The F-Scale is a subjective evaluation by a meteorologist to classify the degree and severity of damage caused by a tornado or high wind event. Since meteorologists are generally not trained structural engineers, most F-scale evaluations are made without explicit reference to the strength and integrity of the original construction. The historic F-scale ratings are implicitly based upon Fujita's faith in the consistent quality of "relatively new suburban structures" (Fujita, 1971). However, experience has shown that this is not a good assumption because construction quality not only varies from community to community, but from builder to builder.

To examine the variability in F-scale estimations, the Storm Prediction Center's Tornado Database is analyzed. This Database contains information on over 42,800 tornadoes that occurred in the contiguous United States during the 52 year span 1950 - 2001. The percentage of tornadoes in various F-scale categories for each individual county (and independent city) is computed, not only for the entire period of the record, but also for several sub-periods.

The variation of tornado F-scale is marked. Individual states can be recognized without reference to state borders because of systematic changes in the proportion of tornadoes rated at various levels. The pattern is not one of urban verses rural construction, but rather one of a change in the individual F-scale raters. Sub-periods examined are: the pre-Fujita scale years when storms were rated based upon newspaper accounts; the NWS state forecast office era when all storms in each state were typically rated by a single individual; and the present time when there are over 110 NWS Warning Coordination Meteorologists that rate all tornadoes that occur near their office. Differences in the pattern are noted and discussed.

These results emphasize that F-scale values are neither objective measures, nor surrogates for a wind speed estimates. A methodology for determining more consistent F-scale estimates which is independent of individual evaluator bias is needed. Some suggestions as to ways to do this are presented.

Reference: Fujita, T.T., 1971:Proposed Characterization of Tornadoes and Hurricanes by Area and Intensity. SMRP Research Paper No. 91. University of Chicago, 42 pp.