3500 years Global mean temperature vs CO 2 forcing CS ~ 4.5 o C CS ~ 3.3 o C Lunt et al, Geology, 2010.

Global mean temperature vs CO 2 forcing Now include vegetation feedbacks (i.e. consider a closer approximation to Earth system sensitivity….) CS ~ 3.3 o C ESS ~ 5 o C Loptson et al, in prep.

Model dependence?? EoMIP: HadCM3 (Lunt et al, 2011), CCSM (Huber and Caballero, 2011), GISS (Roberts et al, 2010), ECHAM (Heinemann et al, 2009), CCSM (WInguth et al, 2010) terrestrial annual mean SST as a function of CO 2 and model..

Model results compared with data (terrestrial) Data points from Huber and Caballero, CP, Leaf-margin analysis and/or CLAMP, + isotopic, geochemical, and pollen estimates. Lunt et al, in prep.

The early Eocene: summary Vegetation feedbacks increase sensitivity by ~50% EoMIP suggests models can produce good agreement with (most) data (but over a wide range of CO 2 ). Future – more coordinated modelling study.

Outlook Dont believe absolute numbers from a single model!! But….. Real potential for the palaeo community to make a crucial contribution to future climate predictions, by quantifying and scoring model performance under warm climate conditions….but…. Need a consistent and comprehensive analysis of uncertainty in both temperature and CO 2 estimates.