Development of an index of storminess as a proxy for dry season severe weather in Florida and its relationship with ENSO

The author has investigated the characteristics of Florida tornadoes since 1989 (http://www.srh.noaa.gov/mlb/research.html ). Since late 1997 the focus has been on the relationship between ENSO and significant tornadoes in the Florida dry season (November through April). Hagemeyer (1998, http://www.srh.noaa.gov/mlb/sls19paper10-1.html ) found that during strong El Nino/La Nina conditions the atmosphere over Florida and the eastern Gulf of Mexico was more/less favorable for significant dry season tornadoes. In particular, strong El Nino's appeared to cause a stronger and more southerly than normal jet stream and a greater than normal number of extratropical cyclones (ET) over the eastern Gulf of Mexico and Florida. La Nina's resulted in fewer ET cyclones, below normal rainfall, and increased chances of wildfire and drought. Recently, a comprehensive statistical investigation was completed to quantify this relationship.

Environmental variables (monthly 250 mb U anomaly, monthly and daily mean sea level pressure) and tornado variables (tornado days, F2 tornado days, total seasonal tornado damage, and number of $5 million tornado damage events) calculated over an area representative of Florida and the eastern Gulf of Mexico (24-31 N, 80-88 W) were correlated with monthly values of NINO 3.4 (5N-5S and 120-170W) for the 1958-1999 dry seasons using multiple regression techniques. The goal was to attempt to forecast significant tornadoes on a seasonal basis.

The statistical relationships between NINO 3.4 and the seasonal tornado data showed some skill, but were not significant at the 95% level. The relationship between NINO 3.4 and seasonal 250 mb U and MSLP was very strong and significant at the 99% level. Nearly all dry season severe weather in Florida is associated with warm sectors of ET cyclones. Extratropical cyclones can bring damaging straight line winds, large hail, flooding rain, and hazardous marine events, as well as tornadoes. It became clear the best approach to seasonal severe weather forecasting would be to define a proxy variable indicative of dry season severe weather potential focused on ET cyclones, rather than tornadoes alone. The author developed the dry season Low Pressure Index (LPI) for Florida and the eastern Gulf of Mexico. The LPI represents a proxy for seasonal "storminess" and is a measure of both season-to-season and day-to-day variability (Hagemeyer 2000 http://www.srh.noaa.gov/mlb/12thclimate-bh.htm ). The LPI values for the 1958-98 dry seasons were regressed on monthly May - April Nino 3.4 and found to be significant at the 99% level. The LPI was closely correlated with seasonal MSLP (R2=.73, 99% significance) and 250 mb U (R2=.72, 99% significance), but was a better statistical proxy for tornado variables. LPI was also considered to best represent the synthesis of the sensible weather effects of ENSO.

The results of this study confirm ENSO has a significant effect on Florida's weather during the dry season. The presentation of this paper will focus on the further refinement of the LPI and investigation of its physical relationship to severe weather potential. Preliminary results were encouraging and indicated successful seasonal severe storms forecast may be possible. Final results for the 1999-2000 Florida dry season, the first full season of experimental forecasting, will be reviewed. An example of how such forecasts might be developed and applied to other regions of the US will be presented. The issues of how to properly communicate such forecasts for use by customers and the risks and benefits involved will also be presented.