Hurricane Irene Model and Forecast Discussion

[Midlo Snow Maker]

landfall is west of 0z by a little hr 36

[phlwx]

eye generally goes along Jersey coastline or just offshore, final landfall in Central LI.

[Jefflaw77]

eye generally goes along Jersey coastline or just offshore, final landfall in Central LI.

hurricane gloria redux

[A-L-E-X]

hurricane gloria redux

At it's nearest approach to me, it's about 10 miles south of Long Beach, so those east winds will be vicious in the northern eyewall. According to Doug, it makes landfall in Freeport, Nassau County, 15-20 miles west of the last run.

[Jefflaw77]

this storm appears to be shaping up to be a repeat in strength and location to Hurricane Gloria... The only difference that I see is that this storm may be slightly larger, moving slower, and come in a bit further west than Gloria... Most importantly though are the tides... If this storm comes in at high tide, w/ the new moon- it will cause much more damage than Gloria did.. That was Gloria's saving grace - it came in at low tide.

It should be interesting how western LI fares, b/c from what i remember, most locations west of Gloria's landfall were spared a huge hit. and many were left wondering what the big deal was in the western Nassau/NYC region.. The question is how wil IRENE be different in that respect, w/ a relatively similar track..

At the end of the day any points east of the eye on Long Island are in serious trouble..

Jeff

[dtk]

The 00z ECMWF initialized on a global data package as close to 00z, 8PM EDT, as possible. 00z is the data cutoff for ALL 00z models. The fact that the model output does not arrive until 1:43 AM EDT and beyond does not change anything, and it gets absoutley no further information after 8 PM EDT. The high resolution Euro model takes that long for the ECMWF to run and produce the grib files that they transmit to their customers around the world.

This is not correct actually. Both the GFS and ECMWF data assimilation packages have observation windows that go beyond the initialization hour. For the GFS, the observation/data dump/prep jobs for all four cycles starts at 2h45m past each of the four cycle times (00z, 06z, 12z, 18z); see here:

Production Status Page Link

For the GFS, we have a 6 hour bin, which for 0Z means that we assimilate any observation that falls within +/- 3 hours (i.e. any observations that was TAKEN between 21z and 03z). The caveat here is that we also have to have received it by the 2:45h after hour start time, obviously (which, some observations take longer than that to arrive, and get assimilated in the late cycle, GDAS, which is run nearly 6 hours after each initialization time).

The ECMWF model initialization does actually start later than us (they wait slightly longer for observations to come in), so I'm sure they use observations up to at least 03z (for a 00z cycle) like we do.

[SnowGoose69]

this storm appears to be shaping up to be a repeat in strength and location to Hurricane Gloria... The only difference that I see is that this storm may be slightly larger, moving slower, and come in a bit further west than Gloria... Most importantly though are the tides... If this storm comes in at high tide, w/ the new moon- it will cause much more damage than Gloria did.. That was Gloria's saving grace - it came in at low tide.

It should be interesting how western LI fares, b/c from what i remember, most locations west of Gloria's landfall were spared a huge hit. and many were left wondering what the big deal was in the western Nassau/NYC region.. The question is how wil IRENE be different in that respect, w/ a relatively similar track..

At the end of the day any points east of the eye on Long Island are in serious trouble..

Jeff

The south shore of Nassau county got hit quite hard with water going about 1 mile inland in some spots. The north shore though as well northern Queens did not seem to get a bad surge....even given the low tide its somewhat surprising...below are the obs from LGA that AM and you can see that they sure had their usual strong ENE funneling wind.

SPECI KLGA 271223Z 10030G42KT 11/2SM RA FG SCT007 OVC012 A2943 RMK RB1155 PRESFR SLPNO

METAR KLGA 271300Z 09032G44KT 1SM +RA FG SCT007 OVC011 23/22 A2936 RMK R04VR35V50 RB1155 PRESFR PK WND 0944/43 SLP942 P0036 T02330222

METAR KLGA 271400Z 09035G47KT 3/4SM +RA FG SCT006 OVC009 23/22 A2921 RMK R04VR30V35 PRESFR PK WND 0948/17 SLP892 P0028 T02280222

SPECI KLGA 271425Z 09032G50KT 1/2SM +RA FG OVC007 A2909 RMK R04VR18V24 PRESFR SLPNO

METAR KLGA 271500Z 10034G55KT 1/2SM +RA FG OVC007 23/22 A2899 RMK R04VR16V20 PRESFR 99173 0NE PK WND 1055/42 SLP817 P0046 T02280217

SPECI KLGA 271525Z 08032G42KT 1/4SM +RA FG OVC004 A2883 RMK R04VR12V14 PRESFR SLPNO

[meteorjosh]

It appears some strengthening may be under way and in fact Irene's eye should wink open in the next hour or two...this makes sense given the warmer Gulf Stream waters she is just now beginning to encounter, around 29C. Next 18 hours are going to be very interesting in this regard. I'm basing this on satellite presentation.

[LakeEffectKing]

GFDL 06z still maintains a just west of Chesapeake Bay to due north into CNY track.....

[lsukev]

There was a time when the GFDL was the NHC's baby. I'm not sure what happened.

[nj2va]

The GFDL solution, while a worse impact down here, would lighten the blow to NJ coast, NYC, and LI for sure. I don't buy it though.

[KEITH L.I]

There was a time when the GFDL was the NHC's baby. I'm not sure what happened.

well it was predicting a Florida landfall 3 days ago..and it's the only model showing this solution..so yea..it's out there

[5speed6]

The GFDL solution, while a worse impact down here, would lighten the blow to NJ coast, NYC, and LI for sure. I don't buy it though.

It would be game over for everybody on the western shore of the Chesapeake--Annapolis, Baltimore, even D.C. By contrast, a repeat of Isabel would be a welcome alternative.

[SnowGoose69]

The GFDL solution, while a worse impact down here, would lighten the blow to NJ coast, NYC, and LI for sure. I don't buy it though.

Its hard to believe the other models are that off at this point. It only makes me worry the EC and UKIE could be onto something with their insane tracks but I'm still leaning towards the GFS and most of the hurricane models or about 30-50 miles east of that.

[LMolineuxLM1]

Its hard to believe the other models are that off at this point. It only makes me worry the EC and UKIE could be onto something with their insane tracks but I'm still leaning towards the GFS and most of the hurricane models or about 30-50 miles east of that.

So the Euro and Ukmet are 30 to 50 miles west of the rest?

[Kicking Up A Storm]

Its hard to believe the other models are that off at this point. It only makes me worry the EC and UKIE could be onto something with their insane tracks but I'm still leaning towards the GFS and most of the hurricane models or about 30-50 miles east of that.

Still trying to learn: I know the GFDL was way off at the very beginning of this storm, but it's consistently shown this current solution for days now, while all the other models keep waffling. Surely the consistency should mean something? (Being located within the DC area is bringing out my inner weenie and support for the GFDL, but still a serious observation/question).

[TL97]

So the Euro and Ukmet are 30 to 50 miles west of the rest?

The Ukmet has a track over Delmarva, while the Euro is between that and the GFS

[SomeguyfromTakomaPark]

Still trying to learn: I know the GFDL was way off at the very beginning of this storm, but it's consistently shown this current solution for days now, while all the other models keep waffling. Surely the consistency should mean something? (Being located within the DC area is bringing out my inner weenie and support for the GFDL, but still a serious observation/question).

Just a thought from a couple of years of tracking. Because a model is consistent doesn't mean its right, it may just be consistently seeing things wrong. I've seen that countless times in in winter. When it has little or no support its hard to take it seriously. Hug at your own risk.

[tombo82685]

I thought someone just posted the HWRF in the Mid-Atl thread with a solution west?

the hwrf was a good bit east at 0z, the 6z run came a good bit west and matches the gfs nicely

[eurojosh]

Still trying to learn: I know the GFDL was way off at the very beginning of this storm, but it's consistently shown this current solution for days now, while all the other models keep waffling. Surely the consistency should mean something? (Being located within the DC area is bringing out my inner weenie and support for the GFDL, but still a serious observation/question).

With all the various factors that have/continue to influence Irene (two northern ridges, ERC, dry air over Hispaniola etc), the only thing consistency suggests is that the model isn't sophisticated enough (compared to Euro or, for that matter, GFS) to factor in all the details. Run-to-run consistency is nice to have, but in something as complex as predicting the track of a hurricane that interacts with the Atlantic coast, perhaps it's also suggestive of insufficient granularity.

[LakeEffectKing]

The experimental HFIP HWRF, has a similar scenario as the GFDL:

[winterymix]

The experimental HFIP HWRF, has a similar scenario as the GFDL:

You have a heck of a good reputation....what does your "gut" tell you?

10% chance

15%

20%...

that Irene takes the CHes. Bay route?

[Jmister]

Interesting, I didn't know there was an experimental HWRF. With all these hurricane models, it seems they are grossly under-forecasting the wind over land. I assume this is because they were programmed with physics for wind predictions over open water? The non-hurricane models have much higher wind forecast than this.

[mike2010]

All i've looked at with the NAM is it's initialization so far. Is it just me,or does it appear it initialized way east of where it should of ?

[LakeEffectKing]

You have a heck of a good reputation....what does your "gut" tell you?

10% chance

15%

20%...

that Irene takes the CHes. Bay route?

Wow...trying to nail me down to within 5%!!

I'd go with 10%....sounds reasonable, ATT, IMO.

[mitchnick]

All i've looked at with the NAM is it's initialization so far. Is it just me,or does it appear it initialized way east of where it should of ?

remember, NAM initialized 2 hours ago

is that sat pic 2 hours old or current?

EDIT: this water vapor loop makes Irene seem further east than your visible pic imho (both are current)

http://www.ssd.noaa.gov/PS/PCPN/DATA/RT/na-wv-loop.html

[mike2010]

remember, NAM initialized 2 hours ago

is that sat pic 2 hours old or current?

About an hour ago, but still. Compare it to the eastern tip of NC. Seems way off.

[Derecho!]

All i've looked at with the NAM is it's initialization so far. Is it just me,or does it appear it initialized way east of where it should of ?

Note that the NAM map projection on the left has due north not pointing straight up, while the satellite image on the right does; if you take that into account, the init looks fine to me.

[BethesdaWX]

All i've looked at with the NAM is it's initialization so far. Is it just me,or does it appear it initialized way east of where it should of ?

The NAM map is deceptive in that you have the "globe" shape/curve there while the satellite image does not have that.

Also, would you mind telling me where you're getting the NAM from? Allan's site isn't running it attm.

[LVblizzard]

All i've looked at with the NAM is it's initialization so far. Is it just me,or does it appear it initialized way east of where it should of ?

It looks like it initialized just fine. The NCEP map takes the curvature of the earth into effect on their maps so you have to tilt your head a little on that NAM map to imagine the lat/lon lines.