My title is a shade misleading, identifying only North Carolina, Massachusetts and Nova Scotia as being under threat from Hurricane Earl. Beach erosion and possible coastal flooding will impact all northeast states and much of the SE facing Canadian Maritimes coastlines. Depending on Earl's precise track, other states/provinces could also see some significant weather from Earl (like southeast VA, eastern Long Island in NY, RI, ME, Newfoundland and others).
On the flip side, by "threat" I mean exactly that... Earl is still just a "threat" for most of these areas. The only one of these three locations where an actual landfall is predicted is Nova Scotia, and Earl will be weaker there than as he passes either of the other two locations (NC or MA). And if he can manage to pass just far enough offshore, even the impacts at NC and MA could be fairly minor... just some showers and gusty winds at Cape Hatteras and Cape Cod.
Anyway, on to the latest with Earl...
I had expected the overnight computer models to lock this thing up for us. A critical piece of energy associated with a trough (a "trough" is just an upper level dip in the jet stream... it drives cold fronts and would act to reach down and "scoop up" Earl) came into the North American West Coast last night. Previously, it was over the Pacific, where lack of weather observations might've made it difficult for the computer models to properly discern this feature. I figured that once it was observed by weather balloons in North America, the models would lock onto the trough and converge on a solution for Earl. Wrong. In the overnight weather model runs, some of them trended east, some west. The biggest shifts were to the west, but the usually-more-superior models were the ones that nudged east. So, it's tough to figure out which ones are correct.
In the models' defense, they're all in reasonable agreement. We're talking about mostly modest differences... it's just that these small differences yield massive impact differences. Also in their defense, the differences are well and easily explained. It's all about speed. The models moving Earl faster (the Navy model, the UK's model and the Navy blend with the hurricane specific GFDL model) naturally take him further west. When you think about it, it's obvious why. The aforementioned trough is marching across the continent from west to east and it is responsible for turning Earl north and northeast. Until its influences reach Earl, he'll continue moving northwest. The models are in good agreement about the progression of the trough. But if Earl moves faster, he will get further northwest before encountering the trough; if he moves more slowly, he will be further southeast (the "east" part of that being critical) when he turns north and northeast. So, the point is, the model differences are neither huge nor obscurely difficult to understand. What is tricky is deciphering which ones are right.
At any rate, here's how Earl looks now (as of about 8AM, when I started writing this):
Earlier this morning the eye had become obscured and Earl looked a bit "ill". He's looking considerably stronger in the image above, though still not as impressive as at his peak. And, in fact, the National Hurricane Center has nudged him down a tick... to a 110kt, Category 3 hurricane, and they even acknowledged that he may be a bit weaker than that. New aircraft recon data does little to clarify things (pressure still low... 943mb, but winds comparatively weak); so, I think the NHC handled this well. Earl may have actually been a bit weaker and may still be, as they said. But given his improved appearance and continued very low pressure, he may rebound some.
Of course, the big question is, where is he headed? Here's a look at the various model solutions:
This plot comes from Allan Huffman's excellent site. These models present a bit of confusion. On the one hand, the greatest cluster is clearly offshore... and that cluserting is far enough offshore such that only coastal Nova Scotia has any chance at hurricane force winds. However, that's a bit misleading. The "clustering" is largely the result of plots from models which run off the same baseline and, therefore, not surprisingly, yield similar results. The "MRFO", "AVNO" and "AEMN" are all based on the American global model. On top of that, the "GFDL", "GFDT" and "HWRF" use that same model's background field to initialize their conditions. Point is, these are not independent solutions in this cluster. They are clusted, in part, simply because they are dependent solutions. That, intrinsically, means we need to give ample consideration to those more western solutions. But, on the other hand, those western models have been some of the poorest performing models of late. The UK's model actually, in general, has a decent track record with tropical systems; but it's been all over the map, literally, with Earl.
As noted above, these differences are all due to the speed with which these models move Earl. The further west ones are all faster moving. Compare the Navy model (the NOGAPS, labeled "NGPS" on the map above) to the American model. By Thursday night, here's where the NOGAPS has Earl:
...at that time it has Earl just south or southwest of Cape Hatteras, NC, about to make landfall on the Outer Banks. On the contrary, here's the American model:
...it is well to the southsoutheast of the Navy's model. And, at that time, Earl is turning to the north, but had been moving northnorthwest. So, that is a difference right along track... entirely a forward speed issue. In other words, this is both a very simple and a very complex forecast. That is, all we really need to determine is the forward speed of Earl... everything else is pretty much lined up. But that's complex because it's pretty obscure to try to figure it out. Those maps about show about a 100 mile or so difference in the model, but that's a two day (48hr) forecast. So, do the math... we're only talking about a 2 to 3mph difference in forward motion. Honestly, it's virtually impossible to nail that down, even at this short-term distance.
But my best call is pretty much the same as it's been and right in line with NHC. I'm looking at current satellite and recon aircraft data. At this time Earl is very near 24.6N, 71.7W. It is difficult to discern maps down to that precision, but the two hurricane-specific models (HWRF and GFDL) produce a text file of actual positions. Their overnight model runs placed Earl at 24.4N and 71.5W for the GFDL and 24.6N, 71.6W for the HWRF. So, this seems to be a bit faster than the GFDL and roughly on the HWRF. Not surprisingly, the HWRF track ended up a little west of the GFDL. (Note that a brand new HWRF track did nudge to the east a bit, but I'm not using that for this analysis since these mid-morning runs are historically less accurate, since less new data was ingested into them.) So, my thinking is pretty much along the lines of the HWRF, which you can see above. That's to the left of the "cluster" but right of the western solutions from the Navy and UK models. It also puts me right on or just a bit west of the NHC forecast:
What does that mean for impacts? First, remember, for impacts, I'm assuming my forecast is correct. If Earl moves faster, and further west, conditions would be considerably worse for NC and southeastern New England. If he moves slower, impacts would be much less for those areas. As for Nova Scotia, impacts would be about the same, but just focused more west or east in the province. So, keeping that huge assumption in mind, the most likely impacts are:
For North Carolina, winds could briefly sustain to hurricane force on the Outer Banks, but winds will primarily be in the tropical storm force range, with several hurricane force gusts. There isn't much "deadwood" in that area, given their frequency of strikes. So, there will likely be some power outages, but not major tree or related damages. However, overwash of the main road on the Outer Banks occurs easily, and would likely occur if this track verifies. So, travel on the Outer Banks could easily become restricted. Elsewhere in eastern NC and southeastern VA conditions will be nasty with heavy rain and winds, but not particularly damaging. Still, folks should hunker down as power outages and a few downed trees and power lines could produce a few hazards. Biggest impacts in this area will occur Thursday afternoon through Friday morning.
In southeastern New England Earl will likely pass a bit further away than it passed NC. Plus, it will be weakening significantly over cooler water (though the wind field may be expanding some). So, I'd anticipate no one experiencing hurricane force winds and, perhaps, just a gust or two to those levels on the Cape and Islands. Tropical storm force conditions will dominate. But there is more "deadwood" here than in NC. So, there will likely be some downed trees (not too many, as winds shouldn't be very severe on this forecast track). That means power outages and potential tree-related damage (trees or large limbs on cars, houses, etc). So, I anticipate no significant, major damage in the area... just isolated damage. But power outages could make the storm a notable inconvenience and folks should be prepared to lose power. Most impact here will occur late Friday.
Nova Scotia will bear the brunt, with an actual landfall likely. Of course, Earl will probably be down to a Category 1 by then or, at worst, a low-end Category 2. Nonetheless, this will come into an area of western Nova Scotia that is slightly less used to these storms. Recall Juan, which was quite impressive, came in further east, leaving western Nova Scotia on his weak side. On westernmost parts of the province - again assuming the track verifies!! - Earl will probably be worse than Juan. It still won't be a major hurricane; so, significant structural damage to homes is extremely unlikely. But tree damage could be considerable and, therefore, so could the collateral damage and power outages. And I should also note that Maine-iacs will also see some heavy rain and gusty winds, with potential power outages, from about Bar Harbor up to Eastport.
I will, of course, try to respond to comments below, and I will try to post updates when the afternoon models come out. Be safe!