There have been several efforts made, over the past few years, to determine the existence of statistical and/or causative links between specific phases of the El Niņo Southern Oscillation (ENSO) and United States tornado climatology. In many of these studies, tornado climatology during the entire year was considered. However, this paper will consider only tornadoes occurring during the cold season, defined as November, December, January, and February. In addition, focus for this study was not on individual tornado events but rather on the frequency of major tornado outbreaks during what is not generally considered to be a favorable time of year for such events to occur. Tornado outbreak climatology over a 45 year period was studied, for an area covering much of the Southeastern and Midwestern United States. The main goal of this study was to examine whether or not certain ENSO cycles may be especially conducive for synoptic-scale patterns which may, in turn, lead to major tornado outbreaks in an otherwise "unseasonable" time of year.

The study was focused around two main queries. First, is there a statistical basis for suggesting that certain phases of ENSO, whether El Niņo, La Niņa, or neutral, entail more favorable conditions for strong winter-time tornado outbreaks than others? Second, if statistics point towards a particular ENSO phase as being more favorable for numerous cold-season tornado outbreaks, does this suggested favorability agree with the actual favorability of the conditions seen during that phase? To approach the first question, cold-season tornado outbreak data in the study area was complied for the period 1950 - 1995. The outbreaks included were those that involved three or more significant (F2-F5) tornadoes. The outbreak data was then broken down into three geographic regions (Ohio Valley, Mississippi Valley, and Deep South), and each region's recorded outbreaks divided by the ENSO phase of the season in which each outbreak occurred. Assignment of ENSO phase to each season was based upon the average Southern Oscillation Index (SOI) for the four months of each season. Chi-squared tests were then run for each region, for the purpose of determining whether cold-season tornado outbreak frequency deviated with respect to ENSO phase. An average outbreak count for the period was determined, and used as the expected parameter in the tests. The chi-squared tests revealed that the greatest deviations from normal levels of cold-season outbreaks were seen during La Niņa seasons, these deviations being towards greater than normal outbreak numbers in all regions. Chi-squared values of less than one were calculated for all three regions when El Niņo season outbreaks were tested, suggesting that little or no deviations from expected numbers of outbreaks may be seen during the El Niņo phase.

When the observed meteorological patterns associated with each ENSO phase were taken into account, the relationship between phase and favorable tornado outbreak conditions started to become evident. La Niņa seasons, identified by the chi-squared tests as having seen the greatest deviations from normal numbers of outbreaks, have normally been associated with atmospheric teleconnections such as warmer than average temperatures in the Southern U.S. (and thus a sharper baroclinic zone between this warmer air and the normally cold arctic air), and an intensified polar jet over mid-latitude North America. These conditions are two very basic prerequisites for strong mid-latitude cyclones which, in turn, may be capable of producing strong tornado outbreaks. The La Niņa phase has been cited in previous studies as an especially active phase for tornadoes, and the data from this study supports that theory. While strong cold-season outbreaks have occurred during seasons that were not recognized as La Niņa, relatively high frequency of cold-season outbreaks appears to be a distinction of the La Niņa phase.