MSU Symposium Notes

A few running notes from today’s Southeast Severe Storm Symposium at Mississippi State…

Alan Gerard of the Jackson NWS office (Alan is the Meteorologist In Charge) is doing a presentation now called “Recent Trends in Tornado Statistics and Associated Warnings; Perceptions and Misperceptions”

Perception 1

“The large majority of fatalities associated with tornadoes are caused by violent tornadoes”

Recent trends do NOT support this.

Only 39 percent of tornado fatalities were caused by F4+ tornadoes during the most recently completed decade.

The number drops to 14 percent from 2001 to 2006.

Data now suggests that only about 40 percent of fatalities are caused by F4 or F5 tornadoes. Recent trends are lower… from 2001 to today only 11 percent of tornado deaths have come from F4/F5 tornadoes.

Possible explanations… “post analysis” ratings for the 1950s and 1960s were based more on fatalities than damage descriptions… ratings have trended more “conservative” as more has been learned about the relationship between wind and damage… impact of QRT and E-F scale?

Perception 2

The number of tornado related fatalities is decreasing due to improving warnings and technology.

MAYBE.

Not a clear answer.

The long term trend is unmistakable… tornado fatalities have decreased dramatically since around 1925.

In the shorter term, things are not as clear. Very little trend since 1975. Annual number of fatalities since 1966 averaging around 55.

Some possibilities:

1996-2005 a more active weather period, so higher number of fatalities in spite of better warnings?
Population has increased
Number of people living in mobile homes… Percentage of people living in mobile homes in the has increased sixfold in the Southeast U.S. from 1960 to 1990. Of course, it has increased as well over the last 16 years.

From Brooks and Doswell research…

“The rate of death among people living in mobile homes in the Southeast U.S. is relatively close to the pre-1925 values in the U.S.”

Perception 3

The false alarm rate for tornado warnings is do high that it could cause a “cry wolf” desensitization among the public.

Do recent trends support this perception?

Research finds that, on average, a county can expect just two tornado warnings per year (mean is 1.93) (based on data from Birmingham, Huntsville, Jackson, Little Rock, Fort Worth, and and Oklahoma City).

Breaking it down to AL/MS/TN/AR… mean number of tornado warnings per county is 2.64 each year.

75 percent of tornado warnings are false alarms.

A study by Barnes says that the false alarm/warned event concept is too simplistic. Making the public aware that a “near miss” occurred helps.

A study by Dow and Cutter looked and hurricane evacuations and false alarms… and found out that the public did not hesitate to evacuate even after recent false alarms.

Bottom line is that the cry wolf syndrome might not be as bad as we think… but more study is needed and we still need to improve the FAR.

Reading suggestion.. “False Alarms and Close Calls: A Conceptual Model of Warning Accuracy” by Barnes..2006…

More to come…

Category: Uncategorized