Decadal Strong Tornadoes

A Preliminary Analysis of Significant Tornadoes and Associated Casualties In the NWS Jackson, MS Service Area

Introduction

A rough statistical analysis has been completed looking at strong (F2 and greater) tornadoes, and associated casualties, by decade since 1916 for the NWS Jackson, MS service area. For data back to 1950, official NCDC/SPC Storm Data was utilized, using the SPC SvrPlot software. Prior to 1950, data from the book Significant Tornadoes by Thomas Grazulis was utilized to obtain the raw data. Obviously, the most recent data is likely the most accurate. This is due to the availability of Doppler radar, which makes detections of circulations capable of producing tornadoes much easier, as well as better communications and increased efforts in storm damage surveying.

Note that all of the statistics discussed below are specifically referring to tornadoes of F2 intensity or greater. Deaths and injuries associated with tornadoes less than F2 in strength are not included in this dataset.

Results

Figure 1

This first figure shows the number of strong tornadoes by decade (blue), and the corresponding number of killer tornadoes (purple). During the last 90 years, the most recent period from 1996 to 2005 was the third least active period with regard to strong tornadoes in the Jackson CWA with only 44 recorded. Only 1956-65 (38) and 1936-45 (35) had fewer strong tornadoes.

The period from 1966 to 1985 stands out as a particularly active timeframe, by far the most active during the 90 year period. Just over 200 strong tornadoes occurred during this period, with 1966-75 being the most active decade with 107. This is well in excess of the mean of 62 strong tornadoes per decade.

With regard to killer tornadoes, not surprisingly the number of killer tornadoes has shown a steady decline over this 90 year period. Even though not particularly active from terms of numbers of strong tornadoes, the twenty year period from 1916 to 1935 had by far the greatest number of killer tornadoes, with 61 reported. The most recent decade of 1996-2005 had the fewest number of killer tornadoes with just 4, and the twenty year period from 1986 to 2005 only had a total of 10 killer tornadoes, which is less than any other single ten year period since 1916 with the exception of 1956-65 (7).

Figure 2

Figure 2 shows the trend of strong tornadoes (purple) and total number of fatalities (blue) for the 90 year period. As one can see, there was a clear decline in the number of fatalities during the first five decades, decreasing from nearly 300 in 1916-25 to less than 20 in 1956-65. However, a pronounced spike in fatalities occurred in the 1966-75 decade, with nearly 250 tornado related fatalities reported. This increase corresponded with a dramatic increase in the numbers of strong tornadoes during this decade, going from 39 to 107. Approximately 150 of these fatalities occurred on two days during this decade - March 3, 1966, when an F5 tornado (the so-called Candlestick Park tornado) killed nearly 60 people in Hinds, Rankin, Scott, and Leake counties; and February 21, 1971, when three violent tornadoes during the Mississippi Delta Tornado Outbreak killed almost 90 persons.

While a similar number of strong tornadoes occurred between 1976 and 1985, associated fatalities were much less. Only 36 fatalities occurred during this time period. The trend in fewer fatalities continued up to the most recent decade, during which only 7 tornado related fatalities occurred in the Jackson CWA.

Interpretation and Speculations

The evolution of technology, socioeconomics, and communications during the past 90 years makes it very difficult to directly compare given decades during this timeframe. However, I will make some comments and speculations based on my analysis of the data and experience after a decade in the region.

The most obvious conclusion is that the improved science of meteorology, better communications, and improved emergency management and preparedness, has resulted in fewer casualties from tornadoes. The fact that fatalities have shown a consistent decrease since 1916, and that the most recent decade had 7 fatalities in comparison to an average of just over 100 per decade for the last 90 year period, would certainly indicate that our efforts are making a difference.

Having said that, it also seems clear that the small number of fatalities in the 1996-2005 period is not just attributable to better warnings and communications, but also to a quite inactive severe weather period. In fact, my estimate is that 1996-2005 was the most inactive decade in the Jackson CWA with regard to significant tornadoes during the past 90 years, with the possible exception of 1956-65.

There are several factors which lead me to this conclusion. First, 1996-2005 was the first decade with Doppler radar and a very active NWS storm damage/verification program. There were a number (between 5 and 10) of strong tornadoes that were documented by NWS personnel based entirely on meteorologists observing a strong circulation on radar, and then going out and looking for damage in the following days. No reports of damage or tornadoes were received from law enforcement or the public with these storms since damage was almost exclusively limited to tree or other rural damage. Hence, in previous decades these tornadoes would almost certainly have went unreported. Removing these tornadoes from the database for 1996-2005 would yield similar numbers for strong tornadoes to the decades of 1956-65 and 1936-45. The fact that 1936-45 had a much larger number of fatalities (101) than the other decades, combined with the fact that the Depression and World War II periods probably resulted in even less effective reporting and recording of tornadoes, leads me to believe that it was a more active period than 1996-2005 or 1956-65.

Another factor arguing for viewing the 1996-2005 decade as a particularly inactive tornado period is the fact that such a large number of the strong tornadoes actually occurred at the very end of the period. Of the 46 strong tornadoes during this decade, nearly half (22) occurred in the last fourteen months of the period. Hence, it seems almost certain that one could characterize the period from 1996 to 2004 as the most inactive period with regard to significant tornadoes in the Jackson CWA back to at least 1916.

Conversely, 1966 to 1975 can be regarded as an aberration in the other direction, particularly with the number of casualties. The total number of strong tornadoes during this decade (107) is nearly twice the long term average (62), and the 7 violent (F4-F5) tornadoes during this period is equaled only by the decade immediately following, 1976-1985. The number of fatalities during this decade (225) is only exceeded by the 1916-1925 period, and the injury total of nearly 3000 is by far the largest of any decade (the next closest is 1679 in 1916-25).

An interesting contrast can be seen when comparing the statistics from the 1966-75 decade with the following decade of 1976-85. The number of strong tornadoes during these decades is very similar (107; 99), and the number of violent tornadoes is identical (7). However, the 1966-75 period had fatality counts (225; 36) and injury counts (2928; 751) vastly larger than the 1976-85 period. While certainly continued improvements to the science of meteorology, communications, and preparedness played a significant part in this difference, one must acknowledge the factor of chance and where tornadoes hit. During the 1966-75 time period, unfortunately, strong and violent tornadoes frequently made direct impacts on significant population areas (e.g., Candlestick Park tornado moving through Jackson metro area; February 1971 outbreak moving through Inverness, Delhi, and Morgan City) during their peak intensity. Such catastrophic overlays of location and peak intensity did not occur as frequently during the 1976-85 period.

Table 1. Tornadoes by decade.

Conclusion

One thing that seems clear from this data is that in spite of some notable tornado events in the last decade (November 24, 2001 outbreak; Hurricane Rita outbreak; Veteran's Day 2002 outbreak; Brandon, MS tornadoes), this time period has been notably unremarkable with regard to strong/violent tornadoes and associated casualties across central Mississippi, northeast Louisiana, and southeast Arkansas, at least in comparison to normal. We in the NWS, as well as our partners in emergency management and media, need to be aware of this fact as we perform our education and outreach efforts. All of us need to stress the fact that whether it starts this year or a decade from now, a more active period of strong/violent tornadoes is likely to occur in the future. Like our counterparts in coastal communities who have not seen a destructive hurricane in many years, we need to work to try to keep awareness of what sort of damage and casualties, normal, and above normal periods of tornado production can bring to our region.