 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
·
Exploit the 15 day operational
NCEP ensembles
|
|
|
|
·
Increase the length of the
ensembles to 30 days
|
|
|
|
·
Use multi-model ensembles
|
|
|
·
Exploit the DERF predictability
due to initial conditions,
|
|
SST and soil moisture in real time
|
|
|
|
·
Do routine SST and soil moisture
anomaly experiments
|
|
|
in real time
|
|
|
|
·
Improve the short-term coupled
forecasts by minimizing
|
|
growing
errors in the initial conditions
|
|
|
|
· (Improve the models!)
|
|