Long Range GFS question...

[Fabian Guerra]

I am not a met or in school but a chaser doing a little personal study on the accuracy of the long range GFS (5-15 days out). I am looking at and logging data on all 4 daily runs 5-15 days out. I will then compare the data to what actually happens during the chase season. I've known for a while that you cannot trust the accuracy of the GFS too far out but it's really something to actually see the way it flip-flops so drastically from run to run on a consistent basis.

My question... Are the 4 daily GFS runs (0z, 6z, 12z, 18z) equal to each other in terms of how much information they each use to come up with their solutions or is there a particular run that is generally more comprehensive and/or accurate? Either way I will still gather data from each run and come to my own theories and conclusions but it would be nice to hear any other opinions or insights on this subject. Thanks.

 

[Chad Cowan]

My question... Are the 4 daily GFS runs (0z, 6z, 12z, 18z) equal to each other in terms of how much information they each use to come up with their solutions or is there a particular run that is generally more comprehensive and/or accurate?

I will let those more familiar with this expand, but it's my understanding that the 6z and 18z runs are considered the "off-hour" runs because they don't use new RAOB data like the 12z and 0z runs do. I've found that the off-hour runs are more susceptible to hiccups, especially beyond 5 days.

 

[Chris Nuttall]

I will let those more familiar with this expand, but it's my understanding that the 6z and 18z runs are considered the "off-hour" runs because they don't use new RAOB data like the 12z and 0z runs do. I've found that the off-hour runs are more susceptible to hiccups, especially beyond 5 days.

That's pretty much it. As far as model accuracy goes, I've found the ECMWF to have a much better handle on the overall pattern beyond Day 5.