Some lessons from SMIP 2 History of SMIP

Some lessons from SMIP 2?

History of SMIP 2 n n SMIP 2 - 1 st and 2 nd season potential predictability - AGCM using observed SST and sea-ice - initialize from reanalysis SMIP 2/HFP - 1 st season actual predictive skill in operational context – no information from the future - 2 -tier forecast or 1 -tier forecast with coupled system - initialized from reanalyses

Why SMIP 2 n n n New knowledge on SP Intercomparison of models in the SP context Measures of potential and/or actual “predictability” in current models Diagnostic subprojects to entrain “outside expertise” to analyze results Collection of results for research into multimodel methods (forerunner/prototype/precursor/exemplar of TFSP 1 st Experiment)

Basics

Declared participants and data

Participation n n Modest participation despite efforts to recruit Reasons…? - problem not of interest - lack of a clear output (i. e. compare IPCC) - funding - timing/other priorities - organizational support

Some problems n n n Scientific interest - SMIP 2 was much like SMIP 1 - SMIP 2/HFP tried to broaden but perhaps too late - no specific scientific output such as an analysis/Report Data/organizational support - PCMDI commitment to central data repository weak … organizational changes + IPCC - IRI commitment modest - passive rather than active data collection Other commitments – espc. IPCC

Funding/competing priorities Cli. PAS/APCC -funded project able to attract participants -funded Workshops (IPRC) -active data collection

Lessons for future WGSIP projects n n Clear scientific goal - includes commitment to basic analysis and a Report - diagnostic subproject support Clear data policy - IPCC shows the virtue of a “dedicated” data repository - active data policy of “recruiting” data is preferred - otherwise KISS including amount of data Funding - commitment of funding a major asset - research, workshops, etc. Organizational support - Website, Workshop(s) etc

End of lesson