Addressing a few tornado questions
Posted May 2, 2011
Given the tornado outbreak we experienced on April 16th, and then the historic outbreak to our southwest around April 27th, we've had a number of questions pop up about tornadoes, their history and how they are categorized on our Facebook page and/or in e-mails and AskGreg submissions. A number of those will make their way into the AskGreg section over time, but I thought I'd touch on a few of them here.
The rating and survey of tornadoes from the most recent outbreak continues, and it may be some time before the numbers are finalized. The number of tornadoes with the outbreak is still a moving target, as some maps from NOAA suggest a total of 40-50 but NOAA also states it tracked 211 tornadoes on April 27th, but it is unclear whether some duplication of reports is involved with that figure, and it could change notably as surveys and analysis continues. That compares to 148 tornadoes in the 1974 Super Outbreak, which also killed over 300 people, but a number of the recent tornadoes were long-tracked and also passed over a number of highly populated areas, as indicated by the paths of supercell rotation centers seen on the first image above, in which many such cells followed the classic southwest to northeast motion across the deep south. As of this morning, the toll from this swarm had unfortunately risen to 345, to surpass the 1974 event and become the second most deadly tornado outbreak in U.S. history. So far, an outbreak of as many as 9 tornadoes on March 18, 1925 remains the worst in terms of fatalities, with as many as 747 lives lost that day. That outbreak included the Tri-State Tornado, thought to have a path length of up to 234 miles.
One question that has arisen comes from the references to damage rating using the "enhanced Fujita" scale, the question being "why was the old Fujita scale replaced?" Well, since that scale was developed and implemented in the early 1970s, years of additional research have led to improved procedures for correlating wind speed with damage to a range of structures, and has allowed development of a more defined and consistent method for assigning ratings based on storm surveys. In the process, it was discovered that less wind than originally thought was capable of producing some of the most severe damage. The entire process was updated in 2007, with "EF" ratings used since then to distinguish survey results published since then from older data. For details on the new scale and how it compares to the old, see www.spc.noaa.gov/efscale.
We've also been asked if there is a resource to check how many historical tornadoes have occurred in a particular county. There is a very nice database online at the National Climatic Data Center that addresses this question and others regarding severe weather of many kinds. It is called the "Storm Events Database" and allows you to search a history of severe weather reports by state, and then filter the results according to county, storm type (tornadoes, hail, wind damage, etc), and within the tornado category you can also narrow it down to certain intensity ranges if desired. This service does a really nice job of answering the question (at least since 1950) of "how many tornadoes have affected my county?" A couple of cautions here - it is best to limit searches to either a county or to a smaller range of years or intensities. A search for all tornadoes in NC, for example, will probably fail because of the very large number of records involved. Also, if you search a certain day or week for tornadoes across all of NC, the list you see is a list of all county tornado segments for the period, not a list of every separate tornado in the state. So, individual tornadoes that track across multiple counties will be listed more than once in the results. Finally, there is a time lag for data to be quality-controlled and added to the database, which at this writing has been updated through January 2011. I've included a link to the database.
We also had someone ask if there had been a shift in "tornado alley" over time, and the answer is not really, along with the fact that there is no "official" set of tornado alley boundaries. It is a more informal term that typically refers to an axis from north Texas up through the great plains into Nebraska or thereabouts, where "significant" tornadoes having intensities of EF2 (or F2) and up tend to be most frequent. Year to year variability in the location of the most intense tornadoes, however, is great enough that the term doesn't have much scientific meaning, and tornadoes are spread over a large part of the central and eastern U.S. - to illustrate this, and to see if there was any apparent shift in distribution on a very broad scale, I plotted two maps of all F2 and higher tornadoes in the United States, the first covering the period 1950 through 1980, and the second 1980 through 2010. If you switch back and forth between the two, you'll see that the overall distribution of these "significant" tornadoes is quite similar through the two periods. with a notable concentration up the eastern side of the "tornado alley" region I mentioned above, along with a clustering in a band across Mississippi into northern Alabama as well as just south of the Great Lakes. You might notice that there are fewer tornadoes shown for the latter period - this may be a result of tornadoes from the 1950s and 1960s, especially, being overclassified in regards to intensity somewhat compared to those in more recent years which are subject to more rigorous post-storm surveys. Otherwise, there is reason to think that tornadoes are actually undercounted in the older period, due to lower population, less robust communication networks and less sophisticated or complete radar coverage. To illustrate, if I include F0 and F1 tornadoes on the maps, there are 35,100 total for the 1980-2010 period, but only 19,414 for the 1950-1980 stretch. The increase in tornadoes associated with greater likelihood of detection and documentation in more recent years has been referred to as "tornado inflation."