Tuesday, March 1, 2016

To Build a Tornado


Not one, but three violent tornadoes struck the Omaha metro area in a single hour Easter Sunday 1913. What weather conditions built those tornadoes? Could they recur? By guest author Evan Kuchera, USAF meteorologist

[Nebraska’s centennial commemorations of the devastating Omaha tornado on Easter Sunday 1913 left Department of Defense meteorologist Evan Kuchera, stationed at Offutt Air Force Base near Omaha, with two burning questions: What meteorological conditions led up to such an extraordinary tragedy? Could those conditions recur today? Below is a condensation of a presentation he gave on his first results—a tour de force of sleuthing! –T.E.B.]

The Omaha tornado of March 23, 1913—rated as a violent F4 with a funnel of destruction a quarter-mile wide—still ranks as Nebraska’s deadliest, single-handedly claiming more than 100 lives (see “’My Conception of Hell’) and as the 13th deadliest twister in the nation. But it did not act alone. Ten minutes 

Map of the modern Omaha metro area with the approximate tracks of what were called the Yutan, Omaha, and Council Bluffs tornadoes. From west to east, all three F4 tornadoes—some of the most violent that occur—struck within 20 miles and 45 minutes. Credit: Evan Kuchera
earlier and 25 minutes later, two other violent F4 tornadoes also struck what is now the same metropolitan area, killing another 50 people.

Given that, on average, there are only maybe 10 such violent F4 tornadoes per year in the entire U.S., to have three of them hit 10 to 20 miles apart in the same metro area within a single hour is truly remarkable. It also begs the important question: could it happen again? 

To ascertain the odds, it’s necessary to figure out the larger context within which they formed.

Eight F4 and F3 tornadoes led to at least 201 deaths in Nebraska, Iowa, and Missouri Easter night, March 23, 1913; they were accompanied by other weaker ones in those states plus several in Kansas (not shown). The three that struck what is now the Omaha metropolitan area are the three tornadoes named “Omaha,” “Yutan,” and “Council Bluffs.” Credit: Trudy E. Bell

First I went back to original 1913 weather reports as well as several dozen newspaper accounts from four states (scans supplied by historian/science journalist Trudy E. Bell) and plotted data by hand; then my meteorologist colleague and coauthor Jeff Hamilton ran the data through a supercomputer simulation tool to see if we could reconstruct a more complete picture of the meteorological events according to current scientific understanding.

Buried data about the freakish storm system
In 1913, the average lay person in Nebraska or other tornado-prone areas recognized that tornadoes occurred on unusually warm and humid days, produced by odd-looking parent thunderstorms that usually moved from southwest to northeast. People knew they needed to go below ground to escape, and often there was a calm before the storm.

Types of observational data Kuchera consulted

But these Easter 1913 tornadoes were unexpected, if not freakish, in many ways. In Nebraska, March is early for tornadoes; peak tornado season is May and June. Quotes from many people in newspapers as well as a meteorologically savvy professor at Creighton College (now University) indicated that Easter Sunday must not have been a very hot or oppressive day. Highs that day approached 60 degrees, and no reports earlier in the day suggest that anyone perceived the weather as being unusually sultry or warm for the end of March—and those informal reports are backed by the recorded official data. That is meteorologically significant because cool air holds less moisture and less heat energy than warm air.

Daily weather map from a NOAA archive shows a high-pressure system to the east and a low-pressure system approaching from the west, drawing up moisture from the Gulf of Mexico.

A daily weather map in the archives of today’s National Oceanic and Atmospheric Administration (NOAA) shows that at 8 AM Easter Sunday, March 23, 1913, a significant high-pressure system was receding to the east, and a low-pressure system was approaching out of Colorado. That’s a standard storm track in the winter. Little arrows indicate the clockwise flow around the high and the counter-clockwise flow around the low, indicating that moisture was being pulled up from the Gulf of Mexico to the south. Also—as you would expect for a storm system that could produce tornadoes as early as late March—the low-pressure system is very strong.

National Weather Service personnel as well as volunteer cooperative observers (the NWS Cooperative Observer Program COOP, started in 1890, still continues today) recorded prescribed observations in the morning and evening, typically at 7 AM and 7 PM Central time, called 1Z and 13 Z on the charts (Z stands for Zulu, or Greenwich Mean Time, a universal reference regardless of time zone; there was no daylight saving time in 1913). Scans of these official reports, archived by the National Climatic Data Center (NCDC) and extending back well over a century, are accessible in an online database. From these reports, we can reconstruct a surprising amount of information about how the powerful storm system developed over Nebraska on Easter Sunday 1913.

Kuchera consulted the handwritten reports from cooperative observers all around Nebraska. Such COOP reports show high and low temperatures, precipitation, prevailing winds, and comments for significant events. In some regions, these reports are quite dense (one per county) for Nebraska, sparsely populated in 1913. This one for the month of March 1913 was from the COOP observer in Osceola in Polk County. Courtesy National Climatic Data Center

For example, the COOP observer in Osceola, about 60 miles west of Omaha, noted that about 4:30 PM, the wind there shifted from the south to the northwest. In 1913, no one yet had the concept of a front—the leading edge of a cold or warm air mass—so they were not able to say that a front came through, and just noted the wind shift. But based on the other data I’ve seen and my knowledge of meteorology, 4:30 PM would have been the time the cold front would have moved through Osceola. Indeed, this wind-shift comment was extremely helpful in placing the position of that cold front at 4:30, which is an actual observation I would not have otherwise had. 

The observer also noted “Omaha tornado.” That notation gives a sense of the weather savviness of these volunteers: they understood that a wind shift was related to thunderstorms and tornado weather. So they knew that the front that came through Osceola at 4:30 PM was likely the cause of what happened to their east an hour or so later.

Mapping the weather details
From the National Weather Service data in Monthly Weather Review, I plotted by hand all the high temperatures across more than half a dozen states on Easter Sunday, March 23 1913. Note that there was a lot of cold air north of this low-pressure system: northwestern Nebraska had 30s for highs, while south of Nebraska temperatures reached the 80s. The temperature gradient is very tight: there was a rapid change of temperature across a short distance—a sure indication of a strong weather system.

Kuchera’s hand analysis of Easter Sunday high temperatures
But my plot of high temperatures reveals another significant feature. Note the little patch of temperatures that exceeded 80 degrees in Kansas and Oklahoma, and how that is nudging toward southeastern Nebraska. That is evidence of the presence of a dry line. The dry line is the demarcation between the mass of moist air from the Gulf of Mexico and the mass of arid inland air from the Rocky Mountains to the west. Typically, the highest temps are right along the dry line. You frequently see a dry line set up in situations like this: the low pressure system draws warm moist air up from the south, but warm dry air is already in place, so a boundary forms between the two even though there is not any strong temperature change across the boundary. 

Kuchera’s summary of observations of dust, high winds, hail, and rain in the COOP reports for Easter Sunday, March 23, 1913, hand-plotted on a map of Nebraska, Iowa, Kansas, and Missouri counties.

From COOP reports and comments, I also able to synthesize a storm report summary as you would see today. On a map of counties in four states, I plotted dust (D), strong winds (W), hail (H), and rain (R). The dust observations reveal the extent of the dry line where warm, dry, dust-laden air made it across Kansas and neighboring states (see “Great Easter 1913 Dust Storm, Prairie Fires—and Red Rains). Kansas had few rain reports, hinting that there were not widespread thunderstorms as there were up in Nebraska. Thus, a lot of the Kansas winds were not related to thunderstorms, but were gradient winds associated with the low-pressure system. On the other hand, the high winds in Nebraska are associated with the thunderstorms. My plot reveals the high concentration of severe weather where the thunderstorms came through, but there was other severe weather as well, notably hail and severe winds.

Kuchera’s map of the Great Plains plotting observed temperatures (T), dew point temperatures (Td), and winds for 7 AM Easter Sunday morning 1913 reveals why tornadoes 10 hours later near Omaha were so unexpected. Winds are indicated by the little barbs with the flags:  The direction the barbs point shows wind direction, and the number of flags on the barb is sustained wind strength (e.g., two flags indicate 20 knots or nautical miles per hour). Temperatures at 7 AM are numbers in red, and dew points are numbers in green.

I also plotted winds, temperatures, and dew points for 7 AM Central (13Z) Easter morning. The dew point is the temperature at which dew can form: the higher the dew point, the more humid the air and the greater the chance of severe weather. The threat of tornadoes increases when dew point temperatures exceed 55 degrees—areas shaded in darker green. The map shows that mass of moist air was nowhere near Nebraska 10 hours before the tornadoes. 

In Omaha itself, it surprised me to see that—for such powerful tornadoes that afternoon—the morning temperature was 40 degrees. There was certainly no clue that morning that there would be severe weather later that day. It was a typical cool March morning. Yes, some winds had picked up from the south, but March in Nebraska is a windy time, so that in itself is not altogether unusual. Several locations show 20-knot winds, there is a 25-knot wind at Amarillo, Texas. But that morning, few features were yet in play.

Kuchera’s map for 7 PM Easter Sunday evening of the Great Plains plotting winds, temperatures, and dew point temperatures reveals the meteorological conditions shortly after most of the deadly tornadoes had passed (another F4 still had yet to form in Missouri). It reveals the moisture from Oklahoma and Texas rapidly moved north that day: note three or even four flags on some barbs, meaning a 30- or even 40-knot sustained wind, which is incredibly strong.

Twelve hours later (7 PM Easter evening in Nebraska, or 1Z March 24), official readings were taken shortly after most of the tornadoes had passed. Plotting that data reveals what we now know was the low-pressure front in western Iowa where the tornadoes had just ended. Observations recorded in Lincoln and Omaha both show that the winds have switched to the north and the temperatures have plummeted. 

But note how the dew points in Missouri, Iowa, and Illinois have all dramatically jumped up in the upper 50s or around 60, whereas the morning dew points in those areas were in the 40s or even in the 30s. That reveals that Gulf of Mexico moisture moved rapidly north up from Oklahoma and Texas, across Kansas, and into Nebraska and Iowa—brought by sustained winds of up to 40 knots. Those are incredibly strong sustained winds. Together, the dew points and winds recorded indicate that this Easter 1913 low pressure system was very intense, drawing up moisture very fast.

Note the similarity of this generic map of conditions ripe for tornadoes to the map of meteorological conditions on Easter Sunday, March 23, 1913, which Kuchera was able to synthesize from 1913 observations to deduce the positions of the warm and cold fronts and dry line
The dashed line indicates the position of the dry line. How do I know where it likely was? Official weather records document that in Dodge City that the dew point was only 15, so we know the dry line is east of there. But in Wichita the dew point is still 58, so we also know that the dry line is still to the west. The dry line could have been anywhere between these two observations. But the newspaper accounts from Trudy allowed me to time when and where weather features moved through. The data also allowed me to provide my best guess for the locations of the cold front (blue triangles) and warm front (red half-circles) were at 7 PM Easter night.

Reconstructing the crime scene
So much changed in under 12 hours. As the low-pressure system intensified, strong frontal systems developed, so it is not surprising some violent tornadoes emerged. Because they developed about two hours before the official reports, we don’t have official observations directly from that time. But important clues in the COOP reports, newspaper accounts, and later scientific articles allow us to reconstruct weather conditions leading up to the major tornado outbreak. 

For example, a 1914 journal article written by G.E. Condra and G.A. Loveland, two professors at the University of Nebraska at Lincoln, stated that the relative humidity in Lincoln jumped from 53% to 78% from 2150Z to 2230Z—that is, from 3:50 PM and 4:30 PM local time. When the relative humidity jumps like that, it’s one of two things: either the temperature dropped so you have the same amount of moisture in colder air, or the temperature stayed the same and the moisture content went up. The cold front could not have gone through Lincoln yet because the wind shift wasn’t recorded in Osceola—which is west of Lincoln—until 4:30 PM; thus, there’s no way the front could have been all the way to Lincoln yet. So the humidity jump must mean that over a 40-min period these high dew pts arrived, and they arrived very quickly. That relative humidity change corresponds to about a 10-degree jump in dew point—a really large, substantial jump in a severe weather situation. 

Now, the Omaha tornado started in Lincoln started right around 23Z, that is, around 5 PM local time. So, essentially, the moisture required to create the tornado arrived in Lincoln an hour before the storms formed. The timing on this intersection of events is just impeccable: the very intense low-pressure system, the arrival of all that moisture, and then the cold front and dry line all came together right like magic in the late part of the day. 

Kuchera’s deduced estimate of weather conditions around 5 PM Easter Sunday 1913—the final map of the “magic moment” when all the conditions met to create the family of violent tornadoes.

The major tornadoes all developed roughly in the next hour. As the cold front/dry line came through, you can imagine all the thunderstorm activity out in front of it. This is what I originally set out to do with this project and sleuthing: to make this map reconstructing the surface chart at that time, based on all the information and data I had and modern meteorological knowledge.

Some descriptions from various people are meteorologically significant. The professor at Creighton noted that immediately behind the Omaha tornado, the sky was clear right up to the cirrus cloud and that the cumulonimbus banked “mountain high” behind the tornado, the highest he’d ever seen. His desciption is painting a very dramatic picture of what the storm actually looked like. He also reported not much rain as the tornado passed, although a heavy thunderstorm followed 15 minutes later. 

Because that same sequence of events was also observed in Lincoln, I’m thinking two things: First, along the cold front itself there was a solid line of thunderstorms. However, about 30 miles ahead of the cold front, discrete supercell thunderstorms formed. (Supercells are rotating storms that give rise to the most violent tornadoes.) Discrete storms are best for tornadoes in favorable environments because there are no nearby thunderstorms that disrupt tornado formation processes due to storm collisions. So the conditions were exactly right for very violent tornadoes to form.

Condra and Loveland reported the low pressure center reached a surface low pressure of 991 millibars. That is pretty intense. They also reported that the cloud level was low with the tornadoes. When the relative humidity is high—as it was here (78 percent) and you lift the air, it makes a cloud at a lower base. We know from modern research that low cloud base is one of the key ingredients to making violent tornadoes. So that observation was an important detail.

Kuchera’s redrawn map of tracks and timing of five of the major tornadoes in the Great Easter 1913 outbreak in the Lincoln/Omaha area, superimposed on a modern map.

Note the southward progression with time in the order in which the tornadoes formed. This ties in neatly with the notion that there was an intersection between the cold front and the dry line: as the cold front swept south and overtook the dry line, their intersection moved farther south. That intersection was where lift was strongest for generating supercells and tornadoes, which would then move northeast. This process continued into northwest Missouri until about 8 PM that night, when the last F4 spun off. 

One last note: the tornado that devastated Omaha was not the strongest one in the family. Condra and Loveland noted “extreme energy” in the Yutan and Berlin (also called Otoe) tornadoes. The only reason those two especially intense tornadoes did not kill more people is that they went through sparsely populated rural areas.

Hopes for the future
I feel I understand this Great Easter 1913 tornado outbreak better than any modern event I’ve examined. All the brain power required to piece together all these puzzle pieces created in my mind a 3D conceptual model of what was actually going on—really surprising, given that it’s a hundred years old and the data are really sparse. Nonetheless, I feel pretty confident that what I’m laying out here is what really happened. Hopefully we can learn from it for future events. 

[This research blog installment represents only the first half of Kuchera’s presentation, covering his hand analysis. He and several colleagues also ran the 1913 tornado data through a major computational tool called the Twentieth-Century Reanalysis Project (20CR) On U.S. Department of Energy supercomputers. 20CR was developed in the last decade by NOAA and various U.S. and international partner agencies to elucidate long-term relationships between weather and climate. However, it also enables today’s meteorologists to use historic surface measurements of atmospheric pressure to reconstruct probable conditions in the atmosphere aloft, thereby gaining insight as to the possible physical causes of historic extreme weather events (see the use of 20CR by Cleveland National Weather Center senior hydrologist Sarah Jamison in reconstructing the events leading up to the 1913 flood in "Be Very Afraid..."). Kuchera and his colleagues intend to prepare their computational work in reconstructing the Great Easter 1913 tornadoes work for scientific publication; the results of this computational modeling will be summarized as a later post at an appropriate time. – T.E.B.]

About the Author

Department of Defense meteorologist Evan Kuchera at his station at Offutt Air Force Base near Omaha, Nebraska. Credit: Trudy E. Bell
Evan Kuchera, a life-long resident of Nebraska, has always been fascinated by the severe weather that strikes each spring and summer, leading to his choice of a career in meteorology by attaining an MS from the University of Nebraska at Lincoln. His current job as DOD meteorologist is to utilize numerical weather modeling to provide forecast information for the needs of the US Air Force as part of the 16th Weather Squadron in the 557th Weather Wing.

Next time: Men of the Hour

Selected references
Descriptions of the entire family of Easter 1913 tornadoes can be found in Bell, Trudy E., “The Devastating Nebraska–Iowa–Missouri Tornadoes of 1913:Harbingers of the U.S.’s Now-Forgotten Most Widespread Natural Disaster,” unpublished research paper presented at the 2007 Missouri Valley History Conference in Omaha, Nebraska. It cites specific newspaper accounts as evidence for the tornadoes being more numerous, more destructive, and more lethal than official figures suggest. Included is a discussion why newspaper reporters, railroad personnel, and farmers would have been reliable and credible observers for tracing additional damage and inferring the full extent of the supercell storm system.

Fujita scale circle diagram Credit: timesonline
Background about violence of tornadoes is at this NOAA page. The Fujita scale circle diagram at left may also help visualize the damage wreaked by tornadoes,

The concept of weather fronts being leading edges of air masses of different temperature and humidity--a concept fundamental to modern meteorology--was not described until 1919, and took a couple of decades to be widely accepted.
Condra, G.E., and G.A. Loveland, “The Iowa-Nebraska Tornadoes of Easter Sunday, 1913,” Bulletin of the American Geographical Society 46(2): 100–107, 1914.
Grazulis, Thomas P., Significant Tornadoes, 1880-1989. St. Johnsbury, VT: Environmental Films, 1991. Classic and fascinating two-volume reference detailing virtually every U.S. tornado F2 and greater for more than a century. Grazulis now runs The Tornado Project.

Information about metropolitan statistical areas used by the U.S. Census Bureau, the Office of Management and Budget, and other organizations is here and here. A map of the Omaha-Council Bluffs-Fremont combined statistical area is here.

A hand analysis of the Great Easter 1913 F4 tornado—separate from the Nebraska-Iowa-Missouri family—that devastated Terre Haute, Indiana after 9 PM that same night, see “Terror in Terre Haute.” 

Various Nebraska centennial commemorations are recounted inHappy 1913Centennial Year!” (January 6, 2013); “1913 Great Easter Disaster Centennial Update” (February 2); “Centennial Month! Events Update” (March 3); “Centennial Update: April through December” (April 13), and "1913 Easter National Calamity: Centennial Highlights--and Legacy" (January 1, 2014).

For how Cleveland-based U.S. National Weather Service hydrologist Sarah Jamison used 20CR to reconstruct the meteorology leading up to the Great Easter 1913 flood in Ohio and Indiana, see “Be Very Afraid….

Bell, Trudy E., The Great Dayton Flood of 1913, Arcadia Publishing, 2008. Picture book of nearly 200 images of the flood in Dayton, rescue efforts, recovery, and the construction of the Miami Conservancy District dry dams for flood control, including several pictures of Cox. (Author’s shameless marketing plug: Copies are available directly from me for the cover price of $21.99 plus $4.00 shipping, complete with inscription of your choice; for details, e-mail me), or order from the publisher.