Hurricane Irene Model and Forecast Discussion

[MarkSC]

Anyone see the Euro ensembles and care to comment?

[am19psu]

Anyone see the Euro ensembles and care to comment?

They're right on top of the Euro ops.

[phlwx]

6z GFS shifts track about 30 miles east, eye across SE Mass (Cape Cod) but misses Hatteras.

[am19psu]

Not much to add this morning. Everyone is doing a good job with analysis in these threads. I like somewhere west of Morehead as a Cat 3, but that's pretty much consensus of everyone's forecast at this point.

[Scott747]

I'll preface by saying I'm well aware of the criticism of the HWRF from earlier. Nonetheless it still could potentially be used by the NHC as a reference in the upcoming disco....

0z HWRF -

6z HWRF -

[Ed Lizard]

Since the water is fairly warm to within about 100 miles of the coast, and a normal speed, slow out of the tropics may imply not terrible shear, a strong enough storm could still be decent by New York/New England.

I know the 50 mph Plus forward speed of 1938 kept it over cold water a very short time,, but I'd bet mean low level flow wasn't 50 or 60 mph, and that system was sheared to heck and mainly got by on slow filling of what was once Cat 5 pressure. Screaming winds aloft no doubt mixed down as well.

[am19psu]

I didn't see this posted anywhere, but we won't have G-IV data for the 23/12z runs, however the 24/00z and 24/12z will have it.

[Scott747]

0z GFDL -

6z GFDL -

[catoctin wx]

Here is the NOAA G-IV track that will be used to assimilate dropsonde data into tonights 00Z model runs.

[mattmfm]

Modeled forward speeds has really picked up as well. The idea of this stalling and/or crawling up the coast appears less of a threat. On 18z GFS Irene didn't reach SNE until 18z Monday. On 6z GFS it makes landfall 18z Sun. So a 24 hour increase in movement, in only two model runs.

[mattmfm]

6z GFDL -

And with that there's now not a single model showing a landfall west of the NC/SC border.

[CT Rain]

And with that there's now not a single model showing a landfall west of the NC/SC border.

06z GFDL is also substantially faster than the 00z run was.

[mattmfm]

06z GFDL is also substantially faster than the 00z run was.

Yup. 00z run at 06z Sun was east of the GA/SC border. 06z run is over HSE. I'm fairly confident climatology of EC storms will play out closer to reality and we're not likely to see a system crawling up the coast north of HSE.

[Sunny and Warm]

Yup. 00z run at 06z Sun was east of the GA/SC border. 06z run is over HSE. I'm fairly confident climatology of EC storms will play out closer to reality and we're not likely to see a system crawling up the coast north of HSE.

However, the synoptic setup is unlike typical climo. The fact that trough #2 misses Irene causes the westerlies to recede into Canada instead of drive her OTS. At that point, the only steering mechanism is the high to the east building back in after the trough lifts out.

[CT Rain]

However, the synoptic setup is unlike typical climo. The fact that trough #2 misses Irene causes the westerlies to recede into Canada instead of drive her OTS. At that point, the only steering mechanism is the high to the east building back in after the trough lifts out.

That is true... the westerlies are awfully far north and let's remember we're still in August. It's easy for a storm to crawl up the coast this time of year than it is even 2 or 3 weeks later in September.

I think the most likely scenario may be a HSE scraper then out to sea to the southeast of Nantucket.

[WxFreak11]

Local Mets here in Charleston just said that new computer models indicates Irene will be much further offshore and maybe OBX will be affected but although not in the clear here yet... it is looking better for us here in Charleston.

[LakeEffectKing]

That is true... the westerlies are awfully far north and let's remember we're still in August. It's easy for a storm to crawl up the coast this time of year than it is even 2 or 3 weeks later in September.

I think the most likely scenario may be a HSE scraper then out to sea to the southeast of Nantucket.

If that pans out, the Euro's 4DVAR assimilation technique will take a hit.

[JamieOber]

CTP had this in their short term AFD:

SFC HIGH PRES WILL SLIDE OFF THE MID ATLC COAST BY WED MORNING, GIVING WAY TO RETURN FLOW/WAA PATTERN. A SHORTWAVE TROUGH AND ASSOCIATED COLD FRONT WILL MOVE FROM THE UPPER MIDWEST THROUGH THEGREAT LAKES INTO THE NORTHEAST STATES THROUGH THE PERIOD. 23/00Z MODELS SHOW RELATIVELY SIMILAR TIMING...WITH THE NAM ON THE SLOWSIDE OF THE DETERMINISTIC CONSENSUS AND THE GFS SLIGHTLY FASTER. THE ECMWF HAS TRENDED SLIGHTLY FASTER THAN ITS PREVIOUS 12Z RUNAND TOWARDS THE GFS. THE ISSUE IS NOT AS MUCH TIMING AS IT IS AMPLITUDE. RECENT HPC VERIFICATION OF THE PREVAILING LONGWAVE PATTERN SHOWS THAT ALL MODELS HAVE UNDER-AMPLIFIED THE TROUGH IN THE EAST. THEREFORE...GIVEN PERSISTENCE OF THE CURRENT LONG WAVE PATTERN THROUGH F84...A SLIGHTLY GREATER AMPLIFICATION OF THETROUGH THAN DEPICTED BY CURRENT GUIDANCE IS RECOMMENDED.

Not sure how that affects things but was wondering what others thought.

[WidreMann]

CTP had this in their short term AFD:

SFC HIGH PRES WILL SLIDE OFF THE MID ATLC COAST BY WED MORNING, GIVING WAY TO RETURN FLOW/WAA PATTERN. A SHORTWAVE TROUGH AND ASSOCIATED COLD FRONT WILL MOVE FROM THE UPPER MIDWEST THROUGH THEGREAT LAKES INTO THE NORTHEAST STATES THROUGH THE PERIOD. 23/00Z MODELS SHOW RELATIVELY SIMILAR TIMING...WITH THE NAM ON THE SLOWSIDE OF THE DETERMINISTIC CONSENSUS AND THE GFS SLIGHTLY FASTER. THE ECMWF HAS TRENDED SLIGHTLY FASTER THAN ITS PREVIOUS 12Z RUNAND TOWARDS THE GFS. THE ISSUE IS NOT AS MUCH TIMING AS IT IS AMPLITUDE. RECENT HPC VERIFICATION OF THE PREVAILING LONGWAVE PATTERN SHOWS THAT ALL MODELS HAVE UNDER-AMPLIFIED THE TROUGH IN THE EAST. THEREFORE...GIVEN PERSISTENCE OF THE CURRENT LONG WAVE PATTERN THROUGH F84...A SLIGHTLY GREATER AMPLIFICATION OF THETROUGH THAN DEPICTED BY CURRENT GUIDANCE IS RECOMMENDED.

Not sure how that affects things but was wondering what others thought.

The models have certainly shown a trend for a deeper and slower trough, probably correcting for their earlier bias. I wonder if that could still result in Irene getting capture at some point, and at least producing large areas of heavy rain to the NW. I would like to see further commentary from the mets on this one.

[Lookout]

And the East trend continues..just amazing!

Folks, posts like this aren't useful. Please refrain from short sentence posts, one word posts, or posting opinions not based on any facts.

[Baroclinic Zone]

If that pans out, the Euro's 4DVAR assimilation technique will take a hit.

And the GGEM would be the winner.

It's been consistently E of all the other globals.

[mattmfm]

CTP had this in their short term AFD:

SFC HIGH PRES WILL SLIDE OFF THE MID ATLC COAST BY WED MORNING, GIVING WAY TO RETURN FLOW/WAA PATTERN. A SHORTWAVE TROUGH AND ASSOCIATED COLD FRONT WILL MOVE FROM THE UPPER MIDWEST THROUGH THEGREAT LAKES INTO THE NORTHEAST STATES THROUGH THE PERIOD. 23/00Z MODELS SHOW RELATIVELY SIMILAR TIMING...WITH THE NAM ON THE SLOWSIDE OF THE DETERMINISTIC CONSENSUS AND THE GFS SLIGHTLY FASTER. THE ECMWF HAS TRENDED SLIGHTLY FASTER THAN ITS PREVIOUS 12Z RUNAND TOWARDS THE GFS. THE ISSUE IS NOT AS MUCH TIMING AS IT IS AMPLITUDE. RECENT HPC VERIFICATION OF THE PREVAILING LONGWAVE PATTERN SHOWS THAT ALL MODELS HAVE UNDER-AMPLIFIED THE TROUGH IN THE EAST. THEREFORE...GIVEN PERSISTENCE OF THE CURRENT LONG WAVE PATTERN THROUGH F84...A SLIGHTLY GREATER AMPLIFICATION OF THETROUGH THAN DEPICTED BY CURRENT GUIDANCE IS RECOMMENDED.

Not sure how that affects things but was wondering what others thought.

Depending on how amplified the trough is could (in the medium term) help lift Irene a bit northward and probably quell any possibility for it to slow/stall before reaching NC.

[LocoAko]

Another question for the group: between the Euro and the GFS, which model has done a better job with Irene since it was named? Has one performed much better than the other up to now?

Bump (Sorry if this was answered already and I missed it). I think this would be a good comparison to do, although I'm not sure it is being done and would be a pain to do manually.

[NCSNOW]

Bump (Sorry if this was answered already and I missed it). I think this would be a good comparison to do, although I'm not sure it is being done and would be a pain to do manually.

Regardless if this brushes OBX or hits SE NC. Got to hand it to the GFS. It's been consistently barking mostly for aGA/SC hit from as far back as I can remeber. GAWX has the stats on this. Evidently these CV long trackers seem to be more in the GFS wheelhouse as oppossed to homebrews.

[ArtRosen]

Pretty good consensus coming in for Irene.

[RU848789]

Just curious if anyone has ever verified that the "common wisdom" that tropical systems approaching the US east coast (from Florida to NC and even up into New England) from the general area of the Caribbean, generally tend to end up going east of the originally forecasted track. Have we had enough storms to make such an observation and if so, is there data on the presumed "bias" (i.e., X miles east of a day 5 forecast, on average)? It's interesting to watch the "cartoon" of the forecasted track of Irene over time and see the eastward shift in the predicted position.

http://www.nhc.noaa.gov/archive/2011/graphics/al09/loop_5W.shtml

The Saturday 7 pm initial advisory had Irene being just about at the middle FL Keys at 120 hours, while today's 8 am forecast has Irene making landfall around Wilmington, NC in about 112 hours or so. While that seems like a gigantic shift, it's magnified by the SW to NE curvature of the east coast in that region, in which eastward movement in the track results in much further eastward landfall along the Carolinas, since the system has more latitude to "bend" NE. Presumably the better comparison is that the current track has the storm bypassing the middle Keys latitude around the NE Bahamas, about 225 miles east of where it was originally forecast to go, about 60 hours ago and if I recall correctly that's within the 250 mile "error cone" for a 5 day forecast (assuming it verifies). At this point, the system is reminding me of Earl from last year, which looked like it was going to be a direct hit on NC even 2-3 days out, but ended up missing Cape Hatteras by 85 miles.

[am19psu]

Just curious if anyone has ever verified that the "common wisdom" that tropical systems approaching the US east coast (from Florida to NC and even up into New England) from the general area of the Caribbean, generally tend to end up going east of the originally forecasted track. Have we had enough storms to make such an observation and if so, is there data on the presumed "bias" (i.e., X miles east of a day 5 forecast, on average)? It's interesting to watch the "cartoon" of the forecasted track of Irene over time and see the eastward shift in the predicted position.

http://www.nhc.noaa.gov/archive/2011/graphics/al09/loop_5W.shtml

The Saturday 7 pm initial advisory had Irene being just about at the middle FL Keys at 120 hours, while today's 8 am forecast has Irene making landfall around Wilmington, NC in about 112 hours or so. While that seems like a gigantic shift, it's magnified by the SW to NE curvature of the east coast in that region, in which eastward movement in the track results in much further eastward landfall along the Carolinas, since the system has more latitude to "bend" NE. Presumably the better comparison is that the current track has the storm bypassing the middle Keys latitude around the NE Bahamas, about 225 miles east of where it was originally forecast to go, about 60 hours ago and if I recall correctly that's within the 250 mile "error cone" for a 5 day forecast (assuming it verifies). At this point, the system is reminding me of Earl from last year, which looked like it was going to be a direct hit on NC even 2-3 days out, but ended up missing Cape Hatteras by 85 miles.

I don't remember a single NHC forecast that called for landfall anywhere ever. Earl's forecasts trended west, not east.

Last year's verification report showed the models have right of track bias in recurving situations - in other words, the models were too quick to recurve systems in 2010.

[WxTurtle]

Bump (Sorry if this was answered already and I missed it). I think this would be a good comparison to do, although I'm not sure it is being done and would be a pain to do manually.

FTR (and ECMWF data isn't being processed for this storm-- probably because Brian doesn't support this as much given that he moved from MIT to NCAR last year) but you can get cyclone track/intensity errors and biases from his guidance page here. Down on the left under verification.

So far HWRF actually has the lowest track errors for the storm to date although the GFS hasn't been far behind it.

[GaWx]

Another question for the group: between the Euro and the GFS, which model has done a better job with Irene since it was named? Has one performed much better than the other up to now?

Josh,

Until we know the actual final track, a complete analysis is, of course, not doable. However, if we assume that it will actually hit near Hatteras or be a fish, I'd then give the Euro the edge, although not a huge edge. I'll just look at the 0Z and 12Z GFS runs since the Euro is run only for those two times. Irene was named at 23Z on 8/20. So, I analyzed the first U.S. hit point for five runs going from 8/21 0Z through 8/23 0Z:

GFS: W. Palm, Beaufort, SAV, Hatt., Hatt.

Euro: GA/FL border, CHS, Hatt., NC/SC border, ILM

They both have generally trended eastward. However, the GFS was way down at WPB when the Euro was at the GA/FL border. Of course, we know that the angle of FL vs. the track probably makes it seem worse than reality. Regardless, the starting edge definitely would go to the Euro. Actually, I'd give the edge to the Euro for the first three as a whole since it was further east for all three. However, the Euro did come back a bit westward for the last two runs whereas the GFS went to Hatt. So, if Hatt. or a fish verifies, the GFS would edge the Euro for the last two.

So, Euro 3, GFS 2...Euro edges out a win if Irene hits near Hatteras or fishes for these five runs, alone.

Edit: It just occurred to me that you may be just talking about the actual track of Irene up til now. If so, I realize this doesn't answer your Q.

[yoda]

12z GFS at 36 has Irene in the SE Bahamas. Intialization point looks good