Testimony Of
Dennis McCarthy
Meteorologist In Charge
Norman Weather Service Forecast Office
National Oceanic and Atmospheric Administration
U.S. Department of Commerce
before the
Subcommittee on Energy and Environment
and Basic Research
Committee on Science
U.S. House Of Representatives
June 16, 1999
Thank you Mr. Chairman and members of the Subcommittees, for the opportunity to testify about tornadoes and tornado forecasting. I am the Meteorologist in Charge of the National Weather Service Forecast Office in Norman, Oklahoma, which serves central Oklahoma and northwestern Texas. The Norman Forecast Office is one of 121 offices strategically placed across the country. These offices are the Nation's first line of defense for all manner of extreme weather and flash-flood events. Supported by the National Oceanic and Atmospheric Administration's (NOAA) remote sensing capabilities, data, and research activities, these offices deliver critical information necessary to get people out of harm's way.
The United States is the most weather hazard prone Nation on earth. Over 1,000 tornadoes occur each year. Tornadoes can occur in every state in the United States, on any day of the year, and at any hour. They are most common across the south and middle part of the country, east of the Rocky Mountains and west of the Appalachian Mountains. Tornado season begins in late winter or early spring in the South and moves northward. Tornadoes are most frequent during the spring and summer, while a secondary maximum occurs in the fall.
My comments today are tempered by thoughts of the May 3rd tornado outbreak that killed 42 people and injured 795 in Oklahoma alone. On that day, the warning program worked very well. Outlooks, watches, and warnings were issued well in advance. They were communicated rapidly. People knew safety rules and responded properly. The media, emergency managers, local officials, volunteer spotter groups, and state agencies worked in partnership to keep people informed.
Our experience on May 3rd leaves us both encouraged with the progress we have made and the lives we were able to save, but saddened at the loss of 42 lives. We need to continue to improve our understanding of tornadoes and tornadic thunderstorms. Better science will lead to more accurate warnings. Improved dissemination will ensure those at greatest risk get the life saving information they need in time to take appropriate action.
NOAA's commitment to apply focused research to the warning process is the key to a safer future for all our citizens. The National Weather Service (NWS) modernization is an example of this synergy between research and operational application. My remarks will briefly describe the May 3rd outbreak and outline how research enabled us to reach this point. I will conclude with additional research I believe necessary for forecast offices, such as Norman, to meet their obligations to provide the best possible warning service to the public.
Summary of May 3, 1999, Tornado Outbreak in Central Oklahoma.
On May 3, 1999, more than 70 tornadoes swept across the Central Plains. Violent tornadoes ravaged the southern portions of the Oklahoma City Metropolitan area and the Wichita, Kansas, area. On that day, more than 50 tornadoes occurred in Oklahoma alone. Killer tornadoes struck Dover, Shawnee, Perry, and Bridge Creek and the Moore and southern Oklahoma City metropolitan areas. This was the largest tornado outbreak ever recorded in Oklahoma.
The warning process worked extremely well. Even though a large and extremely violent tornado moved through densely populated sections of the southern Oklahoma City metropolitan area, the excellent warnings provided by NWS can be credited with the extraordinary low loss of life.
The Storm Prediction Center (SPC) first highlighted the threat of tornadoes at 1:00 a.m. At 11:30 a.m., SPC issued another outlook mentioning the possibility of strong and violent tornadoes. The first tornado watch for Oklahoma was issued by SPC at 4:30 p.m. Twenty five minutes later, the first tornado was reported near Lawton.
Our office issued 116 county warnings for tornadoes and severe thunderstorms. The average lead time for the tornado warnings was 18 minutes. For Moore and southern Oklahoma City, the lead time was 32 minutes. False alarm rates (i.e., crying wolf) for this outbreak were much lower than average.
During severe weather events, NWS offices call back staff on overtime to ensure services are delivered. On May 3rd, we tripled the operations component of our staff to handle the severe weather. Efforts include monitoring weather to issuing warnings, monitoring spotter reports, alerting emergency management officials, etc. This was a total team effort among all office staff.
Operational Warning Improvements Through Research
NOAA supports infusion of research results into forecast operations through many avenues. These efforts include Cooperative Institutes and Joint Institutes with universities and through smaller projects sponsored by the University Corporation for Atmospheric Research/Cooperative Program for Operational Meteorology, Education, and Training (COMET). Since its inception, the COMET Outreach Program has provided monetary support for more than 60 universities and is currently funding 30 active projects. NOAA-sponsored projects have investigated atmospheric phenomena ranging in scale from El Nino and climate prediction to severe thunderstorms, tornadoes, and flash flooding. This research has helped the NWS provide advanced notice to increase awareness ranging from long-term seasonal trends, to daily forecasts and severe weather outlooks, to hourly watches, to short-fused warnings.
The NWS modernization program, built upon the fruits of research, enabled the NWS to have the data and information necessary to forecast tornado occurrence. For example, in the 1980's, tornado and severe thunderstorm warnings were frequently based on reported severe weather. As a result of the modernization, nationwide tornado warning lead times more than doubled during the past 10 years, on average, from less than 4 minutes to more than 11 minutes. The lead time for tornadoes from the Norman office is 15 minutes, well above the national average. Furthermore, there are more frequent instances where tornado warnings are providing lead times in excess of 20 minutes.
Not only are lead times for tornadoes better, but accuracies (percent of tornadoes that are warned for) of the warnings also improved dramatically from about 30 percent to around 65 percent. For all severe storms, both tornadoes and severe thunderstorms, warning accuracies jumped from 57 percent in 1987 to 83 percent in 1998.
Ground-breaking research at the National Severe Storms Laboratory (NSSL) during the late l960's and early l970's led to the first Doppler weather radar detection of circulations within thunderstorms. NSSL scientists demonstrated that Doppler radar capabilities could detect these circulations, called mesocyclones, and other thunderstorm features which often lead to tornado development. Based on this research, Doppler radars were installed nationwide through a Department of Defense, Federal Aviation Administration, and Department of Commerce tri-agency initiative. These Doppler radars are known as NEXRAD (NEXt generation RADar).
Introduction of NEXRAD into operations brought tremendous capabilities to NWS forecasters. These Doppler radars are like CAT scans of the atmosphere enabling forecasters to see wind fields within thunderstorms. These thunderstorm wind fields are precursors to tornadoes. NEXRAD forms the foundation for all NWS severe weather warnings across the United States and is used in close concert with satellite observations and ground-truth supplied by spotter groups. NEXRAD has enabled us to predict tornadoes and where they are headed. In the past, all we could do was report tornadoes and guess where they were going.
In the l970's, NOAA recognized the trend of rapid advances in information systems and formed an Office of Oceanic and Atmospheric Research project to better understand how these advances could be applied to NOAA's weather forecasting and warning mission. The primary goal of this project was to rapidly transfer applicable new technologies and science into NWS operations. In 1988, this project evolved into NOAA's Forecast Systems Laboratory (FSL). Early on, FSL conducted a series of controlled experiments using interactive meteorological workstations which integrated most of the advanced data sets planned for NWS modernization. Operational NWS forecasters participated in these exercises, and the system continued to evolve based on their comments. By the late 1980's, prototype AWIPS (Advanced Weather Interactive Processing System) systems were placed in both the Norman and Denver NWS offices to gain further experience in operational environments and to further evolve the system. Based on experience with these prototype systems in Norman and Denver, FSL developed much of the software available in AWIPS.
An equally important component of new technology is AWIPS. NWS will complete nationwide AWIPS deployment by the end of this month. AWIPS integrates for forecasters all data sets from observing and prediction systems. These data sets include satellites, NEXRAD, Automated Surface Observing Systems, profilers, mesonets, and numerical models. Forecasters typically had to visit six different data sources to develop forecasts and warnings, and perhaps another one or two to compose and issue the message. Now, all data and information are available right at the forecaster's fingertips.
The Norman Forecast Office is on the University of Oklahoma's (OU) north campus, collocated with NOAA's Storm Prediction Center, NOAA's National Severe Storms Laboratory, and NOAA's NEXRAD Operational Support Facility. This collocation generates synergy between research and development done at NSSL and OU, as the latest theories are incorporated into operations by SPC and the Norman Forecast Office. Norman has long served as a test site for new technologies, such as NEXRAD and AWIPS, and currently is testing the Warning Decision Support System (WDSS).
WDSS, developed at NSSL and being tested in Norman, offers access to more data from NEXRAD radars than is available in current NWS operational systems. WDSS provides improved computer software to automatically rank storms according to their software- derived severe weather threats. It enables forecasters to view the intensity of storm signatures without the limitations of current NWS operational systems. During the May 3rd event, changeable menus and tables in WDSS, along with radar displays in AWIPS, allowed us to maintain an overview of the outbreak. WDSS also enhanced forecaster confidence that no severe storms were being overlooked. We tracked tornado signatures with exceptionally high velocity values greater than 210 kts. This increased our confidence to issue strongly worded warnings for storms that occurred in rural areas after dark.
System enhancements, such as those in WDSS, are planned for AWIPS. The System for Convection Analysis and Nowcasting (SCAN) will bring the most recent scientific advances and use of weather radar to AWIPS workstations in every NWS office. However, complete SCAN capabilities depend on planned computer upgrades to the NEXRAD. The NEXRAD system is still using computers and other radar hardware that were delivered with the original units, designed in the 1980's. These components are saturated with existing scientific applications. The Open Systems Architecture program is upgrading the NEXRAD system with components and software that are based on industry standards, commercially available off the shelf, and provide a cost-effective path to future upgrades. The new components will allow a next generation of analyses and techniques, such as those in SCAN, to be implemented nationwide.
NEXRAD and AWIPS were critical to the success of the May 3rd warning process. Those of us who have been issuing warnings since the 1970's or even the 1980's know we could not have conducted such an efficient warning operation without NEXRAD and AWIPS. Our six AWIPS workstations and the additional WDSS workstation allowed us to keep up with data from three NEXRADs and issue warnings and severe weather statements for multiple tornadoes over multiple areas simultaneously. At one point, four of the six workstations were being used simultaneously for warning and severe weather statement preparation.
Future Research to Support Operational Warning Requirements
Earlier I mentioned our improved warning accuracies and lead times. While our false alarm rate for this event was well below the national average, there is much work needed to lower the false alarm rate even further. We need to be able to identify which supercell thunderstorms will produce tornadoes, which will be capable of producing multiple tornadoes, and which will not produce tornadoes at all. Today, with our present technologies and scientific understanding, tornadic and non-tornadic supercell thunderstorms often look alike.
One research effort to bring a better understanding of tornado formation is Verification of the Origins of Rotation in Tornadoes Experiment (VORTEX). Hosted by the NSSL and the University of Oklahoma's Center for the Analysis and Prediction of Storms, collaborators from around the United States and Canada were involved in the VORTEX experiment. VORTEX's goal was to obtain comprehensive data sets to evaluate numerous hypotheses about tornado formation. The data sets obtained by VORTEX are several orders of magnitude richer by comparison than any collected before on a tornadic storm. This information will provide invaluable knowledge to help reduce false alarms, improve predictions of supercell storms, and determine what causes storms to become tornadic.
Next, we need research to assist us in issuing better warnings for severe thunderstorms and heavy rainfall that leads to flash flooding. We need help in discerning which thunderstorms will be hail producers, what the size of the hail will be, or whether the storm will, instead, be a heavy rainfall producer.
As part of the NEXRAD Product Improvement Plan, a new capability planned for NEXRAD radars involves using dual-polarization. This upgrade can only be done once the planned computer upgrades for NEXRAD I mentioned earlier are complete. Dual polarization is a radar signal transmission, reception, and processing scheme that provides information on the relative height and width of precipitation targets. Currently, NEXRAD radars emit a signal in only one plane. Research at NSSL and the National Center for Atmospheric Research over the past 20 years indicates emitting signals in two planes will have significant operational benefits to NWS, the Federal Aviation Administration, and the military. Dual polarized radars make it easier to discriminate between hail and rain as well as between rain and snow. Furthermore, dual polarization can help determine hail size, identify electrically active storms (killer lightning producers), better estimate rainfall amounts in order to improve flash flood warning capabilities, and distinguish between biological scatterers (birds vs. insects, the former being a hazard to aircraft).
Finally, on the frontiers of tornado forecasting, we should set our goals at being able to define when a tornadic supercell will form, when it will produce a tornado, how powerful that tornado will be, and how long it will remain on the ground. It is a tall order, but one that our citizens should expect of a warning service.
Conclusion
Scientific advancements and introduction of new technologies which fueled the NWS modernization have had a profound impact on our ability to predict and warn for extreme weather events. The May 3rd tornado outbreak illustrated the progress this Nation has made in protecting our citizens. Accurate and timely warnings were issued by the NWS, broadcast by the media, acted on by emergency managers, and responded to by the public. The partnership between NWS and the media was especially effective on May 3rd, with each of us doing what we do best.
The May 3rd event also helped identify what more can be accomplished. Focused research will lead to better knowledge of tornadic storms. From better science will flow better operational procedures to anticipate, predict, and warn for tornadoes and other extreme weather phenomena.
We in the Nation's 121 NWS Forecast Offices stand ready to continue improving our ability to deliver critical lifesaving information to our citizens.