Three years later: What made Hurricane Sandy so remarkable

Good evening.

I’ve been so busy lately with school and other events (including a visit to the National Weather Association annual meeting last week in Oklahoma) and the weather in the Northeast has been fairly quiet lately. There was definitely some rain and wind in some areas this week, due to a low pressure system far to the west which absorbed some moisture from the remnants of Hurricane Patricia in the Pacific, but it is nothing like it was three years ago tonight.

Remembering Hurricane Sandy

We all remember what happened on October 29, 2012, since most of our readers dealt with it personally and had their own experience. Hurricane Sandy had fairly modest beginnings in the tropical Atlantic in mid-October 2012, unable to develop into even a tropical depression until October 22 in the western Caribbean due to hostile conditions. The western Caribbean is a traditional breeding ground for late season hurricane activity, so the development of Sandy was nothing unusual. It strengthened into a tropical storm shortly afterward, and into a hurricane on the morning of October 24 south of Jamaica. The rest of the story is well-established and many reports (including the NHC’s official report) tells a great story on what happened meteorologically. Another blog post tells a good story as well.

So what made Sandy so remarkable?

Time of year

Late October isn’t exactly prime time for Northeast hurricanes. Most hurricanes in the database hitting the Northeast occurred from early August to early October. Sea surface temperatures offshore (especially near the shore away from the Gulf Stream) drop rapidly starting in early to mid-October. In the previous 100 years, no hurricane made landfall north of Cape Hatteras after September 28, with most October (and November) hurricanes favoring Florida and the eastern Gulf of Mexico coast. We could certainly be forgiven if we believed that hurricane season was over for us – especially after a fairly quiet 2012 season up to that point (at least for the east coast) – although we know better now.

Unusual track

It is typical for storms in late October to head northward up to roughly 25N latitude, then turn northeast out to sea. However, the atmosphere was charged up in the east coast region, with a blocking ridge to the north and north-northeast stretching from Greenland and Newfoundland southward (which is not typical). Combined with a deep non-tropical low over the central Atlantic (which partially absorbed Tropical Storm Tony), the path eastward out to sea was blocked. Sandy was then forced northward over the weekend, in the western Atlantic between Bermuda and the Southeast coast.

Notice the upper-level pattern. With the blocking ridge stalled due to a nearby strong low, Sandy had nowhere to go but into the jet stream.

Notice the upper-level pattern. With the blocking ridge stalled due to a nearby strong low, Sandy had nowhere to go but into the jet stream.

Unusual structure

There was very high wind shear in the vicinity of the Bahamas on October 25 and 26 due to the presence of a weak upper-level trough. Normally, a storm that enters such an atmosphere would disintegrate very quickly, as we saw with storms this season like Danny and Erika. However, Sandy also had a very large outer footprint, upper-level divergence was very strong and it was actually embedding itself in the trough. The cool air to the northwest (relatively speaking) acted to start an extratropical transition process for Sandy as well, which it did not complete. All that it did was enlarge the storm greatly, which had huge repercussions as explained later.

On October 27, after briefly weakening to a tropical storm, the pressure dropped from 973 to 956 mb and Sandy regained hurricane intensity. About that time, the extratropical transition process aborted and Sandy became more tropical in appearance, in large part due to the record warm Gulf Stream which it traversed for the next 48 hours. That allowed the core of Sandy to redevelop and an eye to form once again, despite being embedded in a very large subtropical-type environment. Sandy was now the largest hurricane in the satellite era in terms of footprint. It deepened slowly until a few hours before landfall, dropping to a minimum pressure of 940 mb, with sustained winds of 100 mph (Category 2), on October 29 as it left the Gulf Stream. It weakened somewhat before landfall and the inner core dissipated over the cooler shelf waters resulting in Sandy becoming fully extratropical, but it was still a very remarkable storm.

The left turn to shore

The hurricane database goes back to 1851 and includes all tropical and subtropical cyclones known in the historical record. Quite simply, there is no storm like Sandy in the record. Few storms have travelled virtually east to west in the high latitudes, and none of them made landfall. However, the presence of a negatively-tilted trough to the southwest helped shift the steering currents into that direction on the afternoon of October 29 and Sandy moved northwest leading up to landfall. It is not uncommon for troughs to take negative tilts and back in itself, resulting in a wrong-way movement in the atmosphere, particularly from late fall to early spring. It is, however, very uncommon for such to occur with a hurricane in the vicinity, which was the case here. That resulted in a worst case scenario for the Jersey Shore and New York City with the storm surge and waves easily able to push into those areas.

Size matters

Even at landfall, the extremely deep storm – at the time of landfall the pressure was recorded at 945 mb (the second lowest pressure ever analyzed north of Cape Hatteras, with only the 1938 Long Island Express having a lower pressure at 941 mb) – was also exceptionally large. At that point, tropical storm force winds extended out upwards of 550 miles from the landfall point, encompassing the entire Northeast, south to coastal North Carolina, east to Nova Scotia and Bermuda, west to the central Great Lakes and north to southern Quebec and Ontario. Few places were spared Sandy’s impact in some form or another, and every hazard imaginable was included. Notice the map below showing the breadth of impacts.

Warning map for Hurricane Sandy. Notice how widespread the wind warnings were. Tropical type warnings were not issued for this event.

Warning map for Hurricane Sandy. Notice how widespread the wind warnings were. Tropical type warnings were not issued for this event.

A hurricane with snow? Yes!

Although not a factor in the Northeast (thanks to warm temperatures in New England and eastern Canada – during the strongest part of Sandy where I am, temperatures were in the upper 60s), the cold air to the west pushed southward into the central and southern Appalachians and entrained into Sandy’s circulation, even while the core was still tropical. As a result of the combination of very cold air and tropical moisture, very heavy snow fell in the higher elevations. Some high elevation areas in West Virginia experienced as much as 3 feet of snow, which is remarkable for late October there! (Even though parts of New England had similar amounts in late October 2011 from “Snowtober”, which was in itself a remarkable storm as well.) Just like with the Northeast storm in 2011, major power outages, crushed trees and even collapsed roofs were a major part of the story in the Appalachians as well, especially in West Virginia. With leaves still on the trees, it only takes 3 to 6 inches of snow to result in major impacts.

While it was too warm to snow in the Northeast, the central and southern Appalachians were battered with SNOW during Sandy.

While it was too warm to snow in the Northeast, the central and southern Appalachians were battered with SNOW during Sandy.

High population

Overall, more than 90 million people live in the area affected by Sandy. That is because it severely affected the entirety of the most heavily populated urban corridor in North America (the corridor from Boston to Washington with over 55 million people), and also spread inland into other urban areas near the Great Lakes and Southeast coast. There was wind damage, water damage and power disruptions in over 20 states, seven countries and two dependencies at one time or another. However, the most populous city in North America (New York City) and its suburbs – together with over 22 million people – took the brunt of Sandy, with coastal houses and buildings destroyed, many other buildings damaged by wind and water, subways and tunnels flooded and infrastructure battered. Overall damage was over $70 billion, the second costliest storm on record after Katrina, and sadly, at least 233 people lost their lives.

Historical precedents?

Although no storm, at least since 1851, has taken the same track as Sandy, there are some historical precedents that were at least somewhat similar.

  • October 29, 1693 Could Sandy have a twin 319 years to the day earlier? Little is known about this storm, but the limited anecdotal evidence suggests a similar track with very high impacts to the Northeast. The population was quite low then but based on the evidence available, it affected a very wide area. A landfall point (if any) cannot be pinpointed, but the high winds in the Northeast suggest landfall farther south due to the structure of extratropical lows.
  • October 9, 1804 – This was the well-documented “Snow hurricane”. The structure was likely similar to Sandy with a hybrid structure and cold air to the west. Heavy snow resulted in severe damage in the higher elevations, particularly from New York to Vermont, similar to the heavy snow during Sandy in the central Appalachians.
  • October 23, 1878Probably the storm most similar to Sandy in the modern database. It raced northward with somewhat of a left bent through the mid-Atlantic as it became extratropical. Despite making landfall in North Carolina, it produced hurricane conditions as far north as New England.
  • October 15, 1954 (Hazel) – A well-known storm that also pushed westward in a trough interaction, but not to the same extent. High winds to hurricane force extended along the entire east coast, while catastrophic flooding occurred in the Toronto region as the front stalled ahead of Hazel.
  • October 28, 1991 (The “Perfect Storm”) – Unlike the other storms on this list, it started as a more typical Nor’easter type low (think “Snowtober”). Like Sandy, a blocking ridge prevented it from escaping to sea, resulting in an east-to-west movement before a southerly turn. However, as it meandered in the Gulf Stream, it actually gained tropical characteristics. Coastal flooding and high waves were the main problems with this storm.
  • September 10, 2001 (Erin)This storm famous for its pre-9/11 turn actually moved NW in the high latitudes for an extended period while offshore from east of Bermuda to SE of Cape Cod.  It turned eastward safely out to sea on September 11 – when the world was changing forever. It was an unusual track for early September given the fact that troughs are weaker in the peak hurricane season.

Three years ago tonight, for almost all of us, it was a day and night that we all remember. We were all hunkered down trying to protect ourselves from the wind, rain, waves, storm surge, snow and (in a couple cases) tornadoes. Sadly, at least 233 people lost their lives either during the storm or in the aftermath (that number includes the Caribbean and other areas). For meteorologists, it was an unforgettable storm unmatched in the known record (although it may have very historical analogs?) due to its strength, track and impacts. We all learned a lot and we are stronger for the next time such a storm occurs.

I think the 1693 storm may be a good analog for Sandy, although little is known on that one. Maybe that is a storm worth studying, since we have good data from a recent storm to make a historical analysis. But it doesn’t mean we will be waiting 319 years for another similar storm!

Forecaster Craig Ceecee

Leave a Reply