*Jump down to the main body for the updates. The intro is unchanged.
With a perpetual eastward shift to the tracks on Monday and Tuesday, I never thought I'd be sitting here talking about Irene possibly tracking inland up the Mid-Atlantic coastal plain. The latest model runs (of course they may not be correct) have even surprised me. 36 hours ago we were hoping Irene could slip out to sea, as a few models had started projecting that to be the case.
Now, instead, we've got the models so far west that they're threatening Wilmington, NC again and more than one tracking so far west that Irene's weakening storm center passes through eastern Pennsylvania (never mind even through New Jersey). Of course, these model runs now may not be right either. Let's step below and examine this...
*UPDATE #2:
The entire write-up below is unchanged. I point that out because, as of Thursday afternoon, the models have shifted around again as they try to get a handle on the details of the various system interacting with Irene. In fact, I've changed the diary title to merely indicate that Irene's track forecast is shifting around. I dropped the "west". Why?
Because this afternoon's spate of model runs have almost universally shifted back to the east somewhat. Only one western model (the GFDL) has stood its ground with an inland track (the inland track is bad for Raleigh, DC and to a lesser extent Philly). And one of the lone exception to trend west was the furthest east model in its previous run... the Canadian model had been the lone outlier out near Cape Cod with its landfall in its previous run. It has backed west to Long Island (and southern New England, of course). But the GFS nudged eastward, the HWRF jumped well eastward, and the Euro pushed a fair bit east (though remains inland from S NJ northward).
All of these acronyms may mean little to you if you don't follow this stuff closely. But their all just computer models run by different governments. Meteorologists rely on these and then TRY to use our knowledge and experience to comb through them and figure out what the correct/best solution is. It's not easy, and sometimes none of them are correct. For example, the almost all swung west last night and, now, today are pushing back east. Well, something there is wrong, and almost every model fell for it... and so did the meteorologists... and, in fact, I'm still not sure which solution is correct. But since the big westward shift appears to have been a one-off, I'm leaning a little more eastward than my thoughts in the main body below (please take note of that when you read down further).
I didn't want to make this a boring text-only update, but the multi-track plot maps have not been updated for these latest models yet. The important take-away is this:
The models appear to be converging on a solution for lanfdall over far eastern North Carolina for Irene. Then she exits from the Virginia Cape and runs up JUST offshore of the Delmarva. Thereafter, Irene either heads north into far eastern New Jersey or northnortheast to a landfall probably no further east than just east of central Long Island and then probably no further east than New London, CT. Only the GFDL is west of this range of solutions, and no model is the east right now. Can this still change? Well, I guess the answer's obvious given how the forecast has jumped around the past 24 hours. Yes, it can still change. BUT, the changes we've seen this afternoon largely reduce the range of options. So, confidence is increased a bit. And as I said, my thinking that you'll read below has changed slightly. With the changes this afternoon - especially on the Euro model - I think perhaps that Irene may nose just a little east of where I was thinking. This could be bad news, as it may allow her center to stay off the NJ coast and, therefore, impact NY/New England as a stronger storm.
Now, as I get ready to post this the 18Z (late afternoon) NAM and GFS confirm the slight eastward trends.
New recon data just in, unfortunately, confirms that Irene is indeed a Category 3 hurricane (NHC, in their 5PM update suggested they may drop her down if recon didn't find higher winds soon... unfortunately, they did... clocking a 110kt flight level wind and a slightly lower cetral pressure).
Note: In their 5PM update NHC did nudge the track west, now putting it over NJ, which would weaken Irene more and reduce impacts - slightly - for NY and New England. I try not to be critical of NHC - they've got a thankless job and generally do superb work. But they've been about one step behind on this track "game". They were always left of the consensus while shifting the track up the coast. By this morning, the consensus - if you account for the quality of the models - was probably along the NJ shore or just inland; they kept Irene just offshore. Since then, there's been almost a universal eastward shift and NHC has pushed her the other way. With one major caveat, expect NHC to kick their track back east - just a hair - at 5AM (after the overnight model runs ...that one major caveat is assuming the overnight runs don't jump back west again).
Anyway, I screwed up trying to publish this at 3PM. So, I've now added these last few paragraphs and I've got a regular tome of an update here!! So, I'll publish now. And any further updates tonight will just come in the comment section below. But, to underscore the highlights of this update. 1) Tracks shift back east, generally, and focus on Morehead City to just off the Delmarva and Jersey coasts then into western or central Long Island and New England. 2) There is some "fluff" in that described track; the GFDL and Euro remain to the left, some others are slightly east. 3) Recon has just reconfirmed Irene as a Category 3 hurricane... and might even be showing hints of slight strengthening (pressure down to 948mb).
*UPDATE:
Just a brief update ...no new pics or anything... just want to keep you up to speed with the latest. The 11AM update from NHC pushes Irene back to the west just slightly, bring her right up along the Delmarva and NJ coast into NYC. Trying to judge how much weakening would occur with such an onshore track is a difficult prospect. So, I'm not even going to try... the forecast track will likely be moved off the coastline (in or out?) as things become more clear. Also of note, the first model run for the mid-day cycle of models (the NAM) actually trended east. This isn't a great model and shouldn't be given much weight, but it's the first piece of data we have from this new cycle of model runs (it keeps Irene off the Delmarva/NJ coast, landfalling her over the eastern half of Long Island late Sunday).
*NOTE: As with yesterday, in the interest of time, I'm posting this without proof-reading it. So, I apologize for any finger-fumbling typos. And if you see anything glaring that I typed wrong, that leaves an incorrect impression, let me know so I can correct it.
First, I want to point out the extremely problematic issue associated with the track forecast. Irene will likely be tracking approximately parallel to the Delmarva and New Jersey coasts. So, the precise point at which Irene makes her turn to the northnortheast will have a tremendous impact on conditions for the coastal Mid-Atlantic, eastern New York and New England. If she makes the turn west of approximately Morehead City, NC Irene will probably never come back offshore as she heads up through the Mid-Atlantic and Northeast. As a result, she'll be degenerating the whole time as she tracks NNE. It'll take her about a day to get up to NY/New England from that landfall. That means that NY/New England would see little impact. They'll be on the dry, east side. And though they're on the windy side, Irene would have weakened dramatically. The south facing coastlines might see some stiff winds, but nothing damaging and most inland areas would only see breezy conditions. Of course, I don't mean to ignore the folks down south... this would be worse for eastern NC, with more of the state on the dangerous east side of the storm. It could also be worse for southeastern VA as they're on the east side too... but they're kind of damned either way; their coastline faces northeastward - so a storm to their east drives the water in on them anyway.
But if you go the other way and keep the storm just off the Mid-Atlantic coast (after a likely NC landfall), this allows Irene to stay over water all the way up to a NYC/Long Island landfall. That would create dramatically worse conditions for New York and New England (and NJ, depending on how tight to the coast the track is). Not only would she stay over water, but the water temps are near their peak this time of year. Check out the sea surface temperatures:
...waters need to be about 80F to support a hurricane. That begins on this map with about the yellow contour. So, Irene might not even start weakening (aside from some North Carolina landfall induced weakening) until she reaches the latitude of southern New Jersey. At that point it's kind of too little too late... sure, she'd weaken some before a NY landfall, but it would be minimal weakening. And she's got a large wind field. As a result, winds out in southeastern New England, though most certainly weaker than what was anticipated yesterday would still be very strong. For example, look at the map below. This is the morning run of the American model (called the GFS). This represents the furthest west the track could go WITHOUT riding inland up the coast of New Jersey:
...that wind field is so expansive to the east that this model has 900mb winds of 75-80 knots hitting the coast of southeastern New England. Surface winds would not be that strong (900mb - milibars - is a pressure level some height above the ground). The conversion to surface winds is a reduction of about 20%. That means winds of 60-65kts. So, though they may be confined to the immediate coastline (friction will further slow things down inland), hurricane force winds would still be plausible even in a track this far west.
So, that's the major question today. Does the track of Irene carry her back offshore near the Virginia Capes, or does she remain inland and slowly collapse as she heads north? Let me be blunt... I honestly don't know. I'm leaning towards an inland track right now, but with extremely low confidence. The NHC forecast (as of 5AM) goes with the offshore track. Here's their forecast:
Though I'm thinking further west - and inland - I really can't argue with their forecast. I think an inland track anywhere as far west as the NJ/PA border eastward to an offshore track with a central Long Island landfall is very reasonable. Unfortunately, even though that's not that wide of a range, as I mentioned above the difference in impacts is massive.
Ok, let's step back a moment and take a look at Irene right now. She's been fighting a battle all night long. Dry air on her southwest side has been occasionally ingested (a bit surprisingly, in my opinion, because it's a small patch of dry air and not THAT dry ...but her satellite appearance has clearly been one of fighting off some dry air). Here's how she looks now (on water vapor, so you can see the dry, dark colors to her southwest):
She's also been dealing with an eyewall replacement cycle. Nonetheless, she's essentially held her own. For a while there a dearth of recon reports supporting her Category 3 status. But, this morning there have been multiple reports from aircraft supporting an intensity of 100kts for Irene. So, unfortunately for the U.S., Irene appears to be putting up a good fight. And she could remove herself from the dry air as she begins to lift north, away from it. So, I think the NHC forecast showing some modest strengthening is reasonable.
That brings us to the problem at hand... when/where does Irene make her turn? Here's a plot of the latest model solutions from Brian Tang's web site:
Basically you've got a the GFS right down the middle along the Jersey coast after a Morehead City, NC hit and a re-emergence into the water off of Norfolk, VA. The NAM and NOGAPS are right there with it, but they're not exactly stellar performers. Then there's one solution close by on either side (HWRF just inland to the west, UKMET - labeled EGRI - just to the east). Then, further out on the limbs there are two more solutions. The GFDL and the GFDL/NOGAPS blend are out to the west, landfalling further down the NC coast, then slicing up, inland, through the DC metro area and into eastern PA and central NY. On the east side are the COTI (irrelevant) and the Canadian (CMCI).
What about other models not on this plot? The Euro model has a track between the GFDL and HWRF, inland along the NJ/PA border and into eastern NY. And of the statistical models (a suite of less complex models the National Hurricane Center runs) the only one of any real consequence (called the BAMD) is basically in agreement with the GFS (and NAM and NOGAPS).
For various reasons, I am leaning - with admittedly low confidence - to the western solutions. First, the shift was pretty decent and almost universal in the overnight runs. That tells me that the models "saw" something new (courtesy of the extra sampling being done by the NOAA and Air Force aircraft, perhaps???), and a continued westward trending wouldn't shock me. Furthermore, the models on the west side (Euro, GFDL, HWRF) are of better quality than those to the east (UKMET, COTI, CMC). In fact, the Euro has been probably the best performer. It was the first model to swing Irene up the coast while the others were still down in Florida, Georgia and South Carolina; but, then, it was the most stubborn holding her inland in the Mid-Atlantic while the other models swung all the way past the Euro to solution over eastern Long Island and SE New England.
So, on balance, the best reasoning, IMO, supports the western solutions. However, an argument can also be otherwise made. The GFS offers the most middle of the road consensus. While some model "jumps" are due to the models "seeing" something new, it can also occur when some bad data sneaks under the QC bar. And, frankly, it's impossible to tell which is the case without going back and doing an extremely detailed forensic analysis of the data to see what culprit caused almost every model to jump west overnight. I am merely guessing it was additional data because we've got both NOAA and the Air Force out there taking extra observations from in and around Irene.
What does that leave us with? Well, probably a track most similar to the HWRF or the Euro. Here are a couple of Euro maps from Ryan Maue's site:
...it has Irene making landfall near Morehead City, NC Saturday evening with maximum sustained winds estimated to be 100-105kts (using a 15% reduction from the Euro's max 850mb winds). The wind field is pretty expansive, with hurricane conditions up and down most of the North Carolina coastline (though, admittedly, the Euro may not handle wind distribution in a hurricane too well).
Then by Sunday night:
...the Euro has trudged northnortheastward, remaining inland, into southeastern PA. Winds have dropped dramatically - to about 60kts, and predominantly out over the open ocean east of New Jersey... though the NJ coast and Long Island would still see some windy conditions. Not seen on these maps, the Euro keeps the excessively heavy rain along and just west of its forecast track for Irene. So, there could be some flooding concerns on the eastern edge of the Appalachians on up into the Poconos.
The HWRF, being a bit further east, has a partially off-shore, partially on-shore track (Irene does come back offshore at the Virginia Capes, but landfalls again in southern NJ). It also has Irene landfalling just east of Morehead City. Here's a map of its track for Irene along with its predicted surface wind speeds:
This allows Irene to stay stronger, longer... bringing severe conditions to the Chesapeake Bay, Delmarva, NJ and - slightly less severe but also bad - in SE NY and coastal CT. Notice that even in this partially inland track, most of the south-facing coasts of eastern New England still sees winds in the 50-64kt range, sustained. Again, Irene's wind field is pretty large.
So, that's my thinking, somewhere in the range of those two options. The highest confidence is in the eastern NC landfall (Saturday night)... though even that's not yet a guarantee (GFDL is down more towards Wilmington). After that, it gets tricky. Even in my small range of solutions the impacts vary greatly for some. But I think coastal VA/MD/DE should see at least hurricane gusts, and coastal NJ close to it... at a minimum. Long Island and south coastal New England would also likely see some high winds, but probably nothing damaging. Excessive rains will fall in eastern NC/VA/MD/PA, DE, NJ and central/eastern NY.
Keep in mind, these impacts are based on an EXTREMELY sensitive track forecast. The 12Z model runs will be critical in determining whether or not the overnight runs represent a trend, an error, or just a proper correction. Those models runs are just starting to come in now (NAM model running as I write this). So, check back in on this diary later as I'll update it - either here in the diary or posting comments below). I also have more info I want to add to this, including links to innundation maps, perhaps... but I'd like a better idea on this track first.