Thunderstorm Outbreak Produces F-5 Tornado in MD

First Posting | Summary
Posting Date: May 3, 2002, 10:30:00 AM

On Saturday and Sunday, April 27-28, 2002, a large outbreak of severe thunderstorms occurred from Kansas to Maryland, as a vigorous upper-level system moved from the Midwest to the East. Warm moist air from the Gulf of Mexico clashed with cold, dry air from Canada to produce damaging tornadoes, hail, and high winds. Information released by the Storm Prediction Center of the National Weather Service (NWS) includes 36 reports of tornadoes, 291 of hailstorms, and 188 of straight-line windstorms over the course of 48 hours on April 27-28. The exact number of discrete events behind these reports will not be known until the data has been analyzed further. An F3 tornado struck in Union and Johnson counties in Illinois on Saturday, April 27, and an F2 tornado struck in southeastern Johnson County. The NWS has confirmed that five tornadoes touched down in Western Pennsylvania Sunday afternoon. Of the five, two struck in Mercer County, two in Butler County, and the fifth and strongest, rated F2, in Indiana County. Another tornado touched down briefly in Marshall County, West Virginia. Preliminary assessment of a tornado in Rutherford County, Tennessee, suggests it was an F3 with a path 3.2 miles long and 0.2 miles wide. Another F2 tornado struck in Shenandoah County, Virginia. The most severe microevent associated with this system occurred in Maryland. The strongest tornado on record for that state struck La Plata, MD, on Sunday night, April 28, 2002, at about 7:10 p.m., EDT (21:10 GMT). The tornado touched down between the towns of Risen and Marbury in Maryland and followed a 24-mile path across Charles County, passing near La Plata, the county seat, before continuing on for another 6 miles in Calvert County. An initial damage survey, undertaken by the NWS office in Baltimore, indicated wind speeds greater than 260 mph (418 kph), resulting in a rating of F5 on the Fujita tornado intensity scale. Damage corresponding to F5 tornado intensity was observed across a path width of about one quarter of a mile and a length of about five miles. This most severe damage lay within a broader swath: the path width corresponding to F1/F2 damage was estimated at about one mile. The tornado lasted about 45 minutes, according to officials at the NWS. According to the National Oceanic and Atmospheric Administration (NOAA), about 2% of all tornadoes nationally fall into the F4-F5 category. Since 1950, there have been about 51 F5 tornadoes in the U.S., but never before has one struck so far east. Maryland has had two F4 tornadoes previously. Damage estimates released by Charles County officials today were nearly twice the previous estimates. Some 194 businesses were damaged, 49 of which were completely destroyed. Most were in and around La Plata. More than 700 homes were damaged, 100 of which were destroyed. At least six homes were completely torn off their foundations. A 125-foot water tower was toppled and power was interrupted to some 25,000 homes at the height of the storm. More than 130 homes were destroyed or damaged in Calvert County. Four people are reported dead as a result of the storm, and over 100 people were injured. A state of emergency has been declared for Charles, Calvert, and Dorchester counties. The AIR severe thunderstorm modeling team has collected the available information on the overall thunderstorm system, or macroevent, which, in these early stages, consists primarily of the location of reports of individual microevents (tornadoes, hailstorms and straight-line windstorms) and, in the case of tornadoes, estimates of Fujita intensity. In general, exact wind speeds associated with localized events are not known, nor is hailstone size. Furthermore, there is duplication in these early reports and the data must be cleaned to avoid double-counting. For example, there may be two reports of hail in close physical and temporal proximity to each other; the likelihood is that this represents one hailstorm, not two. Estimating losses from severe thunderstorm events in real time is an extremely challenging task. Microevents are typically of short duration – sometimes lasting only minutes – and more often than not occur out of the range of weather stations or anemoeters. When no measured wind speeds are available, scientists and engineers must estimate wind speeds from observed damage. The data gathering and analysis process usually takes several months, after which the data enters the official SPC database. AIR has dispatched its own team of engineers to areas most severely affected to conduct a post-disaster field survey. We will be posting preliminary results from this survey over the next several days. Despite the difficulties surrounding the data in its current form, the AIR team has cleaned the data of what appear to be duplicate microevents, and has run the severe thunderstorm model using assumptions regarding model parameters. In the case of reported tornadoes for which F-scale is not yet available, written accounts of damage were mapped to a wind speed using Fujita-scale descriptions. Unless explicitly stated, tornado path length and width was assumed to be the average for that intensity for that state. Hail impact energy (which is a function of hailstone size, the number of hailstones per square meter, and wind speed) was again assumed to be the average for the state, and averages by state were used for straight-line wind speeds except where wind speeds were explicitly reported. As a result of their analysis, AIR estimates that expected losses are between $500 - $600 million. Of this total, over $200 million are attributable to the Maryland F5 tornado. In light of the current uncertainty surrounding the true parameters of this event, however, this estimate itself is subject to uncertainty. All clients can now view loss maps representing several possible scenarios on the Scenarios page of this web site. In addition, an event set is available for download by CATMAP/2 and CATRADER clients. CLASIC/1 and CLASIC/2 clients can contact the AIR Software Services group for assistance in running an analysis using these scenarios.

First Posting | Downloads
Posting Date: May 3, 2002, 10:30:00 AM