Loading...
Please wait, while we are loading the content...
Similar Documents
Reconstructing the Frequency of Tornado Occurrence in the Central United States
| Content Provider | Semantic Scholar |
|---|---|
| Author | Menne, Matthew J. |
| Copyright Year | 2005 |
| Abstract | The official record of tornado occurrences, published monthly in the National Oceanic and Atmospheric Administration (NOAA) periodical Storm Data, is based on a log of thunderstorm and tornado events collected by the NOAA Storm Prediction Center (SPC) and NOAA National Severe Storms Laboratory (NSSL). The annual frequency of occurrence of tornadoes of all intensity classes reported in this archive exhibits an almost linear increase from the beginning of the record through the 1990s, as shown in Fig. 1. Annual totals rise from an average of about 200 reports in 1950 to more than 1000 by the early 1990s. An even more dramatic rise in the number of severe thunderstorm reports is indicated by the archive (Fig. 2). On the other hand, the number of strong (F2 or F3 on the Fujitaor F-scale) to violent (F4or F5-rated) tornadoes recorded in the storm database shows a marked decrease since the late 1970s. Along with any natural variability, the official tornado record reflects changes in public awareness of tornadoes, improvements in detection capabilities and changes in the procedures used to classify storms (McNulty, 1979; Schaefer and Galway, 1982; Doswell and Burgess, 1988; Schaefer and Edwards, 1999; Doswell et al., 1999). These changes have led to biases in the storm log that confound efforts to quantify the true climatological frequency of tornado occurrence as well as any temporal variations. For example, the reduction in the recorded number of strong-to-violent tornadoes since the mid-1970s is likely a consequence, at least in part, of a systematic change in the procedures used to estimate tornado intensity (Brooks and Craven, 2002). Ratings for tornadoes prior to 1976 were assessed retrospectively using published newspaper articles and photographs whereas since the late 1970's most ratings have been made by local National Weather Service personnel via on-site surveys (Kelly et al., 1978; Schaefer and Edwards, 1999). The dramatic rise in severe thunderstorm reports shown in Fig. 2 coincides with the National Weather Service's establishment of requirements for the verification of severe weather warnings (Schaefer and Edwards, 1999). Schaefer and Livingston (2003), in fact, distinguish three periods of storm reporting and survey practices during the period of record covered by the SPC storm log: the “newspaper era” (1950 to 1976), the “state office era” (1977 to 1996) and the “NWS modernized era” (1996 to present). They acknowledge, however, that the change from one set of damage rating procedures to another occurred over the course of a few years and that many elements of modernization were in place before 1996. Given the presence of bias in the observed tornado record, a number of studies have focused on estimating a truer mean annual frequency or climatological “risk” of tornadoes (e.g., Kelly et al., 1978; Schaefer et al., 1986, Concannon et al., 2000; Brooks and Doswell, 2001). However, using the official storm archive exclusively does not afford an evaluation of interannual variability in tornado occurrence or an assessment of low frequency variations. Whether it is likely that at least some of the changes implied by the official record of tornado occurrence are real remains essentially unknown. Consequently, it has been impossible to compare existing studies quantifying variations in numerous atmospheric variables to possible changes in tornado frequency. Evidence of climate change and variability in the U.S. during recent decades includes, for example, increases in surface specific humidity and dew point (Gaffen and Ross, 1999), a rise in surface temperature (e.g., Karl et al., 1993), increases in the frequency of heavy precipitation events (e.g., Kunkel et al., 2003) and a general moistening of the lower troposphere (Elliott and Angell, 1997). Since air parcel-based calculations used to quantify the potential for deep moist convection are sensitive to changes in near surface moisture conditions, it is possible that tornadic thunderstorm frequency may be sensitive to changes in near surface conditions. To determine how useful the tornado log may be in climate and global change analysis and to shed light on the nature of low frequency variations in storm occurrence, a derived frequency of environments favorable to tornado development is compared in this poster to the rate of occurrence implied by the official tornado log. In this study, a reasonably comprehensive set of sounding-based severe weather parameters was used to assess the potential of differentiating tornadic from other types of storm environments. In addition, the goal was to calculate a time series of environments favorable to tornado development that is largely independent of the storm log and thus free of changes in bias that are associated with changes in the reporting practices of severe local storm events. Logistic regression is used to estimate the probably that a sounding is indicative of severe tornadic conditions given a particular configuration of parameter values. The ability of these regression models to discriminate between tornadic proximity soundings and other storm proximity soundings is quantified using skill scores appropriate for the assessment of the prediction of rare events. |
| File Format | PDF HTM / HTML |
| Alternate Webpage(s) | https://ams.confex.com/ams/pdfpapers/100692.pdf |
| Language | English |
| Access Restriction | Open |
| Content Type | Text |
| Resource Type | Article |
