Tornadoes and Climate Change: Huge Stakes, Huge Unknowns

In 2011, a series of violent severe storms swept across the Plains and Southeast U.S., bringing an astonishing six billion-dollar disasters in a three-month period. The epic tornado onslaught killed 552 people, caused $25 billion in damage, and brought three of the five largest tornado outbreaks since record keeping began in 1950. In May 2011, the Joplin, Missouri tornado did $3 billion in damage--the most expensive tornado in world history--and killed 158 people, the largest death toll from a U.S. tornado since 1947. An astounding 1050 EF-1 and stronger tornadoes ripped though the U.S. for the one-year period ending that month. This was the greatest 12-month total for these stronger tornadoes in the historical record, and an event so rare that we might expect it to occur only once every 62,500 years. Fast forward now to May 2012 - April 2013. Top-ten coldest temperatures on record across the Midwest during March and April of 2013, coming after a summer of near-record heat and drought in 2012, brought about a remarkable reversal in our tornado tally--the lowest 12-month total of EF-1 and stronger tornadoes on record--just 197. This was an event so rare we might expect it to occur only once every 3,000 - 4,000 years. And now, in May 2013, after another shattering EF-5 tornado in Moore, Oklahoma, residents of the Midwest must be wondering, are we back to the 2011 pattern? Which of these extremes is climate change most likely to bring about? Is climate change already affecting these storms? These are hugely important questions, but ones we don't have good answers for. Climate change is significantly impacting the environment that storms form in, giving them more moisture and energy to draw upon, and altering large-scale jet stream patterns. We should expect that this will potentially cause major changes in tornadoes and severe thunderstorms. Unfortunately, tornadoes and severe thunderstorms are the extreme weather phenomena we have the least understanding on with respect to climate change. We don't have a good enough database to determine how tornadoes may have changed in recent decades, and our computer models are currently not able to tell us if tornadoes are more likely to increase or decrease in a future warmer climate.


Video 1. Remarkable video of the tornado that hit Tuscaloosa, Alabama on April 27, 2011, part of the largest and most expensive tornado outbreak in U.S. history--the $10.2 billion dollar Southeast U.S. Super Outbreak of April 25 - 28, 2011. With damage estimated at $2.2 billion, the Tuscaloosa tornado was the 2nd most expensive tornado in world history, behind the 2011 Joplin, Missouri tornado. Fast forward to minute four to see the worst of the storm.


Figure 1. Will climate change increase the incidence of these sorts of frightening radar images? Multiple hook echoes from at least ten supercell thunderstorms cover Mississippi, Alabama, and Tennessee in this radar image taken during the height of the April 27, 2011 Super Outbreak, the largest and most expensive tornado outbreak in U.S. history. A multi-hour animation is available here.

Changes in past tornado activity difficult to assess due to a poor database
It's tough to tell if tornadoes may have changed due to a changing climate, since the tornado database is of poor quality for climate research. We cannot measure the wind speeds of a tornado directly, except in very rare cases when researchers happen to be present with sophisticated research equipment. A tornado has to run over a building and cause damage before an intensity rating can be assigned. The National Weather Service did not begin doing systematic tornado damage surveys until 1976, so all tornadoes from 1950 - 1975 were assigned a rating on the Fujita Scale (F-scale) based on old newspaper accounts and photos. An improved Enhanced Fujita (EF) scale to rate tornadoes was adopted in 2007. The transition to the new EF scale still allows valid comparisons of tornadoes rated, for example, EF-5 on the new scale and F5 on the old scale, but does create some problems for tornado researchers studying long-term changes in tornado activity. More problematic is the major changes in the Fujita-scale rating process that occurred in the mid-1970s (when damage surveys began), and again in 2001, when scientists began rating tornadoes lower because of engineering concerns and unintended consequences of National Weather Service policy changes. According to Brooks (2013), "Tornadoes in the early part of the official National Weather Service record (1950 - approximately 1975) are rated with higher ratings than the 1975 - 2000 period, which, in turn, had higher ratings than 2001 - 2007." All of these factors cause considerable uncertainty when attempting to assess if tornadoes are changing over time. At a first glance, it appears that tornado frequency has increased in recent decades (Figure 2). However, this increase may be entirely caused by factors unrelated to climate change:

1) Population growth has resulted in more tornadoes being reported. Heightened awareness of tornadoes has also helped; the 1996 Hollywood blockbuster movie Twister "played no small part" in a boom in reported tornadoes, according to tornado scientist Dr. Nikolai Dotzek.

2) Advances in weather radar, particularly the deployment of about 100 Doppler radars across the U.S. in the mid-1990s, has resulted in a much higher tornado detection rate.

3) Tornado damage surveys have grown more sophisticated over the years. For example, we now commonly classify multiple tornadoes along a damage path that might have been attributed to just one twister in the past.


Figure 2. The total number of U.S. tornadoes since 1950 has shown a substantial increase. Image credit: NOAA/NCDC.


Figure 3. The number of EF-0 (blue line) and EF-1 and stronger tornadoes (maroon squares) reported in the U.S. since 1950. The rise in number of tornadoes in recent decades is seen to be primarily in the weakest EF-0 twisters. As far as we can tell (which isn't very well, since the historical database of tornadoes is of poor quality), there is not a decades-long increasing trend in the numbers of tornadoes stronger than EF-0. Since these stronger tornadoes are the ones most likely to be detected, this implies that climate change, as yet, is not having a noticeable impact on U.S. tornadoes. Image credit: Kunkel, Kenneth E., et al., 2013, "Monitoring and Understanding Trends in Extreme Storms: State of Knowledge," Bull. Amer. Meteor. Soc., 94, 499–514, doi: http://dx.doi.org/10.1175/BAMS-D-11-00262.1


Figure 4. Insured damage losses in the U.S. due to thunderstorms and tornadoes, as compiled by Munich Re. Damages have increased sharply in the past decade, but not enough to say that an increase in tornadoes and severe thunderstorms may be to blame.

Stronger tornadoes do not appear to be increasing
Tornadoes stronger than EF-0 on the Enhanced Fujita Scale (or F0 on the pre-2007 Fujita Scale) are more likely to get counted, since they tend to cause significant damage along a long track. Thus, the climatology of these tornadoes may offer a clue as to how climate change may be affecting severe weather. If the number of strong tornadoes has actually remained constant over the years, we should expect to see some increase in these twisters over the decades, since more buildings have been erected in the paths of tornadoes. However, if we look at the statistics of U.S. tornadoes stronger than EF-0 or F-0 since 1950, there does not appear to be any increase in their number (Figure 3.) Damages from thunderstorms and tornadoes have shown a significant increase in recent decades (Figure 4), but looking at damages is a poor way to determine if climate change is affecting severe weather, since there are so many human factors involved. A study in Environmental Hazards (Simmons et al., 2012) found no increase in tornado damages from 1950 - 2011, after normalizing the data for increases in wealth and property. Also, Bouwer (BAMS, 2010) reviewed 22 disaster loss studies world-wide, published 2001 - 2010; in all 22 studies, increases in wealth and population were the "most important drivers for growing disaster losses." His conclusion: human-caused climate change "so far has not had a significant impact on losses from natural disasters." Studies that normalize disaster data are prone to error, as revealed by a 2012 study looking at storm surge heights and damages. Given the high amount of uncertainty in the tornado and tornado damage databases, the conclusion of the "official word" on climate science, the 2007 United Nations IPCC report, pretty much sums things up: "There is insufficient evidence to determine whether trends exist in small scale phenomena such as tornadoes, hail, lighting, and dust storms." Until a technology is developed that can reliably detect all tornadoes, there is no hope of determining how tornadoes might be changing in response to a changing climate. According to Doswell (2007): "I see no near-term solution to the problem of detecting detailed spatial and temporal trends in the occurrence of tornadoes by using the observed data in its current form or in any form likely to evolve in the near future."


Figure 5. Wind shear from the surface to 6 km altitude in May on days with days with higher risk conditions for severe weather (upper-10% instability and wind shear) over the South Central U.S. has shown no significant change between 1950 - 2010. Image credit: Brooks, 2013, "The spatial distribution of severe thunderstorm and tornado environments from global reanalysis data", Atmospheric Research Volumes 67-68, July-September 2003, Pages 73-94.


Figure 6. Six-hourly periods per year with environments supportive of significant severe thunderstorms in the U.S. east of the Rocky Mountains. The line is a local least-squares regression fit to the series, and shows no significant change in environments supportive of significant severe thunderstorms in recent decades. Image credit: Brooks, H.E., and N. Dotek, 2008, "The spatial distribution of severe convective storms and an analysis of their secular changes", Climate Extremes and Society

How are the background conditions that spawn tornadoes changing?
An alternate technique to study how climate change may be affecting tornadoes is look at how the large-scale environmental conditions favorable for tornado formation have changed through time. Moisture, instability, lift, and wind shear are needed for tornadic thunderstorms to form. The exact mix required varies considerably depending upon the situation, and is not well understood. However, Brooks (2003) attempted to develop a climatology of weather conditions conducive for tornado formation by looking at atmospheric instability (as measured by the Convective Available Potential Energy, or CAPE), and the amount of wind shear between the surface and 6 km altitude. High values of CAPE and surface to 6 km wind shear are conducive to formation of tornadic thunderstorms. The regions they analyzed with high CAPE and high shear for the period 1997-1999 did correspond pretty well with regions where significant (F2 and stronger) tornadoes occurred. Riemann-Campe et al. (2009) found that globally, CAPE increased significantly between 1958 - 2001. However, little change in CAPE was found over the Central and Eastern U.S. during spring and summer during the most recent period they studied, 1979 - 2001. Brooks (2013) found no significant trends in wind shear over the U.S. from 1950 - 2010 (Figure 5.) A preliminary report issued by NOAA’s Climate Attribution Rapid Response Team in July 2011 found no trends in CAPE or wind shear over the lower Mississippi Valley over the past 30 years.


Figure 7. Change in the number of days per year with a high severe thunderstorm potential as predicted by the climate model (A2 scenario) of Trapp et al. 2007, due to predicted changes in wind shear and Convective Available Potential Energy (CAPE). Most of the U.S. east of the Rocky Mountains is expected to see 1 - 2 additional days per year with the potential for severe thunderstorms. The greatest increase in potential severe thunderstorm days (three) is expected along the North and South Carolina coast. Image credit: NASA.

How will tornadoes and severe thunderstorms change in the future?
Using a high-resolution regional climate model (25 km grid size) zoomed in on the U.S., Trapp et al. (2007) and Trapp et al. (2009) found that the decrease in 0-6 km wind shear in the late 21st century would more than be made up for by an increase in instability (CAPE). Their model predicted an increase in the number of days with high severe storm potential for most of the U.S. by the end of the 21st century, particularly for locations east of the Rocky Mountains (Figure 7.) Brooks (2013) also found that severe thunderstorms would likely increase over the U.S. by the end of the century, but theorized that the severe thunderstorms of the future might have a higher proportion causing straight-line wind damage, and slightly lower proportion spawning tornadoes and large hail. For example, a plausible typical future severe thunderstorm day many decades from now might have wind shear lower by 1 m/s, but a 2 m/s increase in maximum thunderstorm updraft speed. This might cause a 5% reduction in the fraction of severe thunderstorms spawning tornadoes, but a 5% increase in the fraction of severe thunderstorms with damaging straight-line winds. He comments: "However, if the number of overall favorable environments increases, there may be little change, if any, in the number of tornadoes or hailstorms in the US, even if the relative fraction decreases. The signals in the climate models and our physical understanding of the details of storm-scale processes are sufficiently limited to make it extremely hazardous to make predictions of large changes or to focus on small regions. Projected changes would be well within error estimates."


Figure 8. From 1995 (the first year we have wind death data) through 2012, deaths from high winds associated with severe thunderstorms accounted for 8% of all U.S. weather fatalities, while tornadoes accounted for 13%. Data from NOAA.

Severe thunderstorms are capable of killing more people than tornadoes
If the future climate does cause fewer tornadoes but more severe thunderstorms, this may not end up reducing the overall deaths and damages from these dangerous weather phenomena. In 2012, the warmest year in U.S. history, the death toll from severe thunderstorms hit 104--higher than the 70 people killed by tornadoes that year. Severe thunderstorms occur preferentially during the hottest months of the year, June July and August, and are energized by record heat. For example, wunderground weather historian Christopher C. Burt called the number of all-time heat records set on June 29, 2012 “especially extraordinary,” and on that day, an organized thunderstorm complex called a derecho swept across a 700-mile swath of the Ohio Valley and Mid-Atlantic, killing thirteen people and causing more than $1 billion in damage. The amount of energy available to the derecho was extreme, due to the record heat. The derecho knocked out power to 4 million people for up to a week, in areas where the record heat wave was causing high heat stress. Heat claimed 34 lives in areas without power in the week following the derecho. Excessive heat has been the number one cause of weather-related deaths in the U.S since 1995, killing more than twice as many people as tornadoes have. Climate models are not detailed enough to predict how organized severe thunderstorm events such as derechos might change in a future warmer climate. But a warmer atmosphere certainly contributed to the intensity of the 2012 derecho, and we will be seeing a lot more summers like 2012 in coming decades. A future with sharply increased damages and deaths due to more intense severe thunderstorms and derechos is one nasty climate change surprise that may lurk ahead.


Figure 9. Lightning over Tucson, Arizona on August 14, 2012. A modeling study by Del Genio et al.(2007) predicts that lighting will increase by 6% by the end of the century, potentially leading to an increase in lightning-triggered wildfires. Image credit: wunderphotographer ChandlerMike.

Lightning may increase in a warmer climate
Del Genio et al. (2007) used a climate model with doubled CO2 to show that a warming climate would make the atmosphere more unstable (higher CAPE) and thus prone to more severe weather. However, decreases in wind shear offset this effect, resulting in little change in the amount of severe weather in the Central and Eastern U.S. late this century. However, they found that there would likely be an increase in the very strongest thunderstorms. The speed of updrafts in thunderstorms over land increased by about 1 m/s in their simulation, since upward moving air needed to travel 50 - 70 mb higher to reach the freezing level, resulting in stronger thunderstorms. In the Western U.S., the simulation showed that drying led lead to fewer thunderstorms overall, but the strongest thunderstorms increased in number by 26%, leading to a 6% increase in the total amount of lighting hitting the ground each year. If these results are correct, we might expect more lightning-caused fires in the Western U.S. late this century, due to increased drying and more lightning. Only 12% of U.S. wildfires are ignited by natural causes, but these account for 52% of the acres burned (U.S. Fire Administration, 2000). So, even a small change in lightning flash rate has important consequences. Lightning is also a major killer, as an average of 52 people per year were killed by lightning strikes over the 30-year period ending in 2012, accounting for 6% of all U.S. weather-related fatalities.

Summary
We currently do not know how tornadoes and severe thunderstorms may be changing due to climate change, nor is there hope that we will be able to do so in the foreseeable future. It does not appear that there has been an increase in U.S. tornadoes stronger than EF-0 in recent decades, but climate change appears to be causing more extreme years--both high and low--of late. Tornado researcher Dr. Harold Brooks of the National Severe Storms Laboratory in Norman, Oklahoma said in a 2013 interview on Andrew Revkin's New York Times dotearth blog: "there’s evidence to suggest that we have seen an increase in the variability of tornado occurrence in the U.S." Preliminary research using climate models suggests that we may see an increase in the number of severe thunderstorms capable of producing tornadoes over the U.S. late this century, but these thunderstorms will be more likely to produce damaging straight-line winds, and less likely to produce tornadoes and large hail. This will not necessarily result in a reduction in deaths and damages, though, since severe thunderstorms can be just as dangerous and deadly as tornadoes--especially when they knock out power to areas suffering high-stress heat waves. Research into climate change impacts on severe weather is just beginning, and much more study is needed.

Other analyses of tornadoes and climate change
Making Sense of the Moore Tornado in a Climate Context by Andrew Freedman of climatecentral.org

Tornadoes, Extreme Weather And Climate Change, Revisited by Joe Romm at climateprogress.org

Seeking Clarity on Terrible Tornadoes in a Changing Climate by Andrew Revkin of the New York Times has links to four other excellent blogs on the subject.


Video 2. Dr. Harold Brooks of the National Severe Storms Laboratory in Norman, Okla., gave a video interview at a tornado and climate research conference held at Columbia University earlier this year. He discusses why we don't issue seasonal tornado forecasts, but doesn't discuss climate change.

References
Brooks, H.E., 2013, "Severe thunderstorms and climate change," Atmospheric Research, Volume 123, 1 April 2013, Pages 129–138, http://dx.doi.org/10.1016/j.atmosres.2012.04.002.

Brooks, H.E., J.W. Lee, and J.P. Craven, 2003, "The spatial distribution of severe thunderstorm and tornado environments from global reanalysis data", Atmospheric Research Volumes 67-68, July-September 2003, Pages 73-94.

Brooks, H.E., and N. Dotek, 2008, "The spatial distribution of severe convective storms and an analysis of their secular changes", Climate Extremes and Society, edited by Henry F. Diaz, Richard J. Murnane,

Doswell, C.A., 2007, "Small Sample Size and Data Quality Issues Illustrated Using Tornado Occurrence Data", E-Journal of Severe Storms Meteorology Vol 2, No. 5 (2007).

Del Genio, A.D., M-S Yao, and J. Jonas, 2007, Will moist convection be stronger in a warmer climate?, Geophysical Research Letters, 34, L16703, doi: 10.1029/2007GL030525.

Kunkel, Kenneth E., et al., 2013, "Monitoring and Understanding Trends in Extreme Storms: State of Knowledge," Bull. Amer. Meteor. Soc., 94, 499–514, doi: http://dx.doi.org/10.1175/BAMS-D-11-00262.1

Marsh, P.T., H.E. Brooks, and D.J. Karoly, 2007, Assessment of the severe weather environment in North America simulated by a global climate model, Atmospheric Science Letters, 8, 100-106, doi: 10.1002/asl.159.

Riemann-Campe, K., Fraedrich, K., and F. Lunkeit, 2009, Global climatology of Convective Available Potential Energy (CAPE) and Convective Inhibition (CIN) in ERA-40 reanalysis, Atmospheric Research Volume 93, Issues 1-3, July 2009, Pages 534-545, 4th European Conference on Severe Storms.

Simmons, K.M., Dutter, D., and Pielke, R., 2012, "Normalized Tornado Damage in the United States: 1950-2011," DOI: 10.1080/17477891.2012.738642

Trapp, R.J., N.S. Diffenbaugh, H.E. Brooks, M.E. Baldwin, E.D. Robinson, and J.S. Pal, 2007, Severe thunderstorm environment frequency during the 21st century caused by anthropogenically enhanced global radiative forcing, PNAS 104 no. 50, 19719-19723, Dec. 11, 2007.

Trapp, R. J., Diffenbaugh, N. S., & Gluhovsky, A., 2009, "Transient response of severe thunderstorm forcing to elevated greenhouse gas concentrations," Geophysical Research Letters, 36(1).

U.S. Fire Administration, “2000 wildland fire season,” U.S. Fire Administration Topical Fire Research Series, vol. 1, Issue 2, 4 pp.

Jeff Masters