How about a primer on ECMWF and CFS multi-month forecasts

[Ed Lizard]

Thread started with the CFS implying blocking over the polar regions and a trough in the Northeast, and then added on relative importance of the ENSO, and the AO/NAO.

But what I, as an interested amateur doesn't know, and I suspect I am not alone, what are the Cliff Notes version of how the Euro and CFS seasonal forecasts? Are they just ensembles of the operational models, perhaps run with a different truncation number/resolution, but run out for months, instead of weeks? Similar physics? Or something different.

And, which is more reliable, the Euro seasonal multi-month forecasts or the CFS, the latest version. Do either have appreciable skill at long range trend forecasting.

Inquiring minds and all.

[dtk]

Thread started with the CFS implying blocking over the polar regions and a trough in the Northeast, and then added on relative importance of the ENSO, and the AO/NAO.

But what I, as an interested amateur doesn't know, and I suspect I am not alone, what are the Cliff Notes version of how the Euro and CFS seasonal forecasts? Are they just ensembles of the operational models, perhaps run with a different truncation number/resolution, but run out for months, instead of weeks? Similar physics? Or something different.

And, which is more reliable, the Euro seasonal multi-month forecasts or the CFS, the latest version. Do either have appreciable skill at long range trend forecasting.

Inquiring minds and all.

I don't have time to get into too many details, but they are simply dynamic model integrations using a model that is typically based on the parent operational NWP models (at least in terms of dynamics). However, because they are long integrations, they are run at much lower spatial resolution and are coupled (two-way) to an interactive ocean model. For example, I believe that the CFSv2 integrations are done at T126 (whereas the operational GFS is currently T574), but uses the same vertical coordinate.

I believe that the ECMWF tries to keep the model parameterizations similar between the operational and seasonal models, but the CFS has some differences from the GFS. The CFS is initialized from its own initial conditions (which uses a coupled model to drive the system) and does not simply use the operational GFS ICs.

Lastly, the CFS has a huge reforcecast database from which to derive calibrated products. I'm assuming the ECMWF does something similar.

[Ed Lizard]

Thanks.

[H2Otown_WX]

http://iri.columbia.edu/climate/ENSO/currentinfo/SST_table.html Good link with all the ENSO model predictions.

[Mashie]

http://iri.columbia.edu/climate/ENSO/currentinfo/SST_table.html Good link with all the ENSO model predictions.

Great link! Thanks