Preliminary Assessment of Climate Factors Contributing to the Extreme 2011 Tornadoes
Draft Assessment - Updated: 08-July-2011
El Niño/La Niña and Tornadoes
The record setting tornado outbreaks during April 2011 took place within a global climate context of lingering La Niña conditions (cold phase of the Southern Oscillation). Below normal sea surface temperatures (SSTs) had developed in the prior year over the tropical central Pacific, and though peaking in winter, persisted into the spring. Even while the Niño3.4 SSTs recovered toward normal during Spring, the overall tropical pattern of SST anomalies continued to be an effective driver of anomalous weather patterns over the United States. This was confirmed by climate simulations using the NOAA Global Forecast System (GFS) model that was forced with the observed Spring 2011 global SST conditions. These simulations were consistent with the actual Climate Forecast System (CFS) forecast runs, and yielded a strengthened jetstream entering the Pacific Northwest and channeling into the central US during late winter and early Spring (not shown). Such a condition would have promoted more active and more frequent upper air troughs in the central and eastern US, such as those seen in the animation of daily 500 millibar heights. A quantification of the contribution of the La Niña conditions on the overall severe weather occurrences over the US during April is a matter requiring further research, but the qualitative evidence from modeling and other observational sources suggests that the tropical SST conditions played a role. To the extent that such global ocean conditions may have played a role in the occurrence of this extreme tornado outbreak, there may exist a long-lead capability to alert the public and decision makers of heightened risk for more violent tornado outbreaks over a season for a given geographical region.
What is the current understanding of La Niña effects on US tornado occurrences? There are suggestions that winter (January-March) tornado counts may change their preferred geographical location depending on the phase of the El Niño/Southern Oscillation (ENSO).11 However, that study's analysis covering 1950-2003 found that neither the frequency of tornado days nor of violent tornado days (days with 5 or more F2+ tornadoes) is affected systematically by the phase ENSO for the US as a whole. A separate investigation that used data for a shorter period of 1950-1992, but covering all seasons, argued that La Niña events increase tornadic activity in the Ohio River Valley and the Deep South during spring, and that La Niña facilitates large tornadic outbreaks and is associated with more destructive storms.12. It should be recognized, however, that small sample sizes do not permit a robust statistical appraisal for either of these studies. Furthermore, results on tornado counts through time are vulnerable to the appreciable trend in tornado statistics that are non-meteorological (see the adjoining Climate Change discussion). More research is required to determine the extent to which La Niña itself is a meaningful and useful early warning indicator for US severe weather.
Table 1: Notable Tornado Outbreaks - 1900-2011
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Time animation of longitude vs. time (Hovmöller) plots of normalized Sea Surface Temperature (SST) anomalies (i.e., change from normal); blue = colder than normal (e.g., La Niña); red = warmer than normal (e.g., El Niño); black dots represent the times of the events listed Table 1. Data Source: NOAA Extended Reconstructed Sea Surface Temperature (SST) V3b |
To illustrate some of challenges in inferring the relation between severe weather and the state of tropical SSTs historically, we explore a record of the most noteworthy tornado outbreaks from 1900-2011. We focus specifically on prominent events, either in terms of fatalities or tornado counts. Documentary evidence of deaths and, when available, tornado counts, were used to identify 20 severe spring (March-May) tornado outbreaks over the contiguous US, and these are listed in Table 1. Of course, it is such events that are of greatest public concern, for which human life most benefits from early warning, and for which society demands adequate explanation of cause and predictability.
It is reasonable to expect that such high impact events often involve an outbreak of numerous, intense tornadoes, and thus these cases may be less susceptible to the sampling problems mentioned above. It is noteworthy, however, that the infamous Tri-State event of 18 March 1925, which has the highest fatality count on record, appears to have involved relatively few tornadoes.
To help visualize the relationship between major tornadic events and ENSO, we have plotted the temporal occurrence of each of these 20 events upon an animation of the tropical Pacific SSTs for 1900-2011. The timing of the events is highlighted with black dots positioned at the center of ENSO monitoring region near 120W. La Niña (El Niño) events are denoted by the darker blue (red) color shades in the tropical Pacific zone from the International dateline (denoted by the vertical line) eastward to the South American coast (far right side). One can readily appreciate, from this long historical perspective, that the relation of ENSO with major destructive tornado events is complicated, and that such a visual inspection alone does not permit a strong statement on the nature of possible causal relationships. We note, for example, that none of the 9 major outbreaks that occurred from the 1925 Tri-State to the 1965 Palm Sunday outbreak were associated with contemporaneous La Niña conditions. On the other hand, 5 of the 8 major destructive events since the Super Tornado outbreak of 1974 have occurred in association with contemporaneous La Niña conditions.
More research is clearly required to better understand how ENSO and its multi-year life-cycle may influence the probability of major, destructive tornado outbreaks over the US. In this regard, we mention again the lingering effects on US climate in Spring owing to the tropical wide SST pattern, and that a simple index of ENSO in the eastern Pacific may not be a suitable indicator. Thus, the relation is likely more complicated than the simple contemporaneous state of the tropical Pacific SSTs as suggested in the above analysis, and some work has begun examining how the synoptic-scale environment is sensitive to the detailed evolutionary history of the ENSO cycle13. And, while it has long been known that the teleconnection patterns of atmospheric circulation anomalies connect the remote ENSO forcing region to US climate conditions in winter and spring, the physical mechanisms by which ENSO may pre-condition a large-scale thermodynamic or dynamical wind shear environment to which strong tornado events are most sensitive is not currently understood.
The body of knowledge regarding the possible role played by large-scale climate forcings in tornado outbreaks is rapidly evolving, and constitutes a field of study that must integrate existing expertise in meso-scale meteorology with expertise in global-scale climate dynamics. Likewise, the methods for conducting attribution science also continue to evolve, and advances on the tornado-climate linkages will require modeling capabilities beyond current tools.
This assessment attempts to summarize a current scientific understanding of the link between ENSO and tornadoes. The assessment is not in lieu of a more rigorous diagnosis, nor does it preempt the need for better quantification of the physical climate factors associated with tornado outbreaks. This must be the topic of future research.