How reliable is the AVN model?

[charleston_hugo_veteran]

Anyone??

[Guest]

the AVN is a very good model. when taken the entire suite of it's own ensemble, and comparing against itself and other models, and historical data, you can really get a beat on developing patterns and systems. the AVN has it's strengths and weaknesses, and it takes some experience to understand where they lie. but generally AVN short and mid range does very well.

[Derecho]

Depends. (BTW, officially it's the GFS)

Beyond 168 hours or so for a tropical system? Basically worthless.

But the same is true of all other models....many of which don't even run out that far.

Of the global models, it was 2nd best for Charley (Canadian, in an idiot-savant moment, was the best, something I doubt you'll see repeated on other storms.)

It probably doesn't quite deserve the constant abuse it gets from a variety of internet meterologists and web posters. Since it runs 4 times a day (twice as often as a lot of other models) it has twice the opportunity to do something screwy.

[charleston_hugo_veteran]

thanks!

[Stormchaser16]

Derecho makes an excellent point about how many times it is run in one day. There are 4 chances for it to screw up, unlike lets say the EC which has a one shot deal(just recently it has switched to two) so when comparing against the EC, it sucks because there are more chances for things to change. Generally speaking however in the midrange the GFS usually performs good, I wouldnt say excellent or anything, as no model IMO is excellent, but it is definately not crap. Beyond 168 HRS tho, like said, it is useless garbage that is good for nothing but a laugh. This is also true of winter, within the short to mid range the GFS is usually good. Beyond that, one must really understand patterns and the GFS itself to understand what it is saying, because you can(even after 168 hrs.) get a general feel for what type of pattern may lie ahead. OF COURSE, the one rule to remember is other model support(in DT's rules to model guidance), if the GFS is the outlier, then you generally dont want to go with it, however if it is somewhat consistent and/or on track with other models then it should be taken into consideration for mid to long range forecasting.

[Guest]

Depends. (BTW, officially it's the GFS)

Beyond 168 hours or so for a tropical system? Basically worthless.

But the same is true of all other models....many of which don't even run out that far.

Of the global models, it was 2nd best for Charley (Canadian, in an idiot-savant moment, was the best, something I doubt you'll see repeated on other storms.)

It probably doesn't quite deserve the constant abuse it gets from a variety of internet meterologists and web posters. Since it runs 4 times a day (twice as often as a lot of other models) it has twice the opportunity to do something screwy.

not worthless. take all of the runs for the 12z for example, today. The actual path of the storm on every single run may be garbage, but the actual grids and information on each one is very useful. even to 240 hours.