Significant Paper: Visualization of a simulated long-track EF5 tornado embedded within a supercell thunderstorm

Visualization of a simulated long-track EF5 tornado embedded within a supercell thunderstorm
Authors: Leigh Orf – Cooperative Institute for Meteorological Satellite Studies, University of Wisconsin, Madison, WI, USA; Robert Wilhelmson – National Center for Supercomputing Applications, Urbana, IL, USA; Louis Wicker – National Severe Storms Laboratory, Norman, OK, USA
Journal: Parallel Computing
Published: July 2016
Significance: Tornadoes are one of nature’s most destructive forces, creating winds that can exceed 300 miles per hour. The strongest tornadoes are produced by supercells, long-lived thunderstorms characterized by a persistent rotating updraft. The sheer destructive power of the strongest class of tornado (EF5) makes these storms the subject of active research. However, very little is currently known about why some supercells produce long-track (a long damage path) EF5 tornadoes, while other storms in similar environments produce short-lived, weak tornadoes, or produce no tornado at all. A breakthrough, ultra-high resolution simulation using the Univ. of Illinois Blue Waters supercomputer demonstrated a supercell producing an EF5 tornado lasting nearly two hours. New computational approaches were required for the simulation at these resolutions and for the visualization of nearly 100 TB of model output. This paper reports on the visualizations illuminating the simulation, which elucidate three-dimensional features thought to play an important role in creating and maintaining the tornado vortex.
Important Conclusions: Volume rendering of vertical vorticity (a measure of spin around a vertical axis) reveals a complex, fascinating turbulent flow regime in the vicinity of the tornado as it forms, grows, and churns for over 90 minutes. Animations of vorticity reveal the presence of dozens of shallow, curved vortices that form along the forward flank gust front and feed into the tornado vortex. The anticyclonic vortices, rather than being absorbed into the tornado’s circulation, tend to maintain their integrity, but are swept around the outer circulation of the tornado, often completing one or more full cycle of rotation while concurrently being tilted and lifted upwards by rising air.


Fig 1. Volume rendered cloud field showing an overshooting top (bulge of cloud atop the storm), anvil (horizontally oriented cloud at the top of the storm), and mesocyclone (rotating updraft manifested as a barrel-like cloud dominating the center of the image). The tornado is discernible but tiny compared to the aforementioned “storm-scale” features. Fig 2. Tornadogenesis, indicated by a descending condensation funnel. Fig 3. Shortly following tornadogenesis, the tornado (T) is visible as an erect tube of cyclonic vorticity (a measure of spin around a vertical axis), while three shallow cyclonic vortices (CV) are visible shortly before merging with the tornado.

Share this:

NOAA, University of Alabama-Huntsville and Partners Kick Off Tornado Study

Tornadoes will be the target as researchers spend the next two months in northern Alabama collecting data as part of the Verification of the Origins of Rotation in Tornadoes EXperiment-Southeast, a research project coordinated by NOAA’s National Severe Storms Laboratory. The goal is to understand how environmental factors in the southeastern United States affect the formation, intensity, structure, and path of tornadoes, as well as determine the best methods for communicating forecast uncertainty to the public.

Tornado in Sardis, MS, December 23, 2015. Credit: Jason Cooley.

The VORTEX-SE field study, set to run March 1 through April 30, will involve 40 physical and social science researchers from 20 research entities, many located in the southeast. They will deploy approximately 13 vehicles, three mobile radars and one fixed radar from  their operations base at the University  of  Alabama-Huntsville. During previous VORTEX field campaigns, researchers roamed the Great Plains, taking the instruments to the storms. This time, some of the instruments are stationary, and the domain is much smaller. As a result, the researchers expect to operate during four to five periods of several days each as the storms come to them.

The number of killer tornadoes in the southeastern U.S. is disproportionately large when compared to the overall number of tornadoes throughout the country. Researchers believe this is caused by a series of physical and sociological factors, including tornadoes at night, in rugged terrain, as well as tornadoes occurring before the perceived peak of “tornado season,” during a time of year when storms typically move quickly. Other variables include the lack of visibility, inadequate shelter, and larger population density that increases the vulnerability of residents in this area.

Damage seen on January 23, 2012 following an EF3 tornado in Center Point, AL. Credit: NWS Birmingham

“In many ways, VORTEX-SE represents a new approach to tornado research in general,” said Erik Rasmussen, VORTEX-SE project manager and Research Scientist for the University of Oklahoma’s Cooperative Institute for Mesoscale Meteorological Studies working at the NOAA National Severe Storms Laboratory. “This is the first field observing campaign in the southeast U.S. to begin to understand how the atmosphere can become locally favorable for tornadoes and how these changes can be better anticipated in the tornado forecast process.”

VORTEX-SE activities are being supported by a special Congressional allocation of more than $5 million to NOAA made in 2015. A similar allocation made this year will support additional activities in the spring of 2017.
Researchers from the following organizations are participating in VORTEX-SE:

  • NOAA Air Resources Laboratory
  • NOAA Atlantic Oceanographic and Meteorological Laboratory
  • NOAA Earth System Research Laboratory
  • NOAA National Environmental Satellite, Data, and Information Service
  • NOAA National Severe Storms Laboratory
  • NOAA National Weather Service
  • University of Oklahoma Cooperative Institute for Mesoscale Meteorological Studies
  • Colorado State University
  • Mississippi State University
  • NASA
  • National Center for Atmospheric Research
  • National Science Foundation
  • North Carolina State University
  • Purdue University
  • Texas Tech University
  • University of Alabama – Birmingham
  • University of Alabama – Huntsville
  • University of Georgia
  • University of Illinois Urbana-Champaign
  • University of Louisiana at Monroe
  • University of Maryland
  • University of Massachusetts
  • University of North Carolina at Asheville
  • University of Oklahoma
  • University of Tennessee
Decatur Daily Photo by Gary Cosby Jr. A massive tornado swept across Limestone County Wednesday afternoon following the same track as a killer tornado in the 1974 outbreak. The storm is located over the TVA power lines and over Ingram Rd. west of Highway 31 in Tanner. The time of the image is 4:21:13 PM April 27, 2011. The picture was shot from Highway 31 in front of Swan Creek Community (trailer court) looking just a couple of degrees south of due west.
A massive tornado in Limestone County, AL, April 27, 2011. Credit: David Cosby Jr., Decatur Daily

Keli Pirtle,, (405) 203-4839

More information

[avl_video id=”3498″]
Watch this video on Youtube




Share this:

VORTEX2: 2009-2010


Spring 2009 brought scientists at NOAA’s National Severe Storms Laboratory back to the field for the largest tornado research project in history. With support from NOAA and the National Science Foundation, more than 100 scientists, students, and staff sought to collect data that would provide better understanding of tornado intensity, longevity, and behavior. The team deployed 10 mobile radars and 40 additional vehicles with custom instrumentation for data acquisition.

vort-2 mobile mesonets
Mobile mesonets deployed on VORTEX2.

NSSL’s Lou Wicker was one of six principal investigators on the project, which included Chris Weiss from Texas Tech, Joshua Wurman from the Center for Severe Weather Research, Yvette Richardson from Penn State, David Dowell from the National Center for Atmospheric Research, and Howard Bluestein from the University of Oklahoma.

VORTEX2 Principal Investigators

Between May 10 and June 13, 2009, researchers traveled more than 10,000 miles across the central and southern plains. A number of storms were analyzed, including one supercell that spawned a tornado. On June 5, researchers were able to deploy all of their mobile research equipment in a tornadic supercell in LaGrange, Wyoming. They collected data on the tornado from 20 minutes before formation until dissipation. This remains the best-sampled storm on record.

Tornado in LaGrange, Wyoming

During the second year of VORTEX2, vehicles logged more than 25,000 miles each. Scientists sampled 36 supercells and 11 tornadoes. The field project resulted in numerous studies published in peer-reviewed journals.

Photo Credit: Susan Cobb

To read some of the publications resulting from VORTEX2 research:

For more pictures:

Share this:

VORTEX: 1994-1995


In 1994, scientists at NOAA’s National Severe Storms Laboratory embarked on a field campaign that aimed to study why certain supercells produce tornadoes. Coordinated by Bob Davies-Jones, Jerry Straka, and Eric Rasmussen, the two-year project sought to address questions about the dynamics and evolution of tornadic storms.

Bob Davies-Jones, Jerry Straka, and Eric Rasmussen

VORTEX took place in the central and southern plains, where conditions are most favorable for severe weather outbreaks in the spring. The region also boasts an easy-to-navigate road network and flat terrain, making it particularly well suited to storm observation.

Mobile mesonet tracking the storm.

Field work was completed between April 1 and June 15, with data analysis taking place during the remainder of the year. Because the project spanned over two years, researchers were able to adapt lessons learned in the first year to improve strategies in the second.

Research vehicle near tornado in Dimmitt, TX.

The first VORTEX field campaign employed a number of vehicles equipped with custom weather sensors, including mobile mesonets, mobile radars, and mobile sounding systems. In addition, the NOAA P-3 participated in both the 1994 and 1995 seasons.

The original fleet of mobile mesonets preparing for VORTEX.

During its two seasons, VORTEX scientists intercepted ten tornadoes. Most notably, researchers followed the Dimmitt, Texas tornado on June 2, 1995, which was the most thoroughly observed tornado on record to that point.

The tornado in Dimmitt was powerful enough to tear up highway pavement!

Each storm was analyzed from multiple angles, some at close range. Research from the VORTEX project was instrumental in improving National Weather Service tornado warning lead time to thirteen minutes. The project also inspired two follow-up campaigns: SubVORTEX in 1997 and VORTEX-99 in 1999.

Share this:

Tornado Warning Decisions Using Phased Array Radar Data

Weather and Forecasting: Early Online Release

Tornado Warning Decisions Using Phased Array Radar Data

Authors:  Pamela Heinselman, Daphne LaDue, Darrel M. Kingfield, and Robert Hoffman

The 2012 Phased Array Radar Innovative Sensing Experiment identified how rapidly scanned full-volumetric data captured known mesoscale processes and impacted tornado-warning lead time. Twelve forecasters from nine National Weather Service forecast offices used this rapid-scan phased array radar (PAR) data to issue tornado warnings on two low-end tornadic and two nontornadic supercell cases. Verification of the tornadic cases revealed that forecasters’ use of PAR data provided a median tornado-warning lead time (TLT) of 20 min.  Precursors that triggered forecasters’ decisions to warn occurred within one or two typical WSR-88D scans, indicating PAR’s temporal sampling better matches the time-scale at which these precursors evolve.

Share this:

NOAA study shows pattern of fewer days with tornadoes, but more tornadoes on those days

A massive tornado approaches Moore, OK on May 3, 1999

Are tornadoes increasing? Not really, the number has remained relatively constant. What is changing is that there are fewer days with tornadoes each year, but on those days there are more tornadoes, according to a NOAA report published today in the journal Science.

NOAA researchers looked at records of all but the weakest tornadoes in the United States from 1954 to 2013 for the study, “Increased variability of tornado occurrence in the United States.” They found that although there are fewer days with tornadoes, when a tornado does occur, there is increased likelihood there will be multiple tornadoes on that day. A consequence of this is that communities should expect an increased number of catastrophes, said lead author Harold Brooks, research meteorologist with the NOAA National Severe Storms Laboratory.

“Concentrating tornado damage on fewer days, but increasing the total damage on those days, has implications for people who respond, such as emergency managers and insurance interests,” Brooks said. “More resources will be needed to respond, but they won’t be used as often.”

Why tornadoes are concentrating on fewer days is still an open question, Brooks said. The pattern may be connected to changes in weather and climate. More research involving climate and tornado scientists is needed.

The study also showed there is greater variability in the starting date of spring tornado season, with more early starts and late starts in recent years. From 1954 to 1997, 95 percent of the time tornado season started between March 1 and April 20. But in the last 17 years, this happened only 41 percent of the time.

Researchers note tornadoes differ from tropical cyclones or hurricanes in the North Atlantic because tornadoes can occur year round. In fact, tornadoes have occurred in the U.S. on every calendar day at some point during the past 60 years.

Recent experience illustrates the study’s findings of variability. The study looked at tornadoes rated EF1 or higher on the Enhanced Fujita Scale, a measure of the damage caused by tornadoes with categories from EF0 to EF5. From June 2010 to May 2011, there were 1,050 EF1 and stronger tornadoes, the most in any 12-month period on record. Shortly after that, the U.S. saw the fewest in a 12-month period, only 236 EF1 and stronger tornadoes occurred from May 2012 to April 2013. November of 2012 had no EF1 tornadoes, but November of 2013 had the sixth most on record, with 66. There have been a relative low number of tornadoes to date in 2014, with an estimated 800 tornadoes of all intensities reported through September, almost 400 tornadoes below what is considered a normal year.

The study’s results are a first step toward understanding the relationship between changing tornado activity and a changing climate. The next step will be for climate scientists and tornado researchers to work together to identify what specific large scale pattern variations in climate may cause, or are related to, clustering of tornado activity.

Co-authors of the study are Gregory Carbin and Patrick Marsh with the NOAA’s National Weather Service Storm Prediction Center.

Share this:

The 40th anniversary of the Union City, OK tornadic storm

Doppler data processed later revealed first Tornadic Vortex Signature

Tornadic vortex signature (stippled) at 2:46 CST on 24 May 1973 at a height of 3.5 km.
Tornadic vortex signature (stippled) at 2:46 CST on 24 May 1973 at a height of 3.5 km. Negative numbers are wind speeds moving towards the radar, positive numbers are wind speeds moving away from the radar.

May 24, 2013, is the 40th anniversary of the Union City, Okla., tornadic storm.   Researchers from the NOAA National Severe Storms Laboratory collected data on the Union City storm using experimental Doppler radar. When they were able to process the data, they discovered a unique pattern now known as the Tornadic Vortex Signature (TVS).

“The TVS revolutionized the NWS ability to warn for tornado activity with sufficient lead time to save lives,” said former National Weather Service Director Joe Friday.

What follows is a narrative by NSSL’s research meteorologist Rodger Brown describing the activities that took place on the day of the Union City tornado.

“On May 24, 1973, the darkened National Severe Storms Laboratory (NSSL) radar room was a typical beehive of activity. Meteorologists, aircraft controllers and coordinators, radar technicians and visiting scientists were monitoring several radar scopes, including one from an experimental Doppler radar. Excitement rose in the room when a phone call was received from members of the Tornado Intercept Project (TIP) team reporting that a large tornado was touching down to their northwest.

“They were positioned 5 km south of the small farming community of Union City, 47 km west-northwest of NSSL.

“As word quickly spread throughout NSSL, a number of staff crowded onto the observation platform atop the building. The tornado was visible in the distance next to a dark rain shaft. With time, the tornado became obscured by the rain.

“An hour earlier, the Doppler radar meteorologist, engineer and technician in the nearby Doppler radar building began sampling the storm at 2:46pm.  

“When the radar data were processed months later, the data revealed the presence of a vortex about 5km in diameter at heights of 5 to 8km above the ground.  By 3:15, there was clear Doppler velocity evidence that a smaller tornado scale vortex was present at mid-levels near the storm’s southwest edge. Researchers compared the data with time-stamped photos and movies.  They found at the same time, the NSSL TIP team observed funnel-like protrusions extending beneath the more rapidly rotating lowered cloud base.

Damage path of the Union City tornado. The funnels and debris clouds were matched to photographs taken by the NSSL storm intercept teams who documented the entire life cycle of the tornado on film. Letters A through H indicate damaged farmsteads.
Damage path of the Union City tornado. The funnels and debris clouds were matched to photographs taken by the NSSL storm intercept teams who documented the entire life cycle of the tornado on film. Letters A through H indicate damaged farmsteads.

“With time, what is now known as the Doppler Tornadic Vortex Signature (TVS) descended to the ground and at the same time, a funnel appeared below the cloud base. From 3:38 to 3:48, while the funnel descended and retracted several times as it moved eastward, a dust cloud continuously was evident on the ground. The TIP team, racing eastward during this development stage, arrived at their final photography site—9km southeast of the tornado—just before the visual funnel made continuous contact with the ground.”

The tornado caused fatalities and extensive damage as it passed through the heart of Union City. The newly commissioned Doppler radar at NSSL observed this tornado, and the Tornado Intercept Project researchers photographed the tornado’s life cycle. The radar, coupled with the photographic evidence of the tornado’s development, revealed previously unknown information about motion inside thunderstorms with a persistent rotating updraft, a type known as supercells.

This event played a major role in the decision to develop and deploy a nationwide network of WSR-88D/NEXRAD radars. The NSSL TIP also proved its scientific worth and paved the way for all the tornado intercept research that goes on today.

The discovery of the TVS and other Doppler velocity signatures led to dramatic improvements in accuracy and lead-time in forecasting severe storms nationwide, and as a result, the ability to save lives and prevent serious storm-related injuries.

Union City Tornado, May 24, 1973

Share this:

Radars capture valuable scans of May 2013 tornadoes

On both May 19 and May 20, 2013, NSSL researchers collected data on storms that produced tornadoes using both the NWRT Phased Array Radar (PAR), and the mobile dual-polarized radar.  The NWRT PAR can scan the sky in less than one minute, five-times faster than current weather radars.  Datasets from the rapidly-updating NWRT PAR will help researchers better understand the evolution of rotating thunderstorms and the tornadoes they produce.

May 19, 2013

The NWRT PAR, a retired Navy surveillance radar adapted for weather, scanned a storm from its first radar echoes through its production of a tornado.  When this storm moved out of range, the PAR was directed to scan the tornado that formed in Norman, Okla. through the time the tornado moved into Shawnee, Okla., killing two people.  Loops of radar imagery are below with still images following.

May 20, 2013

The NWRT PAR scanned the Newcastle-Moore tornadic storm for almost an hour.  This storm produced an EF5 tornado that killed 24 people and injured more than 300.

Here are links to radar loops (still images below):

TOR warning at 1940 UTC

The NWRT was tracking (before 2000UTC) a storm to the south that also had a TOR warning
TDWR: The Moore Tornado from 1930-2059 UTC
Earlier on May 19, 2013 the NWRT PAR scanned this storm from the time it formed, through the formation of a tornado until it died.
Earlier on May 19, 2013 the NWRT PAR scanned this storm from the time it formed, through the formation of a tornado until it died. This radar can complete a full scan of the weather in less than one minute.
WSR-88D scan from May 19, 2013.
WSR-88D scan from May 19, 2013. It takes 4-5 minutes for this radar to complete one full scan of the weather.
The NWRT Phased Array Radar scanned the tornadic storm that hit Shawnee, OK.
The NWRT Phased Array Radar scanned the tornadic storm that formed later, hitting Shawnee, OK.














WSR-88D scan of the May 20, 2013 storms.
WSR-88D scan of the May 20, 2013 storms.
May 20, 2013 Terminal Doppler Weather Radar scanned the Newcastle-Moore tornado from 1930-2059 UTC.
May 20, 2013 Terminal Doppler Weather Radar scanned the Newcastle-Moore tornado from 1930-2059 UTC.
The NWRT Phased Array Radar scanned the tornadic storm over Moore for almost an hour from 2003-2059 UTC.
The NWRT Phased Array Radar scanned the tornadic storm over Moore for almost an hour from 2003-2059 UTC.
Share this:

Scientists re-visit the Tri-State Tornado

desotorbThe March 18, 1925 Tri-State Tornado was unusually severe, killing 695 people while it was on the ground for a record 219 miles crossing parts of Missouri, Illinois and Indiana. Unfortunately, there is only one formal paper regarding the tornado and its meteorological setting.

A team of eight severe storms meteorologists re-analyzed the event using all relevant U.S. Weather Bureau data on the Tri-State Tornado.  The results, published in the Electronic Journal of Severe Storms Meteorology, revealed previous analyses of the surface weather conditions were inaccurate and led to misconceptions about where the tornado formed in reference to the existing weather system.  The authors include retired NSSL Director Bob Maddox, retired NSSL/CIMMS researchers Chuck Doswell, Don Burgess and Charlie Crisp, retired Storm Prediction Center (SPC) meteorologist Bob Johns and current SPC meteorologist John Hart, and Steve Piltz from the National Weather Service Forecast Office in Tulsa, Okla.

The researchers concluded there was no singular feature in the meteorological setting that would explain the extreme character of the Tri-State tornado.  The storms of 18 March were associated with a rapidly moving cyclone that was not unusually intense.  The new analyses show a long-lived supercell that developed very near the center of the cyclone produced the tornado, possibly where a warm front and a distinct dryline intersected.  The south-to-north temperature gradient was very pronounced due to cooling produced by early morning storms and precipitation.  The tornadic supercell tracked at an average speed of 59mph moving farther away from the cyclone center with time.  And, the storm remained very close to the surface warm front.

Researchers did find as the supercell and dryline moved rapidly eastward, the northward advance of the warm front kept the tornadic supercell within a very favorable storm environment for several hours.  It appears this consistent time and space connection of the supercell, warm front, and dryline was extremely unusual.

With reanalysis beginning 70 years after the tornado, it was impossible to confirm the complete continuity of the damage path along the reported path.  Even with extensive field work discovering 2,395 individual damage points, there were 32 gaps of at least one mile in length, but only 7 gaps longer than 2.5 miles in length.  All of the longer gaps were in the Missouri portion of the path; within the sparsely-populated Ozark mountain area.  Assuming that gaps shorter than 2.5 miles might still represent a continuous tornado, the continuous path was at least 174 miles long.  Additional, previously unreported tornadoes were also found before the beginning and after the end of the Tri-State Tornado.  The research also allowed for conclusion that the storm was a supercell; classic in its stages and high-precipitation in the later stages.  The supercell also produced accompanying hail up to baseball size and non-tornadic damaging winds.

Share this: