Global model runs discussion

[blp]

Well so much for consistency on the 18z GFS. Now it shows a 999mb inside Africa at 384hr I don't know why I even bother to look at it.

[ninel conde]

so back to nothing?

[Yellow Evan]

Well so much for consistency on the 18z GFS. Now it shows a 999mb inside Africa at 384hr I don't know why I even bother to look at it.

I guess that's possible. Christine 73 formed over Africa.

[CourierPR]

I believe I recall reading that Donna, in 1960, was a storm when it hit the water off Africa.

[ninel conde]

Well so much for consistency on the 18z GFS. Now it shows a 999mb inside Africa at 384hr I don't know why I even bother to look at it.

I guess that's possible. Christine 73 formed over Africa.

actually from hr 120-384 it has rather low pressure over africa but nothing ever gets going over water.

[Ptarmigan]

Well so much for consistency on the 18z GFS. Now it shows a 999mb inside Africa at 384hr I don't know why I even bother to look at it.

That is why I do not take the 16 day forecast seriously.

[gatorcane]

so back to nothing?

That's what it looks like. The 18Z GFS shows basically nothing developing through Aug. 28th. It even shows the possible Gulf system as a weak and sheared system.

[ninel conde]

so back to nothing?

That's what it looks like. The 18Z GFS shows basically nothing developing through Aug. 28th. It even shows the possible Gulf system as a weak and sheared system.

yea, maybe the models will show something later in sept.

[CYCLONE MIKE]

By the time models start to show something, it will probably be time to start looking north for winter

[Jagno]

so back to nothing?

That's what it looks like. The 18Z GFS shows basically nothing developing through Aug. 28th. It even shows the possible Gulf system as a weak and sheared system.

Although we are in awe of what mother nature can do, I'm not disappointed at "weak or sheared" in the Gulf.

[chaser1]

Well so much for consistency on the 18z GFS. Now it shows a 999mb inside Africa at 384hr I don't know why I even bother to look at it.

That is why I do not take the 16 day forecast seriously.

I disagree. This time of year, I do start taking the model forecast TRENDS seriously. What happened on the 18Z run, a "blip"? The fact that last night's 0Z run and today's 12Z run might have showed a 998mb Tropical Storm - inland over Africa, and one run later either that low either not be there or perhaps is weaker and farther east.... is not in the least bit relevant. If it was, who would need the NHC? We could just have "John Q. Public" run a 6Z GFS run and absolutely know where (or where not) a Tropical Cyclone will be in 384 hrs! It just doesn't work that way.

Try taking a few steps back, and just "smell the sargassum tea leaves for a moment". Now, make a few notes of nuances that you may recently recall from recent GFS runs & start to compare them to the current runs. First of all, think how recently it was when everyone was consistently observing, "model runs not even depicting waves", its been forever since any model has even attempted to carry a low in the Atlantic basin...even our "crazy Canadian uncle" CMC model. Suddenly, the NAM starts developing something from the Caribbean. Sure, its just "the NAM", but wait.... then the Canadian model is jumping in...., and then the FIM model too. Also, remember that deep low on the 4 Million hour GFS , that for one run suddenly vanished? Well, go back to that last GFS run where you last recall seeing it. Now, look at what other weaker lows between Africa and 60W that you see there. Now run the current GFS and pay less attention to the MegaStorm which no longer appears to be coming off the coast at about 280 hrs-320hrs, but look to see if you continue to see the same weaker lows tracking west in the easterly trades. In fact, are there more disturbances starting to show up and perhaps with lower pressures than previous? I am no doubt intrigued by how at times the GFS can latch onto a system way out, and as time progresses, a couple of these systems actually come to fruition. That said, it is almost never quite as far East or as far North as earlier runs quite showed, perhaps the budding Depression didn't form until 10 degrees farther West (or perhaps earlier and farther East). That level of accuracy is entirely inconsistent, but just like using a metal detector searching for old coins or artifacts, when that thing really starts "pinging" consistently, its a pretty good bet that there's some kind of stuff a few feet down.....

Well, from what I can tell....that "pinging" is on an uptick. Two or three days from now, we could be looking at a GFS run AND a Euro run both depicting forming Tropical Storms soon to form. If all this sounds like a lot of unscientific gibberish, well....it is. But while some are drowning in their own tears because the season is nearly over and wont have the chance to track and appreciate how these systems develop, you might want come on back and watch the season start - because it's about to do just that.

On a more scientific note however, take a look at the current GFS model for surface pressures. Find the 1016mb isobar and look how it more or less travels from S. Florida to the Cape Verde Islands. Now, look for the 1012mb isobar which is largely limited to the S. Caribbean. As you progress forward in time, notice how the pressures start to fall and it becomes nearly common for the 2012mb line to traverse much of the Caribbean and E. Atlantic around 15N most hr. forecasts. Mother Nature is quite soon going to shrug off the dust, take advantage of the lower surface pressures, tap into the consistently warming "go-go" juice (indicative of all recent SST analysis) and pump out "a big fattie" named Erin. My guess is when that switch suddenly goes off, we'll likely see at least a couple more within days or a week at the most. Remember, 6 or 7 days ago...., the most any models were showing in the Caribbean was perhaps a weak wave or "pouch". Now, in well less than 384 hours we might be looking at a forming Tropical Depression in approx. 72 hours.

[weatherwindow]

Good morning, Chaser et al...speaking of our crazy uncle, the CMC/GEM 00Z..finally a little love, drum roll, our first modelled CV hurricane...a sweet 980 mbs north of Northern Leeward at 240hrs, and a 1000mb landfalling TS for the Panhandle at 114hrs ...a little sanity in a model world gone mad

http://www.tropicaltidbits.com/analysis ... opics.html

[chaser1]

....speaking of our crazy uncle, the CMC/GEM 00Z..finally a little love, drum roll, our first modelled CV hurricane...a sweet 980 mbs north of Northern Leeward at 240hrs, and a 1000mb landfalling TS for the Panhandle at 114hrs ...a little sanity in a model world gone mad

Now, THATS funny! But you wanna know something?? Its periods of a week or so like this when I almost have to just stick my head in the sand and feel like hiding, because no different than a few here, my friends or family will also start to call and start the "guess its gonna be a slow season after all...chant LOL. Too funny that we're almost a bit relieved to see the CMC of all models, start doing what it does so well!

[cycloneye]

The 00z ECMWF at least has a small green circle in MDR.

[weatherwindow]

The 00z ECMWF at least has a small green circle in MDR.

Good morning, Luis....two models indicating honest-to-goodness development in the MDR...grab the Valium ...Grtz from KW, Rich

[gatorcane]

The 06Z GFS is showing a weak Cape Verde low staying in tact for many days where it heads slowly west. Once we lose the finer resolution after 180 hours it loses it but this could be something to watch to see if later GFS runs show more favorable condtions once this low gets farther west. You can see the low at about 18N, 45W on Aug 20th:

[Ptarmigan]

I disagree. This time of year, I do start taking the model forecast TRENDS seriously. What happened on the 18Z run, a "blip"? The fact that last night's 0Z run and today's 12Z run might have showed a 998mb Tropical Storm - inland over Africa, and one run later either that low either not be there or perhaps is weaker and farther east.... is not in the least bit relevant. If it was, who would need the NHC? We could just have "John Q. Public" run a 6Z GFS run and absolutely know where (or where not) a Tropical Cyclone will be in 384 hrs! It just doesn't work that way.

Try taking a few steps back, and just "smell the sargassum tea leaves for a moment". Now, make a few notes of nuances that you may recently recall from recent GFS runs & start to compare them to the current runs. First of all, think how recently it was when everyone was consistently observing, "model runs not even depicting waves", its been forever since any model has even attempted to carry a low in the Atlantic basin...even our "crazy Canadian uncle" CMC model. Suddenly, the NAM starts developing something from the Caribbean. Sure, its just "the NAM", but wait.... then the Canadian model is jumping in...., and then the FIM model too. Also, remember that deep low on the 4 Million hour GFS , that for one run suddenly vanished? Well, go back to that last GFS run where you last recall seeing it. Now, look at what other weaker lows between Africa and 60W that you see there. Now run the current GFS and pay less attention to the MegaStorm which no longer appears to be coming off the coast at about 280 hrs-320hrs, but look to see if you continue to see the same weaker lows tracking west in the easterly trades. In fact, are there more disturbances starting to show up and perhaps with lower pressures than previous? I am no doubt intrigued by how at times the GFS can latch onto a system way out, and as time progresses, a couple of these systems actually come to fruition. That said, it is almost never quite as far East or as far North as earlier runs quite showed, perhaps the budding Depression didn't form until 10 degrees farther West (or perhaps earlier and farther East). That level of accuracy is entirely inconsistent, but just like using a metal detector searching for old coins or artifacts, when that thing really starts "pinging" consistently, its a pretty good bet that there's some kind of stuff a few feet down.....

Well, from what I can tell....that "pinging" is on an uptick. Two or three days from now, we could be looking at a GFS run AND a Euro run both depicting forming Tropical Storms soon to form. If all this sounds like a lot of unscientific gibberish, well....it is. But while some are drowning in their own tears because the season is nearly over and wont have the chance to track and appreciate how these systems develop, you might want come on back and watch the season start - because it's about to do just that.

On a more scientific note however, take a look at the current GFS model for surface pressures. Find the 1016mb isobar and look how it more or less travels from S. Florida to the Cape Verde Islands. Now, look for the 1012mb isobar which is largely limited to the S. Caribbean. As you progress forward in time, notice how the pressures start to fall and it becomes nearly common for the 2012mb line to traverse much of the Caribbean and E. Atlantic around 15N most hr. forecasts. Mother Nature is quite soon going to shrug off the dust, take advantage of the lower surface pressures, tap into the consistently warming "go-go" juice (indicative of all recent SST analysis) and pump out "a big fattie" named Erin. My guess is when that switch suddenly goes off, we'll likely see at least a couple more within days or a week at the most. Remember, 6 or 7 days ago...., the most any models were showing in the Caribbean was perhaps a weak wave or "pouch". Now, in well less than 384 hours we might be looking at a forming Tropical Depression in approx. 72 hours.

I understand your point. On the topic of 16 day forecast, the GFS is the only one that does it besides CFSv2 (more of a climate forecast model). CMC and EURO go up to 10 days, while NAM goes up to 4 days. It would be nice if CMC, EURO, and even NAM have forecast up to 16 days like GFS, so they can be compared. I do not know why CMC, EURO, and even NAM do not go up to 16 days.

A little off topic, but related, I remember the GFS model saw the February 2011 Freeze 16 days in advance and it turned out to be right. In that case, there was consistency. Another thing about GFS is when they forecast something 16 days, it disappears or changes 10 to 14 days and than reappears in 8 days or less. I know it does not always happen, but I have seen it happen in the past.

[USTropics]

I understand your point. On the topic of 16 day forecast, the GFS is the only one that does it besides CFSv2 (more of a climate forecast model). CMC and EURO go up to 10 days, while NAM goes up to 4 days. It would be nice if CMC, EURO, and even NAM have forecast up to 16 days like GFS, so they can be compared. I do not know why CMC, EURO, and even NAM do not go up to 16 days.

A little off topic, but related, I remember the GFS model saw the February 2011 Freeze 16 days in advance and it turned out to be right. In that case, there was consistency. Another thing about GFS is when they forecast something 16 days, it disappears or changes 10 to 14 days and than reappears in 8 days or less. I know it does not always happen, but I have seen it happen in the past.

Two reasons. First, the GFS is actually 2 models that were combined together, the medium-range forecast, or MRF, and the aviation model, or AVN (in actuality, the GFS model is a coupled model composed of four seperate models: an atmosphere model, an ocean model, a land/soil model, and a sea ice model). Thus the model is run in 2 parts, with the first part having a higher resolution out to 192 hours (8 days), and the second part is run from 192-384 hours (16 days) at a lower resolution. This takes a tremendous amount of computing power to do 4, 16 day runs each day.

[CalmBeforeStorm]

http://mag.ncep.noaa.gov/Imageanis.php

The 12z GFS tracks a 1009-1012 low in the MDR that ends up at about 18n 42w but moving west under a nice ridge. Hard to believe that a low in that position at this time of year wouldn't develop.

[Blown Away]

I'm still puzzled why no significant activity being shown in these long range GFS/Euro models given the above average predictions...GFS 384 gets you to September, so any CV system at that point would be 10 days off before any land threat...I guess I'm use to past models showing so many phantom storms and maybe these updated models are getting so precise that they are only supposed to show @15 storms this season...Plenty of time for that...