Sandy… Please Start Your Jog East!

When you see an exceptional weather model run or two on a day 7 or 8 forecast you expect it to go away. Every once in a while we see a bit of long range weather porn pop up on the computer models only to quickly disappear.

12z GFS Ensemble Forecast / Courtesy: Weather Underground

Sandy is a bit of a pain, however. She’s not behaving like most tropical storms or hurricanes do. The “spaghetti plot” of this afternoon’s GFS ensembles really tells the story with 2 distinct possible paths for Sandy.

This is one of the oddest spaghetti plot I’ve ever seen for a storm threatening the northeastern U.S.! While half of the GFS ensemble members curve the storm east (including the operational GFS in white which is in the out to sea track)  the other half of the models take Sandy to just east of Cape Hatteras and sling shot it west into New England. That would be one of the most unusual paths ever documented for a New England tropical storm or hurricane!

The operation European model agrees with the left hook bringing Sandy ashore in southern New England as a hurricane from the southwest. Exceptionally unusual.

12z Euro QPF/MSLP / Courtesy: WSI

There is no doubt in my mind that the European model is overdoing the strength of this storm. It has a tendency to strengthen storms like this that are undergoing extratropical transition a bit too much. Needless to say a track like this is concerning.

So what should we make of it all? At this point we are still at least 36 or 48 hours from really getting a good handle on Sandy. The reason for all these funky looking computer model solutions is that the weather pattern is very amplified and all blocked up!

Euro Ensemble Mean 500mb Heights (156 hour forecast) / Courtesy: WSI

The Euro Ensemble mean shows a monstrous ridge over Newfoundland and Greenland which may act as a block for Sandy and prevent it from moving out to sea. The positive height anomalies near Newfoundland are nearly 4 standard deviations above normal! The large rex block that develops in the next 72 hours will transition to a large omega block which is pictured above (the height contours make on outline of the Greek letter omega). How this block evolves will determine whether Sandy can escape to the east like the GFS shows or get shoved up the coast like the Euro shows.

Hurricane Esther Best Track / Courtesy: Unisys

While we have seen truly anomalous New England hurricane tracks in the past – they are few and far between. Here’s a look at Esther from 1961. What the heck was Esther doing out there???

The category 4 hurricane approached New England from the south and then the storm did a loop south of Cape Cod. The storm made 2 landfalls in New England.

Other storms have done odd things. Most notably, the Great New England Hurricane of 1938 which was in the process of being absorbed by a large cut-off low that was diving south of the Great Lakes.

1938 Hurricane Best Track / Courtesy: Unisys

This hurricane tracked due north or maybe even a hair west of due north and then hooked left south of Montreal toward Toronto.

The fact that we need to look back to 1961 or 1938 to find hurricanes that impacted New England that took really unusual paths says something about just how hard it is to get a storm to get absorbed at just the right time by the jet stream and hook left into the region and not head out to sea or continue due north.

Even if Sandy takes a path that will indirectly or directly impact us it’s not clear how strong the storm will be. While there is a very outside chance of a direct hurricane hit the more likely scenario is a brush or a hit with a hybrid system that will act like a fall-nor’easter that is spinning around in the north Atlantic.

No need to panic! Let’s watch things and see how they shake out over the next day or two. My gut feeling is that the computer models will trend away from the extreme solutions (they almost always do) and we’re left with something in the middle.


Fujiwhara and Northeast Hurricanes

On Saturday I attended the Tri-State Weather Conference out at WestConn in Danbury. Dr. Bob Hart, a North Branford native who’s a professor at Florida State University, gave a superb talk on the 1938 hurricane.

We know a lot about the 1938 hurricane. It’s been looked at and researched quite a bit. We are now able to get a new look at the storm thanks to the 20th Century Reanalysis which uses historical surface observations to recreate a  “six-hourly, four-dimensional global atmospheric dataset spanning 1871 to 2010 to place current atmospheric circulation patterns into a historical perspective.”

Now we’re able to get a better idea what the synoptic pattern looked like during at after the 1938 storm. Here’s a look at the sea level pressure and 500 mb ensemble mean height reanalysis for12z 9/21/1938.

No big surprises here with a deep, anomalous 500mb cut off low essentially capturing the 1938 hurricane and rocketing it north. A large portion of Dr. Hart’s presentation was about 38’s unusual (and dangerous) extra tropical transition to warm seclusion and associated problems with predictability.

He also talked about the 1893 hurricane which brought a devastating storm surge to New York City. The 1893 storm took an unusual path. Instead of losing longitude as it moved north of the Carolinas is continued due north to New York City.

1893 Hurricanes #4 and #6

While 1893 Hurricane #4 barrelled toward New York City there was another hurricane nipping on its heels. The best track data shows 1893 hurricanes #4 and #6 on very different paths. On August 24 at 00z hurricane #4 was just east of Norfolk, VA while hurricane #6 was just north of Puerto Rico as caetgory 2 and 3 storms, respectively. The 20th century reanalysis data shows an impressive parade of Atlantic hurricanes but a relatively unimpressive 500 mb trough over the Great Lakes (though there is strong ridging over the North Atlantic).


So what lead the 1893 hurricane to smash right into New York City and not head out to sea? Dr. Hart argues the Fujiwhara effect kept Hurricane #4 close to the coast as the interaction between the 2 powerful cyclones resulted in Fujiwhara behavior. Wild stuff, right? While Fujiwhara can result in beautiful interactions in the west Pacific it is more unusual (but certainly not unheard of)  in the Atlantic. While these 2 hurricanes didn’t make a complete loop Fujiwhara appears to have influenced the hurricanes track.

Connie and Diane in 1955 lead to one of the state’s worst natural disasters on record (see here).

While I had always noted their unusual track burps with Connie’s sharp right south of Hatteras and Diane’s sharp left southeast of Bermuda I had never really given it much thought. Hart cited this as an example of Fujiwhara interaction between the 2 cyclones.

While Fujiwhara is fun to say it can lead to a model meltdown. One of the most sobering parts of Dr. Hart’s talk was about the breakdown in predictability of a tropical cyclone’s track when Fujiwhara interaction is involved. Using the 20th century reanalysis as an initial condition the model Dr. Hart ran (I believe it was a WRF ensemble) showed exceptional spread and run-to-run variability in the 1893 case. The potential for dramatic track errors in a short period of time is alarming – particularly for the kind of storm that brought a 25 foot storm surge to New York City.

Getting a hurricane into New York or southern New England is unusual. While many follow a rather predictable pattern there are exceptions with highly anomalous flow regimes or Fujiwhara interactions that, according to Dr. Hart, can lead to breakdowns in predictability. Certainly a sobering thought from an emergency management and operational meteorology perspective.

73 Years Ago Today We Had “Our Katrina”

It’s been 73 years since a “major” hurricane has hit Connecticut. The September 21, 1938 hurricane was arguably New England’s greatest natural disaster of the 20th century and to this day remains the benchmark for modern New England hurricanes.

Since 1938 we’ve seen our fair share of hurricanes. Hurricane Carol produced category 2 conditions in the state. 1944, Donna, Gloria, and Bob produced category 1 conditions in portions of the state. None have risen to the level of ’38.

What made the ’38 hurricane so powerful is that as the storm was rocketing north (at nearly 60 mph) the jet stream was helping to enhance the storm’s power. While many storms weaken at this latitude the 1938 storm was holding steady. The tropical system was being fueled by extratropical processes which lead to a monster in our own backyard.

The hurricane made landfall in New Haven with a pressure of 946mb. The sustained wind in some parts of the Connecticut shoreline reached 115 m.p.h. There were higher gusts. Compare that to Gloria’s 75 m.p.h. sustained winds or Irene’s 50 m.p.h. sustained winds and you can see what made the ’38 storm so remarkable.

In the past week we’ve heard from utility company executives talk about what a extraordinary storm Irene was. I don’t think Irene was extraordinary at all. The duration of damaging wind gusts and the heavy rain made the damage worse than you’d expect from other 50 m.p.h. tropical storms but let’s be real here. The damage from Irene was not even a tiny fraction of what a major hurricane can do.

Irene’s 5 foot storm surge came at the worst possible time – during an astronomically high tide. How would we be prepared for a 15 foot storm surge like we had in ’38 (and during Carol, for that matter in southeast Connecticut)? Are we?

The 1938 hurricane was exceptional but it wasn’t unprecedented. There are some indications that the 1635 Great Colonial Hurricane was even stronger with a pressure around 938 mb!

When the utility companies testified about their storm response in front of the legislature earlier this week the president and COO of CL&P mentioned that Irene was about the size of Katrina when it made landfall in Connecticut. Yes, the official radius of gale force winds was similar, but we had about 6 hours of marginal tropical storm conditions in Connecticut from Irene. Winds never even approached hurricane force.

If the government, utilities, or citizens think that there is any appropriate comparison between Katrina and Irene they’re truly not prepared for a major hurricane. Hell, they’re not prepared for a category 1 hurricane! 1938 was our Katrina. Irene was barely a blip on the historical radar. I am truly concerned that making Irene out to be some extraordinary freak set of circumstances that crippled the power grid means we aren’t even close to prepared for the day the 1938 hurricane returns. It will.

Irene brought down a couple trees in many neighborhoods. 1938 flattened whole forests. The Guilford Green lost 2 trees during Irene. In ’38 it lost 80 percent of them. The trees still have their leaves this year but after ’38 most trees in the state were completely stripped bare.

If a tropical storm can knock out power for 9 days and destroys dozens of beachfront homes I can only imagine what a major hurricane like 1938 would do.