Under SMIP, two experimental protocols using atmospheric general circulation models were set up up to investigate: (1) potential seasonal predictability (SMIP-2) using observed SST, and (2) actual predictability using forecast SST (SMIP-2/HFP). These projects include international contributions from DEMETER, PROVOST, DSP and APCC. These coordinated numerical experiments clearly demonstrate that successful seasonal prediction is a multi-model and multi-institutional problem.
1st and 2nd season potential predictability based on initial conditions and the AGCM response to specified observed values of SST and sea-ice.
1st season actual predictability based on inital conditions and the AGCM response to predicted values of SST and sea-ice or based on the results of a coupled atmosphere-ocean forecast system.