The Effect of Climate Change on Tornado Frequency and Magnitude
Michael Pateman and Drew Vankat

Abstract
Our research project concerns the phenomena of tornadoes, their frequency and magnitude, and possible correlations between that and climate. Investigations will be made into the effect of events such as El Nino and La Nina in regards to tornado occurrence and strength. Research will be conducted with data from 1950 to 1999, and in several geographic regions: Texas, Nebraska, and Ohio. We hope this will give us both a broad overview of the topic, as well as more localized data showing what happens to locations regularly experiencing tornadoes, and those that lack much pronounced tornado activity. Relationships in our data could assist authorities in preparation for aid to families and businesses, as well as further strengthen the belief that we must reduce emission of harmful pollutants into the environment.
Results from data show a stronger correlation in tornado magnitudes by state
over the 1950-1999 time period, as well as in El-Nino and La-Nina years. This
suggests that those events affect our three states in similar manners.
Results pertaining to frequency are less conclusive, as p-values tend to either be split a slight tendency to show a significant difference in data sets and rank orders, as detailed in our report. This means that there is a lack of strong evidence to relate tornado frequency to El-Nino or La-Nina years. Variations in this generalized conclusion are detailed in the results and conclusion sections.
Introduction
Tornadoes have struck every U.S. state, including Alaska and Hawaii. But most
tornadoes form in a belt from Nebraska southward through central Texas known
as Tornado Alley and in the Southeast. Wind speeds in tornadoes can vary from
72 to almost 300 mph.
When the El Nino/Southern Oscillation peaked in 1997 and 1998, much talk was
given to the theory of global warming and its possible connection to an increase
in extreme weather. In some areas of the United States, tornadoes are and have
been a serious risk to property and lives. Are tornadoes a part of this web
of phenomena possibly related to the theory of global warming? Could a possible
increase or decrease (depending on the area) be an indicator of global warming?
Our hypothesis is that the increase in greenhouse gases plays a role in the formation of tornadoes. This means that we are experiencing more tornadoes and tornadoes of greater magnitude as a result of both global warming and strengthened cycles such as El Nino and La Nina. We are cautious, however, not to underestimate the advances in forecasting and detection technology which may have led to an increase especially in the number of tornadoes reported.
Definitions
What is El Nino?
El Nino is a disruption of the ocean-atmosphere system in the Tropical Pacific
having important consequences for weather and climate around the globe. These
consequences can include increased rainfall in some areas (which can lead to
flooding) and extreme drought in others.
What is La Nina?
La Nina is characterized by unusually cold ocean temperatures in the Equatorial
Pacific, compared to El Nino, which is characterized by unusually warm ocean
temperatures in the Equatorial Pacific.
What is a tornado?
A tornado is officially defined as an intense, rotating column of air extending
from the base of a thunderstorm cloud to the ground. Most tornadoes form under
the southwestern section of the thunderstorm cloud. Air moves very rapidly upward
around a tornado center. This distinguishes tornadoes from microbursts, which
often do tornado-like damage and are often mistaken for tornadoes. Microbursts,
on the other hand, features air blasting downward from thunderstorms. The large
hail that often precedes tornadoes forms as a result of the intense updraft
feeding the thunderstorm.
The United States is the world capital for tornadoes as conditions favorable for tornado development most often occur over the Plains during spring and summer. A typical tornado outbreak often features an intense upper-level disturbance moving across the Plains during spring. This disturbance provides the strong vertical wind shear that gives an updraft its twisting motion, turning a normal thunderstorm into a potentially tornado spawning supercell. Although, the United States has the most tornadoes of any nation in the world, tornadoes do occur in other locations such as Australia and Europe.
What are favorable conditions for tornado formation?
A convective cap is a layer of hot, dry air in the middle layers of the atmosphere above the surface. Often, temperatures increase with height in this layer and relative humidities are extremely low. As you can see in the graphic above, warm humid air in the lower layers of the atmosphere near the surface is heated by the sun, but is not allowed to rise and initiate clouds and precipitation because of the hot, dry air above it. As the air near the surface continues to heat up, it builds up an enormous amount of energy much the same way boiling water in a pot with a heavy lid on it would. If a triggering mechanism, such as a cold front or dryline, moves into the area, the convective cap may weaken enough to allow the heated, humid air near the surface to burst through the cap and initiate extremely violent convection. Supercells, along with intense tornadoes, often form as a result of this violent convection.
How powerful/dangerous do tornados get? (Tornado
Facts)
Tornado strength and intensity are measured in the Fujita Tornado Damage
Scale, that ranges from F0 (weakest) to F5 (most powerful).
Category F0: Light Damage (<73 mph); Some damage to chimneys; branches
broken off trees; shallow-rooted trees pushed over; sign boards damaged.
Category F1: Moderate Damage (73-112 mph); Peels surface off roofs;
mobile homes pushed off foundations or overturned; moving autos blown off road.
Category F2: Considerable Damage (113-157 mph); Roofs torn off frame
houses; mobile homes demolished; boxcars overturned; large trees snapped or
uprooted; light-object missiles generated; cars lifted off ground.
Category F3: Severe Damage (158- 206 mph); Roofs and some walls torn
off well-constructed houses, trains overturned; most trees in forest uprooted;
heavy cars lifted off ground and thrown.
Category F4: Devastating Damage (207- 260 mph); Well-constructed houses
leveled; structure with weak foundations blown off some distance; cars thrown
and large missiles generated.
Category F5: Incredible Damage (261- 318 mph); Strong frame houses lifted
off foundations and swept away; automobile sized missiles fly through the air
in excess of 100 meters (109 yds); trees debarked; incredible phenomena will
occur.
How do tornadoes form?
Many of the strongest tornadoes are produced in supercell thunderstorms. A supercell
storm consists of one very large cell with dimensions of up to 20km high and
20-50km in diameter. These cells have a much longer life than ordinary thunderstorms;
they can persist for many hours. Because of their size, supercells are quite
complex in nature. Wind being lifted in the updraft may start off from the south,
but then shift to the west at a higher elevation, causing a rotation to form.
The area of rotating air is called a mesocyclone, and this is where tornadoes
often form. Mesocyclones can become nearly vertical, and stretched lengthwise,
reducing girth, which increases wind speeds. As the mesocyclone extends downward
in the storm, part of it may protrude from the bottom of the cloud. If this
'funnel cloud' touches the ground, it is classified as a tornado.
Tornado Structure
As mentioned, tornadoes are a rotating column of air extending downward from
a cumulonimbus cloud. Weaker tornadoes may consist of only one vortex, but in
many larger tornadoes, several smaller 'suction vortexes' are present within
a main vortex. These suction vortexes are very concentrated, and often no larger
than 10 meters in diameter. Extremely low pressures are located within a tornado.
Some estimates are on the scale of 10 percent less than the immediate surrounding
area. Because of this steep pressure gradient, we can experience wind speeds
of up to 300 miles an hour.

Figure - Tornado structure with multiple suction vortex.
Relevance
Tornadoes are a real concern for real people. If a relationship with climate cycles can be found, then perhaps emergency funding can be set aside in advance for victims. Also, if there is a relationship, perhaps these findings can be translated to other weather phenomena such as thunderstorms, flooding, or hurricanes. Knowing that an increase in occurrences is coming, authorities can take necessary precautions to minimize economic and human losses. Another benefit of proving a relationship is to provide a more convincing argument for the need to rethink the way we treat our environment.

Methods
We will download tornado data from The
Tornado Project and NOAA.
We analyzied this data using Statview and Excel.
Our timeline for research will was:
Week 8: turn in first draft of proposal.
Week 9: finalize proposal and collect all tornado and climate data
Week 10: Post revised proposal
Week 10: Begin to enter data into Statview and familiarize ourselves
with formation of tornadoes.
Week 11: Enter data. Post article(s) for class discussion.
Week 12: Research and prepare for class discussion.
Week 13: Finish data entering and begin interpretation.
Week 14: Interpret data and write results, discussion and conclusion.
Week 15: Present project to class and turn in final copy.
Research Problem and Data
We intend to investigate tornadoes, their frequency, and magnitude from 1950-1999.
Data was taken from the entire United States, which we hope will give us a large picture and convey any overall themes, as well as several states. These states include (1) Texas, which historically experiences the most tornadoes, (2) Nebraska, a state consistently ranking in the top 10 for tornado frequency, and (3) Ohio, our state and one in which tornadoes are known to happen, but is not considered a 'hotspot.' It is our intent that these states will give a complete picture of tornado occurrence, and show what may happen to hotspots and areas of lower activity when climate is changed.
Tornado Data for the United States from 1950 - 1999.
| 1950's | 1960's | 1970's | 1980's | 1990's | |
| Total # Tornadoes | 4796 | 6813 | 8579 | 8141 | 10696 |
| Yearly Avg | 533 | 757 | 953 | 905 | 188 |
| Monthly Avg | 44 | 63 | 79 | 75 | 99 |
| El Nino | La Nina |
| 1958 | 1951 |
| 1964 | 1952 |
| 1966 | 1956 |
| 1969 | 1965 |
| 1973 | 1971 |
| 1983 | 1974 |
| 1987 | 1989 |
| 1988 | |
| 1992 | |
| 1995 | |
| 1998 |
FREQUENCY DATA
Tornado Frequency By Decade
| 1950's | 1960's | 1970's | 1980's | 1990's | |
| Texas | 647 | 1195 | 1484 | 1492 | 1836 |
| Nebraska | 319 | 302 | 405 | 377 | 701 |
| Ohio | 71 | 148 | 202 | 165 | 251 |
There is a significant difference in the rank orders of data for Texas and Nebraska. There is no significant difference in the Texas-Ohio and Nebraska-Ohio tests.
Spearman Rank Correlation p-values:
Texas-Nebraska p=.1096
Texas-Ohio p=.0719
Nebraska-Ohio p=.0719
Tornado Frequency in El Nino Years
|
Years
|
Texas | Nebraska | Ohio | United States |
| 1958 | 75 | 55 | 12 | 608 |
| 1964 | 78 | 46 | 9 | 760 |
| 1966 | 77 | 10 | 3 | 606 |
| 1969 | 131 | 20 | 21 | 650 |
| 1973 | 152 | 19 | 55 | 1199 |
| 1983 | 190 | 16 | 10 | 995 |
| 1987 | 135 | 26 | 6 | 695 |
| 1988 | 92 | 23 | 0 | 773 |
| 1992 | 203 | 77 | 63 | 1404 |
| 1995 | 204 | 64 | 10 | 1479 |
| 1998 | 134 | 93 | 35 | 2149 |
The only two tests which revealed no significant difference between rank orders were Texas-USA and Nebraska-USA. All others revealed a significant difference.
Spearman Rank Correlation p-values:
Texas-Nebraska p=.4902
Texas-Ohio p=.1837
Texas-USA p=.0170
Nebraska-Ohio p=1983
Nebraska-USA p=.0785
Ohio-USA p=.1223
Tornado Frequency Non-El Nino Years
| 1950's | 1960's | 1970's | 1980's | 1990's | |
| Texas | 572 | 909 | 1332 | 1075 | 1295 |
| Nebraska | 264 | 226 | 386 | 312 | 467 |
| Ohio | 59 | 115 | 147 | 149 | 143 |
| United States | 4188 | 4797 | 7380 | 5678 | 5664 |

Only the Texas-USA and Ohio-USA tests revealed no significant difference in rank orders.
Spearman Rank Correlation p-values:
Texas-Nebraska p=.1096
Texas-Ohio p=.1615
Texas-USA p=.0719
Nebraska-Ohio p=.3173
Nebraska-USA p=.2301
Ohio-USA p=.0719
Tornado Frequency La Nina Years
| Texas | Nebraska | Ohio | United States | |
| 1950 | 15 | 9 | 3 | 269 |
| 1952 | 13 | 11 | 2 | 272 |
| 1956 | 56 | 36 | 10 | 567 |
| 1965 | 110 | 48 | 39 | 995 |
| 1971 | 193 | 52 | 15 | 963 |
| 1974 | 118 | 36 | 25 | 1123 |
| 1985 | 166 | 41 | 19 | 921 |
| 1997 | 237 | 43 | 18 | 1735 |

Texas-Ohio and Nebraska-Ohio were the only two rank orders which show a significant difference. The majority of tests showed a conformity in rank orders, or no significant difference.
Spearman Rank Correlation p-values:
Texas-Nebraska p=.0422
Texas-Ohio p=.1474
Texas-USA p=.0376
Nebraska-Ohio p=.1117
Nebraska-USA p=.0751
Ohio-USA p=.0438
Tornado Frequency Non La Nina Years
| Texas | Nebraska | Ohio | United States | |
| 1950's | 563 | 263 | 56 | 3688 |
| 1960's | 1085 | 250 | 109 | 5818 |
| 1970's | 1173 | 317 | 162 | 6493 |
| 1980's | 1326 | 336 | 143 | 7220 |
| 1990's | 1599 | 658 | 233 | 8961 |

Only Nebraska-Ohio revealed a significant difference in rank order correlations. All other pairs were not significantly different.
Spearman Rank Correlation p-values:
Texas-Nebraska p=.0719
Texas-Ohio p=.0719
Texas-USA p=.0455
Nebraska-Ohio p=.1096
Nebraska-USA p=.0719
Ohio-USA p=.0719
MAGNIGTUDE DATA
Magnitude counts total 1950 - 1999
| Magnitude | Texas | Nebraska | Ohio |
| F0 | 2823 | 758 | 185 |
| F1 | 1976 | 622 | 335 |
| F2 | 1112 | 267 | 177 |
| F3 | 330 | 87 | 55 |
| F4 | 78 | 46 | 31 |
| F5 | 6 | 4 | 9 |

Our Spearman rank correlation tests reveal p-values of less than .10: we can reject the null hypothesis, which for Spearman rank tests says there is a significant difference in the two sets of data. In lay terms: the rank orders of the sets of data tested here are statistically similar.
Spearman Rank Correlation p-values:
Texas-Nebraska p=.0253
Texas-Ohio p=.0350
Nebraska-Ohio p=.0350
El nino magnitude counts 1950 - 1999
| Magnitude | Texas | Nebraska | Ohio |
| F0 | 668 | 120 | 48 |
| F1 | 407 | 105 | 67 |
| F2 | 195 | 46 | 38 |
| F3 | 64 | 5 | 25 |
| F4 | 15 | 2 | 1 |
| F5 | 2 | 4 | 0 |

Here we also see p-values of less than .10, so we reject the null hypothesis.
We say there is no significant difference between any of the sets of data for
total magnitude counts.
Spearman Rank Correlation p-values:
Texas-Nebraska p=.0350
Texas-Ohio p=.0350
Nebraska-Ohio p=.0476
La nina magnitude counts 1950 - 1999
| Magnitude | Texas | Nebraska | Ohio |
| F0 | 168 | 100 | 11 |
| F1 | 261 | 74 | 44 |
| F2 | 185 | 20 | 34 |
| F3 | 40 | 16 | 11 |
| F4 | 8 | 9 | 10 |
| F5 | 0 | 0 | 0 |

Again, we see p-values of less than .10, which means we reject the null hypothesis: there is no significant difference between any sets of data in these tests. The rank orders of magnitude in la nina years for Texas, Nebraska, and Ohio are statistically similar.
Spearman Rank Correlation p-values:
Texas-Nebraska p=.0639
Texas-Ohio p=.0275
Nebraska-Ohio p=.0905
Conclusion
There is an obvious increase in tornado frequency between 1950-1999. This could
be due to increased detection. Also this could be due to changing climatic conditions.
Looking at the raw data we have seen that there are generally less tornadoes
in El Nino years compared to La Nina Years. But, since we were unable to get
climate data, we were unable to see if the change in the frequency was due to
climate factors.
Our data has failed to show a strong correlation in increase in tornado frequency
and magnitude during El Nino and La Nina events. P-values indicate that there
is some relationship between the frequency of tornadoes in La Nina years between
states and the USA. For example, the p-value for rank orders between Texas and
Nebraska is .0422therefore there is no significant difference between
them.
For El-Nino years, only two of six p-values are under .10, revealing that there
is a majority of rank orders which are significantly different. For example,
Texas and Nebraska had a p-value of .4902, meaning that there is a significant
difference.
In non-El-Nino years, we again saw 2 of 6 tests producing p-values under .10.
Therefore, 4 of 6 rank orders produced a significant difference. Texas and Nebraska
rank orders had a p-value of .1096, meaning that there is a significant difference
in that test.
In non-La-Nina years p-values indicate a strong state and national correlation.
Because all but one p-values were under .10, we could reject the null hypothesis
that there is a significant difference in rank orders. For example, Texas and
Nebraska yielded a p-value of .0719.
For total magnitude counts, we see p-values of less than .10: there is no significant
difference in rank orders. Texas and Nebraska yielded a .0253 p-value.
For magnitude in El-Nino years p-values were also less than .10 and we see no
significant difference in rank orders. Texas and Nebraska yielded a .0350 p-value.
For magnitude in La-Nina years p-values were again less than .10 and there is
no significant difference in rank orders. Texas and Nebraska yielded a .0639
p-value.
Suggestions for future research would include an investigation into tornado
frequency by latitude, rather than region alone, in El Nino and La Nina years.
Because of the latitudinal shift in the jet stream, an important ingredient
in tornado formation, we believe this could reveal some interesting results.
See our first presentation when we present tornadoes to the class.
See our tornado powerpoint presentatin where we present our final results.
Bibliography
C. Church, D. Burgess, C. Doswell, R. Davies-Jones, Editors. American Geophysical
Union. The Tornado: ItÍs Structure, Dynamics, Prediction, and Hazards.
Washington, D.C., 1990.
Allaby, Michael. Dangerous Weather: Tornadoes. Facts on File, Inc. New York,
1997.
Grazulis, Thomas P. The Tornado: Nature's Ultimate Windstorm. University of
Oklahoma Press. Norman, 2001.
Lutgens, Frederick K. The Atmosphere: Eighth Edition. Prentice-Hall, Inc. Upper
Saddle River, New Jersey, 2001.
Rampino, Michael R., Sanders, John E., Newman, Walter S., Konigsson, L.K. . Climate: History, Periodicity, and Predictability. Van Nostrand Reinhold. New York, 1987
http://ecosystems.wcp.miamioh.eduhttp://ecosystems.wcp.miamioh.edu/studentresearch/climatechange02/tornado/articles/
Pine, Devera. Chasing
Twister. Scholastic, Inc., 1998
Gives good background info and descripton of Doppler radar and detection.
Monastersky, Richard. Don't
Blame La Nina. Science Service, 1999.
Investigates why La Nina cannot be tied to formation of individual tornadoes.
Schaefer, Joseph T., and Tatom, Frank B. The
Relationship Between el Nino, la Nina, and U.S. Tornado Activity. Prepared
for 19th Conference on Severe Local Storms, 1999.
Investigates these relationships very well. Link between climate phenomena and
tornadoes.
No Author. Southern
Tornadoes: Hallmark of La Nina. Meredith Group, 2000.
Ties between La Nina and tornadoes that are relevant to our purpose.
No Author. La
Nina May Shift Tornado Activity. ENN, 1999.
More information on a possible connection.
Corfidi, Stephen F. Some
Thoughts on the Role Mesoscale Features Played in May 27, 1997 Central Texas
Tornado Outbreak. Prepared for 19th Conference on Severe Local Storms, 1999.
Talks about storms which can lead to tornadoes. Very helpful in describing formation.
Ferguson, Edward, and Ostby, Frederick. Tornadoes
of 1990: An All-time Record Year. Weatherwise, April 1991.
Explains why we might find peak tornado activity and the factors behind that.
Zimmer, Carl. The El Nino Factor. Discover, 1999.
Mazza, Patrick. The
invisable hand: as human activity warms the earth, El Nino grows more violent.
Sierra, May-June 1998.
Websites:
Tornado and Storm Research Organization
Tornado Science Charged Sheath Vortex Tornado
Nebraska Weather and Climate Information
Storm Chaser - Warren Faidley's storm chaser and natural disaster Homepage
WEATHER & EARTH SCIENCE RESOURCES |
|
OTHER ACADEMIC COURSES, STUDENT RESEARCH, OTHER STUFF
|
|
TEACHING TOOLS & OTHER STUFF
|