NSSL SWAT Case Study - 23 February 1998

Central Florida Tornadoes

A powerful storm system approached Florida on Sunday, February 22, 1998, and spawned a tragic outbreak of damaging tornadoes.  One wave of severe storms crossed northern Florida early in the day, and then several supercells formed in Central Florida during the afternoon, but the storms which defined this tornado event took place in the early morning hours of Monday, February 23, 1998.  The result was a devastating loss of life and property for which three long-track, tornadic supercells were mainly responsible.

Presented here are various WSR-88D radar images from the Melbourne, FL radar (KMLB).
The radar is always to the south and east of the images presented here with the exception of the earlier storms at the bottom of this page.  Range Folded echos have been removed from all images to improve clarity.

Included in some of the WSR-88D images is output from NSSL's Mesocyclone Detection Algorithm (MDA) and Tornado Detection Algorithm(TDA).  A yellow circle depicts a mesocyclone.  The red-in-yellow circle depicts a mesocyclone whose base is at the lowest radar scan (where it is a more likely tornado threat).  A red inverted triangle depicts a Tornadic Vortex Signature (TVS) detection.

These images are from NSSL Radar and Algorithm Display System (RADS) that is part of NSSL Warning Decision Support System (WDSS).  The WDSS is a real-time operational warning guidance system that contains images and overlays of NSSL enhanced WSR-88D algorithms and novel display concepts.  The WDSS was running in real-time during this event at the
National Weather Service Office in Melbourne, Florida, and offered a unique perspective on the event that very few NWS offices in the U.S. have yet to enjoy.  These links contain more information about this event:



  • THE DEVASTATING TORNADOES (2/23/98, 0300 UTC - 0700 UTC):

  •   Description:  This storm tracked across nearly the entire Florida peninsula, beginning in Pasco County on the west coast and moving offshore near Daytona Beach only some two hours later.

    The Two Panel TVS image demonstrates the presence of an intense mesocyclone in the absence of a reflectivity "hook.."  The NSSL MDA and TDA use velocity data, not reflectivity data, when searching for mesocyclones and tornadoes.  This particular velocity frame has a dealiasing error, meaning the radar could not properly measure the radial component of the wind.  The affected area is the bright green stripe underneath the mesocyclone.  Dealiasing errors show up frequently enough to present a significant threat to algorithm performance.  However, this error occured in a place where it did not hinder the algorithms.

    Description:  This storm developed a significant mesocyclone around 0426 UTC in Lake county.  Much like the Daytona Beach storm, the Winter Garden storm produced a long track of damage across Central Florida before moving out to sea.

    The reflectivity loop follows the lifetime of the Winter Garden storm; Winter Garden is near Ocoee which is just left of the center of the frame.  The TVS associated with this storm moves from the middle-lefthand side of the image to the upper-righthand corner.  Midway through the loop, the Kissimmee tornado touches down and shows up as a second TVS further south and west.

    The Tornado Touchdown image shows that the MDA and TDA did their jobs with pinpoint accuracy in identifying the mesocyclone and tornado.  It is said to be a (likely) tornado because efforts are still underway by the Melbourne, FL, NWS to verify the precise track of each tornado.

    The Tornado Northeast of Winter Garden was again detected despite a dealiasing error.  The dealiased data has been removed from the image, hence the blank stripe immediately southeast of the mesocyclone.

    Description:  This storm developed a significant mesocyclone around 0520 UTC.  Tornado damage associated with this long-lived mesocyclone primarily affected the area from Intersession City through Kissimmee to near Titusville and Cape Canaveral where the storm headed out to sea (NWS, Melbourne, FL).

    The Tradgedy at Kissimmee panel shows radar products at the time at which the tornado was only one half mile noth of downtown .  The 4-Panel Mesocyclone Images  show that the mesocyclone became even more ominous as viewed by the WSR-88D when it had moved east of Kissimmee.

  • THE EARLIER STORMS (2/22/98, 2000 UTC - 2/23/98, 2300 UTC):
  • Description:  Several small supercells developed as a result of afternoon heating.  The Melbourne NWS office issued a tornado warning for the storm which has  a mesocyclone indicated in the velocity data on the first image.  However, none of these cells produced significant tornadoes, a fact which is most likely linked to two factors:

    1) The violent, overnight storms formed on and moved along a clearly defined surface
        Outflow Boundary.

    2) The violent, overnight storms took place in an environment which had more directional as well as speed shear due to the presence of a very strong jet max.

    The Near Range Storm  exhibits a pronounced reflectivity "hook," on it's south side, and a divergence signature northeast of the "hook," both of which are consistent with supercell structure.  A Closer Look  depicts these supercell characteristics, along with associated boundaries.  Despite the impressive structure, this storm did not produce a tornado.

    A similar storm nearby did, however, produce a tornado/waterspout near Indiatlantic.  The "hook" is again impressive, and is outlined in the image.  The velocity portion of the image confirms the existence of a broad, well established mesocyclone which is not present in the other Near Range Storm.

    Back to NSSL SWAT Case Study Table of Contents Page.