skyline385 wrote:Steve wrote:You have to pick and choose how you use it as well as every other model. No one in their right mind is going to use ICON for an 8 day out intensity prediction - or any of the globals (except for generalizations). OTOH, it has the position of Earl pretty damn good from over 8 days out.
ICON has proven to me that it's best used for predicting genesis of weaker and smaller systems. Once you learn or accept certain biases inherent in each model, you can avoid throwing out the proverbial baby with the bathwater and adjust accordingly.
Thats pretty much why the GFS v16 has such high cyclone spawn rates, it was made that way to detect TCG earlier than previous versions even though it has gone a bit overboard this year. Obviously the GFS is now taken with caution knowing its high TCG bias when out in the future and ICON is now in that category as well. IIRC ICON used to be very conservative and this seems to be a recent change.
Also regarding the position of Earl, a 960mb Cat 3 system and a 995mb TS aren't going to be under the influence of the same steering current and Coriolis effect so more than likely the ICON got the position right simply by chance. It also had a ton of runs where it parked Earl near the Bahamas including one with a WSW dip. So this run is just one out of many possible locations it showed which is expected 8 days out.
Yeah, I've been running it and the others every day a few times while we were waiting. Like all of them, there's been a plethora of solutions and evolutions. Earl could be accidental luck. But it gets some stuff recognized out front which is why I mentioned it had the 1004mb low hitting NWFL this Friday.
^^ FWIW, the new ECMWF also has a low in the North Gulf but closer to the mouth of the Mississippi River in the same timeframe. So ICON definitely first on that, EC next in case anything shows up or spins up there in a couple days.