percentage of total variance over decade – associated with forced component – associated with internal variability p and p tend to be inverses of one another so p = p + p is more uniform than either pp p p Potential predictability of temperature for (“next decade”) Boer 2008
CMIP5 Experiment Design “Long-Term” (century & longer) TIER 1 TIER 2 CORE “realistic” diagnostic “Near-Term” (decadal) (initialized ocean state) prediction & predictability CORE TIER 1
Decadal forecast results to 2015 CCCma
U. Miami
© Crown copyright Met Office We are exchanging very basic quantities: Global Annual Mean Temperature One file for each year, each member Exchanged once per year around November Equal ownership Example diagnostics:
© Crown copyright Met Office Annual global temperature (20) (9) (3)
© Crown copyright Met Office Annual global temperature
© Crown copyright Met Office Surface temperature: 2011 relative to
© Crown copyright Met Office Surface temperature: 2012
© Crown copyright Met Office Surface temperature: 2013
© Crown copyright Met Office Surface temperature:
© Crown copyright Met Office Surface temperature:
© Crown copyright Met Office We’ve had an overwhelmingly positive response: GFDL – Tony Rosati Uni. Tokyo – Kimoto Masahide MRI – Masayoshi IshiiSMHI – Klaus Wyser,Colin Jones KNMI – Wilco HazelegerIC3 – Francisco Doblas- Reyes MPI – Daniela MateiRSMAS – Ben Kirtman CCCMA-EC – George Boer IfM-GEOMAR - Mojib Latif CERFACS – Laurent Terray UKMO-Hadley – Doug Smith Are others planning to make regular decadal predictions? If so, would you like to take part?