Perturbed physics ensemble using the MIROC5 couple

Authors: Hideo Shiogama Masahiro Watanabe Masakazu Yoshimori Tokuta Yokohata Tomoo Ogura James D Annan Julia C Hargreaves Manabu Abe Youichi Kamae Ryouta O’ishi Rei Nobui Seita Emori Toru Nozawa Ayako AbeOuchi Masahide Kimoto

Publish Date: 2012/07/22

Volume: 39, Issue: 12, Pages: 3041-3056

PDF Link

Abstract

In this study we constructed a perturbed physics ensemble PPE for the MIROC5 coupled atmosphere–ocean general circulation model CGCM to investigate the parametric uncertainty of climate sensitivity CS Previous studies of PPEs have mainly used the atmosphereslab ocean models A few PPE studies using a CGCM applied flux corrections because perturbations in parameters can lead to large radiation imbalances at the top of the atmosphere and climate drifts We developed a method to prevent climate drifts in PPE experiments using the MIROC5 CGCM without flux corrections We simultaneously swept 10 parameters in atmosphere and surface schemes The range of CS estimated from our 35 ensemble members was not wide 22–32 °C The shortwave cloud feedback related to changes in middlelevel cloud albedo dominated the variations in the total feedback We found three performance metrics for the present climate simulations of middlelevel cloud albedo precipitation and ENSO amplitude that systematically relate to the variations in shortwave cloud feedback in this PPEClimate sensitivity CS which is defined as the global mean surface air temperature response to a doubling of the atmospheric CO2 concentration is a crucial piece of information that informs the adaptation and mitigation policies for anthropogenic climate change Despite the considerable efforts of climate scientists and technical advances the ranges of the CS have not been narrowed Knutti and Hegerl 2008 In the multimodel ensemble MME of general circulation models GCMs used for the Intergovernmental Panel on Climate Change Fourth Assessment Report the range of CS was 21–44 °C Randall et al 2007 The variation in the CS in the MME is caused by the use of different model structures ie different physical parameterization schemes and resolutions Therefore this is known as “structural uncertainty” Murphy et al 2004 2007The “parametric uncertainty” is another substantial uncertainty The present climate biases and future climate changes in a single model may be sensitive to changes in parameter values in the model physical schemes Murphy et al 2004 The Met Office Hadley Centre’s project “Quantifying Uncertainty of Model Predictions” QUMP and http//climatepredictionnet are the first and most comprehensive projects to investigate the parametric uncertainties of climate responses to external forcing Murphy et al 2004 2007 Stainforth et al 2005 Webb et al 2006 Collins et al 2006 2007 2011 Brierley et al 2010 Jackson et al 2011 These investigators constructed several perturbed physics ensembles PPEs in which they swept uncertain parameters within the HadCM3 model Gordon et al 2000 The variation in the CS in their PPEs was comparable to or greater than the variation in the CS in the MME Murphy et al 2004 Stainforth et al 2005 Collins et al 2011A PPE that uses a different GCM ECHAM5 also has a CS variation that is comparable to the MME Klocke et al 2011 However these wide ranges in CS are not universal The PPEs used versions of the CAM3 GCM Jackson et al 2008 Sanderson 2011 and the EGMAM GCM Niehörster and Collins 2009 consistently yielded a CS of less than 3 °C whereas the PPEs of MIROC3 produced a CS greater than 4 °C Annan et al 2005 Yokohata et al 2010 Yoshimori et al 2011 The distributions of CS in PPEs are contingent upon the model structures as well as the experimental design Intercomparison studies of multiPPEs have only recently begun and can facilitate further understanding of the structural and parametric uncertainties of climate responses to external forcing Yokohata et al 2010 Sanderson 2011Although previous PPE approaches have been useful they have limitations Most of the previous PPE studies used atmosphere/slabocean mixed layer ocean GCMs ASGCMs rather than coupled atmosphere/fullocean GCMs CGCMs Murphy et al 2004 Stainforth et al 2005 Annan et al 2005 Sanderson 2011 Klocke et al 2011 One of the reasons for this is that the computational costs required to reach equilibrium for the ASGCMs are lower than those for the CGCMs However the climate feedback may differ between ASGCMs and CGCMs Boer and Yu 2003a Yokohata et al 2008 Williams et al 2008A few studies have performed CGCM PPEs to move beyond this limitation of PPEs in ASGCMs Collins et al 2006 2007 2011 Brierley et al 2010 Jackson et al 2011 Rowlands et al 2012 However another problem remains Changes in values of atmosphere and surface parameters can lead to larger net radiation imbalance at the top of atmosphere TOA and the climate drifts To prevent large climate drifts most previous CGCM PPE studies have applied corrections for ocean surface heat and salinity fluxes note that ASGCM PPEs also require flux corrections However flux corrections can affect the CS because of changes in the climatology of the sea surface temperature cloud distribution sea ice and other parameters in the control simulationJackson et al 2011 performed a QUMP CGCM PPE without flux corrections using atmosphere and surface parameter values in the members of the ASGCM PPE with a relatively small TOA imbalance Because their approach was based on the existence of the ASGCM PPE other modeling groups cannot easily apply itIn this study we developed a CGCM PPE without flux corrections Our method utilizes a preliminary ensemble of atmospheric GCMs AGCMs which have much lower computational costs to reach equilibrium than ASGCMs or CGCMs Therefore this method would be useful for other modeling groups for performing CGCM PPEs without flux corrections

Keywords: