A look at weather extremes and the big-picture climate connections.

How Fujiwhara Effect Will Toss Hurricane Sandy Into U.S.

By Adam Sobel
Guest Blogger

A vortex is a flow pattern in a fluid that has rotation about a center: water spiraling down the bathtub drain, the swirling eddies made by a canoe paddle, or a hurricane.

A marker dropped into the flow near a vortex -– a cork dropped into the water, for example -- will orbit about the vortex center. At the same time a vortex itself can also behave like a marker dropped into the fluid and move, like a cork, with the larger-scale flow in which it’s imbedded. If you paddle your canoe in a river, the vortices your paddle leaves behind will float downstream just like a cork would.

Infrared satellite image of Hurricane Sandy as it transitioned into a hybrid tropical/extratropical storm off the Carolinas on Oct. 27.
Credit: NOAA.

If two vortices come close enough to each other to get caught in each other’s flows, then each one acts like the cork; its center moves in the flow swirling around the other one. At the same time, if one is moving, the center about which the other is moving is itself moving, and vice versa, and the two carry out a joint maneuver. Or a dance. Hurricane Sandy is moving into position to do this dance with the upper-level trough whose eventual predicted merger with it has led to the “Frankenstorm” nickname. 

How the dance goes depends on the two vortices’ directions of rotation and their relative strengths. If they are spinning opposite ways but are identical in every other way, they will move in parallel straight lines, perpendicular to the line through the two vortex centers. If they have the same spins, they will circle each other, both orbiting about their “center of mass” (if they are equal in strength, this is just the midpoint of the line between them). This do-si-do is known as the “Fujiwhara effect,” named after Sakuhei Fujiwhara, an early 20th century Japanese scientist who first explained it in meteorological literature.

Occasionally, two tropical cyclones get close enough to one another to do the Fujiwhara dance. It happens only rarely in the Atlantic. Computer models predict that Sandy is going to do it, not with another tropical cyclone, but with the vortex center associated with the larger extratropical storm. This dance move is going to be happening right at landfall. In fact, it is what’s going to bring Sandy ashore, leading to so much risk of flooding and even possibly wind damage inland.

Because the two vortices are neither equal in size or strength, nor simple in structure, they will merge partway through their first mutual orbit. The dance of courtship will end with hybridization, making the resulting storm either “perfect” or “Franken,” as you prefer.

NOTE: The movie above (via Weatherbell Models) shows temperature at 500 hPa (5-6 km above the surface) from a recent run of the GFS model. The period shown begins at 12 UTC on October 28 – early Sunday morning east coast US – and ends two days later at 12 UTC on Tuesday the 30th, with frames separated by 6 hours. Sandy, being a tropical system, is warm-cored, thus red, while the extratropical trough is cold-cored, and blue.

Those who wish to see the Fujiwhara effect in terms of the most natural variable —vorticity — can watch a short animation showing the geopotential height (contours), wind barbs, and vorticity (shading), from the same model for the same period. Vorticity is just what it sounds like:  it measures the spin or vortex-ness in the fluid. A vortex is just a compact blob with high vorticity.  The first frame sits still for a while so you can see the high vorticity region (reddish) associated with the extratropical trough, and the even higher vorticity region associated with Sandy. Images in the latter animation from Kyle Griffin.

Adam Sobel is a professor at Columbia University, in Earth and Environmental Sciences and Applied Physics and Applied Mathematics. He is an atmospheric scientist who specializes in the dynamics of climate and weather, particularly in the tropics, on time scales of days to decades. He is author or co-author of more than 85 peer-reviewed articles and has received the Meisinger Award from the American Meteorological Society and the Excellence in Mentoring Award from Lamont-Doherty Earth Observatory of Columbia University. He lives in New York with his wife and two sons.

Related Content 
Hurricane Sandy’s Five-Fold Flood Threat, with Local Maps 
Historic Sandy Poised to Blast Mid-Atlantic, Northeast 
How Hurricane Sandy Can Become a Frankenstorm 
East Coast Facing Major Threat From Hurricane Sandy 
Hurricane Sandy Poses Growing Threat to East Coast 
Grim Storm Scenarios Loom For Mid-Atlantic, Northeast

« Extreme Planet


By Stu Ostro (Atlanta, GA)
on October 27th, 2012

Hi Adam,

Isn’t it just getting captured/absorbed by the trough, not really doing a Fujiwhara?

Reply to this comment

By Adam Sobel (New York, NY)
on October 27th, 2012


Thanks for the comment.  I would argue the pair does about half a rotation of a Fujiwhara before being absorbed.  You can see how Sandy’s circulation pulls the ET system down to its south, so that the ET’s circulation is then more easterly (onshore) at Sandy’s location, which helps to catapult it onto land.  That’s mutual rotation by the standard dynamics of the Fujiwhara effect, even if it doesn’t last long in this case before merger occurs.


PS the video didn’t come across quite right in youtube, we are working on that.

Reply to this comment

By Werner Loell (02871)
on October 28th, 2012

We need to resuscitate our air shelters from the 50s as climate change will demand more in-place protection from weather insurgencies!

Reply to this comment

By Aaron D Rodriguez (Kingsville,Texas)
on October 28th, 2012

The hurricane is projected to dissipate more or less Thursday. Wouldn’t this effect strengthen the storm?

Reply to this comment

By Tim Hillman (West Warwick, RI)
on October 29th, 2012

Of course they built those fallout shelters underground….I don’t wish to walk down into a tank on the coast of RI and hope the storm stays above me and the shelter doesn’t leak. Yikes.

Reply to this comment

By Lewis Cleverdon
on October 30th, 2012

Adam - very good to see this clear explanation and graphic. Re tweaking the latter, I’d recommend running it slower, so it can be appreciated, and having a good pause before the start of each run. These changes would raise its cogency.



Reply to this comment

By Stu Ostro
on October 30th, 2012

Was extremely busy with the storm and am catching up ...

Seeing the video, I can understand why you are suggesting it, but it still looks to me like just a trough absorbing the hurricane, and that animation is showing the temperature field, not circulations.  And even looking at the 500 mb vorticity field, it still looks like the trough’s and Sandy’s vorticity commingled, not Fujiwhara-ed (sp?), and there was no second surface cyclone, only Sandy, unlike when two typhoons Fujiwhara.

Reply to this comment

By Adam Sobel (New York, NY 10025)
on October 31st, 2012


I define the Fujiwhara effect as the orbiting of two same-signed vortices about their center of mass due to the advection of each by the flow field associated with the other.  If you accept this definition I don’t see how you can argue that it wasn’t operative here to some extent.  I think if you look at the vorticity and flow maps at 500 hPa (for example) it’s clear that Sandy’s circulation helped to pull the ET vortex to the south and then east below her, and then that the ET vortex’ circulation, now pointing westward at Sandy’s location, helped push Sandy onshore, and that all this happened before the merger was fully complete (though it was certainly in process in the latter phases). 

It’s true that temperature is not really the right field if one wants to diagnose this precisely, but it is a rough marker of what was happening.  The right variable would be potential vorticity in some weighted vertical average.  But by PV invertibility, temperature is like a smoothed version of the baroclinic component of PV.  To really make this argument precise we would also have to quantify the relative influence of the flow at different levels on the advection of Sandy - that is, define the steering level.  But I think the relevant flow features here were occurring over a deep enough layer and are apparent enough by eye to be able to call it pretty clearly.

It’s also true that the ET system didn’t have a TC-like surface circulation, but I don’t see how that matters unless you want to argue that the steering level relevant for Sandy was at the surface.  It’s true that we often associate Fujiwhara with TCs, but the effect is generic to any vortices (I believe Fujiwhara himself demonstrated it by experiments in a water tank) so I don’t see the value in restricting the definition in ways that exclude non-TC vortices when the same essential dynamics is operative.

For what it’s worth, it was also called Fujiwhara by the forecaster who wrote the discussion for the Upton NWS office on Monday.


Reply to this comment

Name (required):
Email (required):
Enter the word "climate" in the box below:

[+] View our comment guidelines.

Please note: Comment moderation is enabled. Your comment will not appear until reviewed by Climate Central staff. Thank you for your patience.