The Brooklyn, NY Tornado of August 8th, 2007  

  by Philip Lutzak August 2007

 

The Resultant Floods and Tornado

 

 

   From the evidence previously discussed, the merging of these two systems resulted in a backward propagating SCS, with its maximum convergence zone and thus heaviest precipitation at the rear of the system.  This type of slow-moving MCS often produces flash flooding, as indeed occurred here - rainfall was 3.47 inches at JFK, most of which (3.18 inches) fell in the 2 hours from 6-8AM (JFK metars). In addition, the back (northwestern) edge of the system eventually intersected an area of strong vertical wind shear and significantly high SRH, causing supercell thunderstorms to develop, one of which produced the EF2 tornado in New York City.

  In Figures 1 and 2 below are the reflectivity and velocity radar images from the Mount Holly, NJ radar, 76 miles southwest of New York City, as the tornado moved through Staten Island and Brooklyn.  These images have been enlarged for close-ups of the New York City area. Understanding that green colors represent winds moving towards the radar (northeast) and reds represent winds moving away from the radar (southwest), Figures 1b and 2b clearly show the signature of a rotation of the mesocyclone within the supercell thunderstorm. In the same images we can also see the eastward motion of the mesocyclone as it moved over extreme northeast Staten Island, across New York Harbor, and into western Brooklyn. A tornado warning was issued at 6:28AM, just as the tornado was moving across Staten Island as an EF1. After touching down on Staten Island, the tornado lifted up briefly over the waters of New York Harbor, but touched down again as an EF2 in Brooklyn two minutes later at 6:30AM.

 

     
  Figure 1a. Radar reflectivity 2007-08-08 1026Z, or 6:26AM EDT, showing heavy thunderstorms with very high rainfall over the area.  Courtesy NOAA NCDC archives and Mark Thornton.       Figure 1b. Radar velocity image 2007-08-08 1026Z, or 6:26AM EDT. Green tones are towards the radar (to the southwest), while reds are away from the radar site. Courtesy NOAA NCDC archives and Mark Thornton.  
     
  Figure 2a. Radar reflectivity 2007-08-08 1030Z, or 6:30AM EDT. As in Figure 1a, high rainfall rates associated with an HP supercell were occurring. Courtesy NOAA NCDC archives and Mark Thornton.       Figure 2b. Radar velocity image 2007-08-08 1030Z, or 6:30AM EDT. Green tones are towards the radar (to the southwest), while reds are away from the radar site. Courtesy NOAA NCDC archives and Mark Thornton.  

 

  It should be noted here that the supercell that produced the tornado was an HP supercell, or High Precipitation supercell. From the 2007-08-08 1030 radar, it appears that there may have been as many as three HP supercells in close proximity. Note that they are not discrete supercells, but rather embedded within the surrounding convection. Note also the "kidney bean" shape, a common configuration of HP supercells on radar reflectivity images. These are particularly dangerous when accompanied by a tornado, since the heavy precipitation usually obscures the tornado from view. Eyewitness accounts from Brooklyn, NY support this observation.

 

Warm Front Lift One interesting factor that may help explain the development of both the flooding and tornado over New York City is an interaction of this MCS with a developing warm front over the area. Figure 3 at right shows that the HPC chose to add a second warm front to the surface features at 12Z (compare to the 9Z surface map). Looking at the larger 12Z sfc version, it appears this warm front represents a wind shift due to the the deepening surface trough extending from the Canadian Low rather than any appreciable temperature or humidity change. We know from the OKX skewT and JFK meteogram that winds were very light southerly at the surface at 10Z or so over New York City and Long Island but then shifted to faster westerly during the MCS passage. So this warm front feature may have been developing over the NYC area before the HPC drew it in further east at 12Z. Figures 4a and b below, the 10Z 850 and 700mb charts, are annotated to show the wind shift line from south-southwest to west above the surface. Over the New York City area this line was coincident with the convergence zone at the back end of the MCS. Was it possible that there was some enhanced lifting due to a developing warm front? Certainly the nose of the fast westerly LLJ was impinging upon the north-south axis of where the warm front would be developing at 1030-11Z, and this is the area where the supercell and tornado occurred. So it is possible that a developing warm front enhanced lift in the rear flank of the MCS.

Warm Front Shear Finally, the wind directions and speeds in Figures 4a and b below explain the vertical wind shear that was present. Note the veering with height from 850 to 700mb ahead of the wind shift line/convergence zone. We also previously observed that there was a small inversion over the New York City area,  corroborated by the light southerly winds at the surface, and that the most unstable CAPE was quite high (Figure 5a - page 4), more evidence of the developing warm front in the area. Figure 4c, below right, shows the resultant area of 40 knot wind shear intersecting the northwestern, rear edge of the MCS. So it seems likely that much of the low level shear that aided in producing the supercells was caused by the developing warm front and deepening trough extending southward from the Canadian low pressure area.  

 

Figure 3. Surface weather map showing conditions at 12Z or 8AM EDT on 08-08-2007. Courtesy HPC. Larger version.

 

 

Figure 4a. 850mb heights/temp/dewpoint and wind, annotated to show the wind shift line in purple. Note the southerly wind flow ahead of the wind shift line. Larger image.

 

 

Figure 4b. 700mb heights/wind/temp, , annotated to show the wind shift line in purple. Note the winds are more southwesterly ahead of this line than the 850mb winds in Figure 4a. Larger image.

 

Figure 4c. 0-6km shear values. Note the area of 35-40 knot shear abutting the NW side of the MCS.  Larger image.

 

 

Summary  The difficulty in forecasting this situation was trying to ascertain the resulting conditions when a fast-moving DCS, associated with derechos and downbursts, was intercepting a slower-moving SCS associated with heavy, possibly flooding rains. From the evidence presented here, the final, merged MCS was an SCS, or Severe Convective System, characterized by slower movement, abundant mid level moisture, and convection concentrated at the rear of the system. It is likely that the faster moving cold pool and gust front of the DCS enhanced convection along the gust front of the SCS as it intercepted it.

  As the already energized, highly convective back edge of this system moved over New York City shortly after 5AM, a wind shift from southerly to strong west-southwesterly due to a developing warm front in the low to mid levels may have enhanced the convection along the convergence zone. At this time, as the flooding rains began, an area of vertical shear of 35-40 knots intersected the northwestern side of the SCS, causing supercell thunderstorms to develop within the already heavy convection. As was previously shown, the SRH values were high enough to produce tornadogenesis in one of the supercells moving over New York City. Also, because this was an SCS type of MCS, the system moved more slowly, causing higher rainfall amounts than otherwise might have occurred. What is interesting about this scenario is that while most MCSs that produce flash flooding events occur in low-shear environments, in this case it appears that New York City was in the unfortunate and somewhat rare location during the summer season where an MCS with flooding rains in a generally low-shear environment became impinged upon by enough vertical shear on its back edge to cause a tornado along with the flash flooding. 

 

  As a final note, I would like to add a quote/comment from Steve Corfidi of SPC: "Concurrent flash flood and tornado events are most common during the cool season when the mean flow may be strong but largely unidirectional. In these situations deep shear may be limited, but low level shear may be heightened near storm outflow boundaries. Throw in a nearly saturated environment (weak cold pools), and you have the recipe for isolated tornadoes in addition to back-building and/or training storms with flash flooding."  So it should be emphasized that this was an event more commonly expected in the cool season, although in this particular situation the impetus for the shear in a nearly saturated environment with strong unidirectional mean flow came from a developing warm front rather than from storm outflow boundaries.

 

Unusual aspects of this event:

1. The merging of two MCSs of different types (DCS and SCS).

2. Flash Flooding with a strong tornado due to the juxtaposition of low and high shear environments. 

3. Frontogenesis (in this case warm frontogenesis) within the area of an MCS.  

4. An event during August more commonly associated with the cool season. 

 

  For improvements in future forecasting techniques, it seems that a case study of situations where one MCS has overtaken another could be quite fruitful if it could help to determine rules on the outcomes when different MCS types combine.

 

 

 

Author's Note:  Many thanks to Stephen Corfidi of SPC for the use of his analysis on MCS types and especially for his review of this work. Also many thanks to Professor Lee Grenci of Penn State for his keen insights into flash flooding and low-shear environments. Finally, thanks to Lee Grenci and Mark Thornton for providing critical radar images to me. I couldn't have done the correct analysis without them.

 

References  

Forecasting MCS Mode and Propagation, Corfidi 1998

Corfidi on Cold Pools and MCS Propagation, 2003

MCS Lecture by Chuck Doswell

MCS Systems and Squall Lines - Scott Dimmich

New York State Tornadoes by County

Arkadelphia AR, 1997 - A Cool Season Example of Flash Flooding with Tornadoes

 

 

Previous Page: The Intersection of Two Different MCS Types

Pages 1 2 3 4 5 6