23rd Conference on Hurricanes and Tropical Meteorology. Dallas, TX, Amer. Meteor. Soc., 415-418.
EL NINO AND SIGNIFICANT TROPICAL AND HYBRID CYCLONE TORNADO EVENTS IN FLORIDA
Bartlett C. Hagemeyer e-mail: bart.hagemeyer@noaa.gov. National Weather Service Melbourne, Florida 32935 1. INTRODUCTION Hagemeyer (1998) investigated significant tornado events associated with tropical and hybrid (T/H) cyclones in Florida. Significant T/H tornado events were defined as those with at least one F3 tornado, or any killer tornado, or 5 or more tornadoes in 24 hours with at least one being rated F2. A total of sixteen T/H cyclone tornado events were found to be significant. Information on these 16 cases is contained on Table 1. The track segments of the 16 T/H cyclones during their tornado-producing phase are shown on Figure 1. Fifteen of 16 cyclones were in the eastern Gulf of Mexico, and produced tornadoes in outer rainbands in the right front quadrant with respect to cyclone motion. Fourteen of the 16 cyclones were moving between 360 and 60 degrees during tornado production. Table 1 Florida Significant Tropical and Subtropical Cyclone Tornado Events (1882-1996)
case # date 9/9/1882 10/20/41 9/19/47 9/22/47 9/4/48 6/8/57 10/14/64 6/18/72 5/8/79 9/3/79 6/17/82 9/26/82 10/3/92 11/15/94 10/4/95 10/7/96 system type hurricane t. storm hurricane t. storm hurricane t. storm h. isbel h. agnes hybrid h. david st storm hybrid hybrid ts gordon h. opal ts josephine #tors 5+ 6 9+ *15+ *19 10+ 10 *6 * 10+ 20 f# time (est) eve-am 10th afternoon 0030 eve-eve23rd late eve evening aft-eve am-am19th am-pm am-am 4th am-am 18th early am am-pm evening afternoon am - aft death/ injury 6/17 1/1 2/100 0/1 2/12 0/0 0/48 7/119 1/47 0/0 1/13 1/7 4/77+ 1/40 1/3 0/0 location/stm ctr relative nw-n central/rf n central/ rf franklin/ rf wc-n east/rf panhandle/rf nc-n east/ rf se-e central/lf-rf keys-ec/ rf centrl/rf e coast/lf sw-e central/rf southwest/rf w central/rf e central/ rf panhandle/rf-cntr central-ne/rf
Figure 1. Track segments during tornado-production phase of the 16 tropical and hybrid cyclones that produced significant tornadoes in Florida (from Hagemeyer 1998). Hagemeyer (1998) noted that T/H cyclone central pressure, organization, and amount of deep central convection were poor predictors of significant tornado potential. Indeed, the best indicator of tornado potential was found to be the existence of relatively intense convective elements in primary outer rainbands in the right front quadrant of T/H cyclones. Hagemeyer also noted the importance of hybrid cyclones in producing significant tornadoes. Half of the 16 cases were arguably hybrid cyclones and nearly all showed some hybrid influence. The concept of "hybridization" became a focal point for further investigation. Basically, a hybrid cyclone is one that has both tropical and extratropical elements. Hagemeyer found that, for Florida, the hybrid cyclones typically had tropical characteristics at low levels and extratropical influences such as a 500 mb shortwave aloft, and were in the early stages of hybrid evolution. Often high pressure to the east resulted in a highly asymmetrical low-level wind field with strong low level winds and intense rainbands well east of the center and weaker winds and minimal convection west of the center. The track segments of nearly all of the significant tornado producing T/H cyclones effecting Florida (Fig 1.) hint at the influence of westerly steering currents and shear, and thus a high probability of hybrid influences, it is likely that there are certain favorable synoptic patterns that can be identified. This paper presents results of a preliminary investigation into the seasonal synoptic aspects of T/H cyclones that produced significant tornadoes in Florida. 2. SEASONAL CHARACTERISTICS OF SIGNIFICANT TORNADO PRODUCING TROPICAL AND HYBRID CYCLONES A review of seasonal tropical cyclone tracks for the 13 years with 16 significant tornado events revealed interesting results. The seasonal track plots of three of the worst years for T/H cyclone tornadoes, 1972 (2a), 1982 (2b), and 1992 (2c) indicated only two Cape Verde type cyclones developed in these three years. Further investigation revealed that 8 of the 13 years (10 of 16 total cases) had similar seasonal patterns. A composite track map for these 8 years is shown as Figure 3a. Remarkably, in these 8 years only 5 tropical cyclones formed

south of 30 N and east of 60 W, while T/H cyclone formation was concentrated over the Gulf of Mexico and northwest Carribean Sea. These 8 years accounted for 90% of T/H cyclone tornado deaths.

Figure 2a.

Figure 2b.

Figure 2c.

Hagemeyer and Hodanish (1995) looked at all tropical cyclone tornado cases in the history of Florida and found that in only 4 out of 67 cases the cyclones originated in the Atlantic. Therefore, an increase or decrease in Atlantic cyclone activity would not necessarily affect tornado odds in Florida, unless the factors that lead to the Atlantic decrease either increased the chances of Gulf or Carribean T/H cyclone development and/or increased the chances that these Gulf and Carribean cyclones would produce significant tornadoes. Clearly, more cyclones forming in the Gulf and Carribean would mean a greater chance of rainbands affecting Florida and a greater chance of hybrid influences which would increase the chances of significant tornadoes. Five of the 13 years with significant tornadoes (6 of 16 cases) had relatively active Atlantic seasons. A composite of the cyclone tracks for these 5 seasons is shown as Figure 3b. Of the six cases in these 5 years only one resulted from an Atlantic TC, the rest were from the Gulf of Mexico and Caribbean. The 6 cases in these 5 years were responsible for only 10% of T/H cyclone tornado deaths. Figure 3b shows that Gulf and Carribean cyclone activity was also high. Five of the 6 cases in these years occurred very early or late in the season, either before or after peak Atlantic activity had occurred, when hybridized cyclones are most likely to develop in the Gulf or Carribean. There is certainly no indication that more active Atlantic seasons would necessarily decrease T/H cyclone tornado activity in Florida.

Figure 3a.

Figure 3b. Figures 3a-b. Composite seasonal tropical cyclone tracks for 1882, 1941, 1947, 1948, 1957, 1972, 1982, and 1992 (3a), and for 1964, 1979, 1994, 1995, and 1996 (3b).
3. POSSIBLE RELATIONSHIP OF EL NINO TO SIGNIFICANT TROPICAL AND HYBRID CYCLONE TORNADOES It is generally accepted that El Nino's, particularly strong El Nino's, inhibit northern Atlantic tropical cyclone activity, especially the so-called Cape Verde storms, by increasing westerly shear at mid and upper levels. Figure 3a would seem to indicate that conditions that favor reduced tropical Atlantic activity might favor increased activity in the Gulf of Mexico or northwest Caribbean and that El Nino's might therefore favor increased T/H cyclone significant tornado activity in Florida. The worst T/H cyclone tornado event in Florida history occurred with Hurricane Agnes in June 1972 (Fig 2a) when strong El Nino conditions were rapidly increasing as measured by the Multivariate El Nino/Southern Oscillation (ENSO) Index. However, both Atlantic and Gulf/Caribbean cyclone activity was reduced in 1972. The only season with two killer T/H cyclone tornado events, 1982 (Fig 2b), was also during the early stages of a strong El Nino. The 3rd deadliest T/H cyclone tornado event, 3 October 1992 (Fig. 2c), occurred during the later stages of a strong El Nino when the ENSO Index was the second highest for October ever. There would appear to be a relationship between strong El Ninos and significant Florida tornadoes, and one might have expected a significant T/H cyclone tornado event to have occurred in 1997. However, the onset of the 1997-98 El Nino was one of the fastest and strongest ever with the greatest intensity ever during the tropical cyclone season of 1997. Atlantic cyclones were greatly reduced in 1997, but no T/H cyclones formed at all in either the Caribbean or Southern Gulf of Mexico. These regions were actually more suppressed than the Atlantic. Perhaps there is a crucial El Nino strength threshold that inhibits all lower latitude T/H cyclone development. To further illustrate possible El Nino influences two brief case studies of two of the most devastating T/H cyclone tornado events (6/18-19/1972, 7 killed, and 10/3/1992, 4 killed) are presented. Figures 4a-b show the mean monthly 200 mb wind anomaly for June 1972 and October 1992 respectively. Figure 4a shows a strong westerly wind anomaly from the Gulf of Mexico across the Atlantic to Africa. The ENSO Index for June 1972 was surpassed only by June 1982 and June 1997 (For June 1997 a mean northwesterly 200 mb anomaly extended across the southern Gulf and Carribean). Figure 4b shows a very similar, but even stronger westerly wind anomaly extending from the Gulf Mexico to Africa. The ENSO Index for October 1992 was surpassed only by the El Nino of 1986-87.

Figure 4a.

Figure 4b. Figures 4a-b. Mean monthly 200 mb wind anomaly (m/s) for June 1972 (4a) and October 1992 (4b).
4. CONCEPTUAL MODEL OF SIGNIFICANT TROPICAL/HYBRID CYCLONE TORNADO ENVIRONMENTS To successfully diagnose significant T/H cyclone tornado events in operational practice it is necessary to understand the conceptual model without undue concern to the "official" categorization of the cyclone or the state of ENSO. Hagemeyer and Matney (1993), Hagemeyer and Hodanish (1995) and Hagemeyer (1997 and 1998) have all noted the importance of hybrid influences and outer rainbands to significant tornado production. Three brief examples will be presented to illustrate this concept. Figures 5a-b show mean daily synoptic analyses for 18 June 1972 and 3 October 1992 respectively. Figure 5a shows Hurricane Agnes (category 1) centered just north of the Yucatan Peninsula on 18 June 1972. Agnes exhibited deep central convection, but it was very much a hybrid cyclone characterized by extensive cloudiness and deep convection extending more than 800 miles to the east, southwest upper shear, and a 500 mb shortwave to the west (see Fig. 6a). Typical of a hybrid cyclone, an 850 mb pseudo warm front extended out of the northeast sector with an extensive rainshield to the north and an 850 mb jet developed well east of the center. Organized outer rainbands with intense convection formed in this environment characterized by strong lowlevel helicity and shear in the "warm sector." From the morning of the 18th to early on the 19th the most devastating T/H cyclone tornado outbreak in Florida history occurred as extreme outer rainbands spread from the Keys to northeast Florida while Agnes moved north making landfall in the Florida Panhandle. Tornadoes to F3 intensity were reported up to 300 miles northeast of Agnes, killing 7.

Figure 5a.

Figure 5b. Figures 5a-b. Mean daily 850 mb synoptic analyses with 850 mb jet axes and 500 mb steering currents indicated by small and large arrowheads, respectively, for 18 June 1972 (5a) and 03 October 1992 (5b). Figure 5b shows a "no-name" hybrid cyclone in the central Gulf of Mexico on 3 October 1992. There was absolutely no deep convection near the circulation center (white "L" on Fig. 6b) of this cyclone, but an environment favorable for significant tornadoes developed several hundred miles to the northeast just as in the case of Agnes. Low level winds were weak around the circulation center, but a 50-knot 850 mb jet developed over Florida Peninsula and produced intense bands of convection where it intersected the pseudo warm front in an environment of very high low level shear and helicity. Extensive cloudiness and heavy rain occurred north of the front. Tornadoes to F3 intensity killed 4 near Tampa making this the 3rd deadliest event. Figure 6c shows Hurricane Opal (category 3) nearing the Florida Panhandle on 4 October 1995. Opal had no discernable hybrid characteristics, however the mesoscale environment of the detached primary outer rainband in the right front quadrant on Fig. 6c was quite similar to that of the unnamed cyclone of October 1992, Hurricane Agnes of June 1972, and Tropical Storm Gordon of November 1994 (see Spratt et. Al. 1996). An F2 tornado developed in this outer rainband, killing 1 person in a trailer. In general, the strongly sheared hybrid environment is a more dangerous tornado situation. Hybrid cyclones are similar to intense extratropical cyclones except they exhibit less baroclinicity, upper shear, and instability. However, they generally exhibit greater low level convergence in narrow rainbands and more extreme values of low level shear, streamwise vorticity and helicity.

Figure 6a.

Figure 6b.

Figure 6c.

Figures 6a-c. Visible satellite image of Hurricane Agnes, 18 June 1972 at 1800 UTC (6a), IR satellite image of a hybrid cyclone in the eastern Gulf of Mexico (white "L"), 03 October UTC (6b), and an IR satellite image of Hurricane Opal, 04 October UTC (6c) 5. CONCLUDING REMARKS Strong El Ninos which inhibit cyclone development in the tropical Atlantic appear to increase the chances of significant T/H cyclone tornado events in Florida. However, extreme El Nino events like 1997 may inhibit nearly all cyclone formation at lower latitudes in the Gulf of Mexico and Carribean as well as the tropical Atlantic. During strong EL Nino years there are generally fewer T/H cyclones in Gulf of Mexico and northwest Carribean, however, those that do develop in a westerly shear environment generally unfavorable for pure tropical cyclone development are more likely to be hybrid cyclones, move northeast and affect Florida, and produce significant tornadoes. Continued efforts should be made to improve understanding of the hybrid cyclone and refine the conceptual model. Efforts should also be continued to improve our understanding of the mesoscale environment of T/H cyclone outer rainbands. 6. REFERENCES see http://www.srh.noaa.gov/mlb/?n=research