1 |
heimbach |
1.1 |
Improving ENSO simulations and predictions through |
2 |
|
|
the ECCO ocean state estimation |
3 |
|
|
|
4 |
|
|
Dietmar Dommenget and Detlef Stammer |
5 |
|
|
|
6 |
|
|
The global ECCO ocean state estimation procedure is used to simulate |
7 |
|
|
and forecasts of tropical Pacific SST and sub-surface fields on seasonal |
8 |
|
|
time scale. As compared to a traditional ENSO assimilation and prediction |
9 |
|
|
procedures, the hindcast of the constrained ocean state is much closer to |
10 |
|
|
observed conditions. The skill of the 12 month lead SST forecast in the |
11 |
|
|
equatorial Pacific appears to be similar in both assimilation, but |
12 |
|
|
the optimizations leads to better skills in the SST anomaly correlations. |
13 |
|
|
Although our current forecast system is strongly limited by the estimation |
14 |
|
|
of atmospheric forcing corrections, the results see to suggest that |
15 |
|
|
the forecast system would benefit if adoint optimizations are performed |
16 |
|
|
with coupled ocean-atmosphere systems in which the coupling is |
17 |
|
|
part of the control vectors of the adjoint optimization. |
18 |
|
|
|
19 |
|
|
|