• Hello, please take a minute to check out our awesome content, contributed by the wonderful members of our community. We hope you'll add your own thoughts and opinions by making a free account!

Wintry 1/9-12 Winter Potential Great Dane or Yorkie

So I went through weatherbell and pivotal to look at thermals for mby and it was all snow. 0.6-0.7” of liquid fell. Dewpoints in the single digits to start with temps falling between 29-31 for the duration. Thermals through the entire column were actually solid. Likely 10:1-12:1 ratios if I had to guess. Probably a solid 6-8” event.
 
jeremyt said:
I understand your sentiment but I think to be fair it needs a few years of results to take it either way!
Click to expand...
Folks have expressed various opinions regarding whether they like the Euro AI / don't like it. The published data over the last 3 winters have the GFS Graphcast and Euro AI as the top performing models at 500mb at Day 5. On the chart, AIFS is the Euro AI. HRES/IFS is the Euro

pJ7PJ0q.jpeg
 
griteater said:
Folks have expressed various opinions regarding whether they like the Euro AI / don't like it. The published data over the last 3 winters have the GFS Graphcast and Euro AI as the top performing models at 500mb at Day 5. On the chart, AIFS is the Euro AI. HRES/IFS is the Euro

pJ7PJ0q.jpeg
Click to expand...
i think that some of that perception people have is because of how coarse the ai maps are, "it nailed this snowline" is easier to remember than "it was marginally better than everyone else at predicting this block". 500 mb is much easier to measure
 
I know it's just a computer generated forecast, but TWC app has a chance of rain/snow all around the Chattanooga area next Friday with their standard "Watching a potential winter storm" to go along with their basic forecast. Same wording in Dalton. Ga. Knoxville, Dunlap, Tn., etc. Let's see where this takes us. 🤔
 
Uncle Nasty said:
I know it's just a computer generated forecast, but TWC app has a chance of rain/snow all around the Chattanooga area next Friday with their standard "Watching a potential winter storm" to go along with their basic forecast. Same wording in Dalton. Ga. Knoxville, Dunlap, Tn., etc. Let's see where this takes us. 🤔
Click to expand...
Chattanooga is in a good spot.
 
Chazwin said:
To me this ultimately what we want as opposed to the fully phased amped up monster that we’ve seen on some model runs.
Click to expand...
Agree. Pull in the moisture, then slide it west to east (colder). Better view here from the WBell images. Surprised a bit at the deep tap of moisture (baja low helping in that regard)

C7p1omS.gif
 
griteater said:
Folks have expressed various opinions regarding whether they like the Euro AI / don't like it. The published data over the last 3 winters have the GFS Graphcast and Euro AI as the top performing models at 500mb at Day 5. On the chart, AIFS is the Euro AI. HRES/IFS is the Euro

pJ7PJ0q.jpeg
Click to expand...
I developed a Pangu-Weather ensemble for tropics.. it did really well with cyclones this year. About to fire it up for this upcoming system. According to that, has comparable accuracy to ECMWF OP/IFS.
 
ILMRoss said:
i think that some of that perception people have is because of how coarse the ai maps are, "it nailed this snowline" is easier to remember than "it was marginally better than everyone else at predicting this block". 500 mb is much easier to measure
Click to expand...
Ross - what are you referring to with respect to coarse maps on the AI? These look pretty detailed here. One thing I do notice is that all I see are maps at 6 hour intervals, and not 3 hour. Oh, I suppose you mean the model resolution?

tWPjtmB.png


FBkdmPx.png
 
jeremyt said:
A lot of Mets that I have followed have said they don’t put much stock in/ like the 6z 18z models but seem to like the 0/12 better. Not sure why , I have asked and never got an answer if any one time frame is better than another.
Click to expand...
At one time, this was answered and the given answer was that even though the 6z ran off of the 00z and the 18z ran off of the 12z, there was no negligible difference as far as which one verified correctly more of the time.
 
Psalm 148:8 said:
At one time, this was answered and the given answer was that even though the 6z ran off of the 00z and the 18z ran off of the 12z, there was no negligible difference as far as which one verified correctly more of the time.
Click to expand...
Only hires cams run from previous global runs (ie 18z HRRR/RAP initializes with 12z GFS forecast, nudged towards observations made at 18z with data assimilation).. Global models are fresh runs at each 18z, 00z, 06z, 12z.

And the most recent run is considered the most accurate. Any inter-run differences in accuracy due to 'off-run' data availability are far exceeded by gains in accuracy as you reduce lead-time.
 
griteater said:
Ross - what are you referring to with respect to coarse maps on the AI? These look pretty detailed here. One thing I do notice is that all I see are maps at 6 hour intervals, and not 3 hour. Oh, I suppose you mean the model resolution?

tWPjtmB.png


FBkdmPx.png
Click to expand...
sorry, yeah, resolution. better way to say it. just a theory. but when maps get shown here they remind me of the old CMC maps from that french language site we had to use in the early/mid 2010s lol. when i look at them on tidbits they give me navgem vibes
 
bouncycorn said:
Only hires cams run from previous global runs (ie 18z HRRR/RAP initializes with 12z GFS forecast, nudged towards observations made at 18z with data assimilation).. Global models are fresh runs at each 18z, 00z, 06z, 12z.

And the most recent run is considered the most accurate. Any inter-run differences in accuracy due to 'off-run' data availability are far exceeded by gains in accuracy as you reduce lead-time.
Click to expand...
i think the perception of the 00z and 12z being of higher mettle is because this is when balloons get launched and that data is ingested
 
IMG_0583.gif
GFS has been consistent as well for this upcoming OH valley/MA storm, aside from the 18z run on the 31st. Don’t have a gif but euro looks to have been pretty steady as well.

Obviously different storms, different conditions before, different stream interactions etc. So, this may be a more “predictable” ( less complex) setup. Certainly one or some of the models are out to lunch on our yorkie/Great Dane system, but just wanted to show that the euro AI was not alone in its consistency in depicting a snow footprint (that may not even verify).

Is that one run of the gfs that was too far south (18z 12/31) noteworthy potentially? The euro ai had already locked on while it took a few more cycles for the gfs to do so, and we’re about at that range right now. TBH may be wise to just check back at 18z tomorrow, and see where things lie.
 
You must log in or register to reply here.
Back
Top