• Baer, V. E., 1991: The transition from the present radar dissemination system to the NEXRAD Information Dissemination Service (NIDS). Bull. Amer. Meteor. Soc., 72 , 2933.

    • Search Google Scholar
    • Export Citation
  • Brock, F. V., , Crawford K. C. , , Elliott R. L. , , Cuperus G. W. , , Stadler S. J. , , Johnson H. L. , , and Eilts M. D. , 1995: The Oklahoma Mesonet: A technical overview. J. Atmos. Oceanic Technol., 12 , 519.

    • Search Google Scholar
    • Export Citation
  • Crawford, K. C., , Morris D. A. , , McPherson R. A. , , Johnson H. L. , , Shafer M. A. , , Wolfinbarger J. M. , , and Hughes T. W. , 1998: OK-FIRST: A decision-support system for public-safety agencies. Preprints, 14th Int. Conf. on Interactive Information and Processing Systems for Meteorology, Oceanography, and Hydrology, Phoenix, AZ, Amer. Meteor. Soc., 5–9.

    • Search Google Scholar
    • Export Citation
  • Crawford, K. C., , and Lewis D. E. , 1999: OK-FIRST: A NIDS success story. Preprints, 15th Int. Conf. on Interactive Information and Processing Systems for Meteorology, Oceanography, and Hydrology, Dallas, TX, Amer. Meteor. Soc., 44–47.

    • Search Google Scholar
    • Export Citation
  • Doswell, C. A,I. I. I., , Moller A. R. , , and Brooks H. E. , 1999: Storm spotting and public awareness since the first tornado forecasts of 1948. Wea. Forecasting, 14 , 544557.

    • Search Google Scholar
    • Export Citation
  • James, T. E., , Long P. O. , , and Shafer M. A. , 2000: An independent evaluation of the OK-FIRST decision-support system. Preprints, Second Symp. on Environmental Applications, Long Beach, CA, Amer. Meteor. Soc., 40–47.

    • Search Google Scholar
    • Export Citation
  • Klazura, G. E., , and Imy D. A. , 1993: A description of the initial set of analysis products available from the NEXRAD WSR-88D system. Bull. Amer. Meteor. Soc., 74 , 12931311.

    • Search Google Scholar
    • Export Citation
  • McPherson, R. A., , and Crawford K. C. , 1996: The EARTHSTORM project: Encouraging the use of real-time data from the Oklahoma Mesonet in K–12 classrooms. Bull. Amer. Meteor. Soc., 77 , 749761.

    • Search Google Scholar
    • Export Citation
  • Morris, D. A., 1998: The design and implementation of a Web-based decision-support system for public-safety agencies. Preprints, 14th Int. Conf. on Interactive Information and Processing Systems for Meteorology, Oceanography, and Hydrology, Phoenix, AZ, Amer. Meteor. Soc., 10–14.

    • Search Google Scholar
    • Export Citation
  • Morris, D. A., , and Duvall C. , 1999: An evaluation of the use of real-time weather data by public-safety agencies,. Preprints, 15th Int. Conf. on Interactive Information and Processing Systems for Meteorology, Oceanography, and Hydrology, Dallas, TX, Amer. Meteor. Soc., 380–383.

    • Search Google Scholar
    • Export Citation
  • Morris, D. A., , Lewis J. , , Rowe L. , , King D. , , Honigsberg M. , , and Conaghan C. , 1999: OK-FIRST in the field: How public-safety agencies use and apply weather data,. Preprints, 15th Int. Conf. on Interactive Information and Processing Systems for Meteorology, Oceanography, and Hydrology, Dallas, TX, Amer. Meteor. Soc., 384–387.

    • Search Google Scholar
    • Export Citation
  • Morris, D. A., , Crawford K. C. , , Johnson H. L. , , McPherson R. A. , , Shafer M. A. , , and Wolfinbarger J. M. , 2000: OK-FIRST: An innovative information-support system for public-safety agencies. Preprints, Second Symp. on Environmental Applications, Long Beach, CA, Amer. Meteor. Soc., 32–37.

    • Search Google Scholar
    • Export Citation
  • Morris, D. A., , Kloesel K. A. , , and Wolfinbarger J. M. , 2001: OK-FIRST: A meteorological information system for public safety. Bull. Amer. Meteor. Soc., 82 , 19111923.

    • Search Google Scholar
    • Export Citation
  • National Research Council, 1980: Technological and scientific opportunities for improved weather and hydrological services in the coming decade. National Academy Press, 87 pp.

    • Search Google Scholar
    • Export Citation
  • NOAA, 1977: Kansas City flash flood of September 12–13, 1977. Natural Disaster Survey Rep. 77-2, 49 pp.

  • NOAA, 1980: Red River valley tornadoes of April 10, 1979. Natural Disaster Survey Rep. 80-1, 60 pp.

  • NOAA, 1991: Shadyside, Ohio, flash floods June 14, 1990. Natural Disaster Survey Rep., 27 pp. [NTIS PB91-204966INZ.].

  • NOAA, 1994: Southeastern United States Palm Sunday tornado outbreak of March 27, 1994. Natural Disaster Survey Rep., 75 pp. [Available from National Weather Service, 1325 East–West Highway, W/OS52, Silver Spring, MD 20910-3283.].

    • Search Google Scholar
    • Export Citation
  • NOAA, 1998a: The central Texas tornadoes of May 27, 1997. Service Assessment, 99 pp. [Available from National Weather Service, 1325 East–West Highway, Silver Spring, MD 20910-3283.].

    • Search Google Scholar
    • Export Citation
  • NOAA, 1998b: Central Florida tornado outbreak of February 22–23, 1998. Service Assessment, 37 pp. [Available from National Weather Service, 1325 East–West Highway, Silver Spring, MD 20910-3283.].

    • Search Google Scholar
    • Export Citation
  • NOAA, 1999: Oklahoma/southern Kansas tornado outbreak of May 3, 1999. Service Assessment, 51 pp. [Available from National Weather Service, 1325 East–West Highway, W/OS52, Silver Spring, MD 20910-3283.].

    • Search Google Scholar
    • Export Citation
  • Thompson, R. L., , and Edwards R. , 2000: An overview of the environmental conditions and forecast implications of the 3 May 1999 tornado outbreak. Wea. Forecasting, 15 , 682699.

    • Search Google Scholar
    • Export Citation
  • Winston, H. A., , and Ruthi L. J. , 1986: Evaluation of RADAP II severe storm detection algorithms. Bull. Amer. Meteor. Soc., 67 , 145150.

    • Search Google Scholar
    • Export Citation
  • Wolfinbarger, J. M., , Young R. A. , , and Stanley T. B. , 1998a: Interactive software for viewing NEXRAD level 3 data on the World Wide Web. Preprints, 14th Int. Conf. on Interactive Information and Processing Systems for Meteorology, Oceanography, and Hydrology, Phoenix, AZ, Amer. Meteor. Soc., 208–212.

    • Search Google Scholar
    • Export Citation
  • Wolfinbarger, J. M., , Young R. A. , , and Stanley T. B. , 1998b: Delivering real-time interactive data from the Oklahoma Mesonet via the World Wide Web. Preprints, 14th Int. Conf. on Interactive Information and Processing Systems for Meteorology, Oceanography, and Hydrology, Phoenix, AZ, Amer. Meteor. Soc., 213–217.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    Communities served by the OK-FIRST program as of Aug 2001

  • View in gallery

    Approximate damage paths and Fujita-damage-scale ratings for several of the tornadoes in central OK on 3 May 1999. Gray circles locate emergency management offices discussed in the text. Numbered circles denote areas impacted by decisions based on OK-FIRST information: 1) Chickasha Municipal Airport, 2) town of Stroud and Tanger Outlet Mall, 3) town of Dover, 4) town of Mulhall, and 5) I-40 highway exit near Shawnee. Adapted from NOAA (1999)

  • View in gallery

    Radar product usage by OK-FIRST participants during the period from 1800 UTC 3 May 1999 through 1800 UTC 4 May 1999

  • View in gallery

    Base reflectivity from the Oklahoma City (Twin Lakes; KTLX) WSR-88D at 1755 LT, indicating a hook echo near the town of Chickasha. Emergency Management Director S. Chapman used this image to decide to evacuate the airport, which is located a few miles northwest of Chickasha

  • View in gallery

    Base radial velocity from the Oklahoma City WSR-88D at 2215 LT. A couplet of inbound and outbound velocities (with respect to the radar location southwest of the couplet) was located between the towns of Chandler and Stroud, approaching the Tanger Outlet Mall. The mall was located at the highway exit on I-44 near Stroud

  • View in gallery

    (left) Base reflectivity and (right) vertically integrated liquid from the Oklahoma City WSR-88D at 2057 LT. Emergency Management Director D. Mastalka used these images to note that a tornadic supercell was west of Kingfisher and headed for the town of Dover (between Kingfisher and Hennessey) and to note the hail potential of the storm. The largest values of vertically integrated liquid [a reliable indicator of hail (Winston and Ruthi 1986)] were also located west of Kingfisher, indicating a strong updraft with suspended precipitation particles

  • View in gallery

    Storm-relative radial velocity images from the Oklahoma City WSR-88D at 2217, 2312, and 2340 LT. Circled areas were added for clarity in this article to highlight tornadic supercell thunderstorms (i.e., radial velocity couplets) in Logan County (indicated by heavy solid line). Light thin lines denote various highways. The circled storm at 2217 LT first hit the Crescent area and subsequently hit the town of Mulhall. Crescent was sandwiched between two tornadoes at 2312 LT that followed similar tracks. The southwestern storm at 2312 LT (just entering Logan County) was the same storm that was near the Mulhall area at 2342 LT. These images were guidance for the decisions made by Logan County emergency management personnel as discussed in the text

  • View in gallery

    OK-FIRST images used by Seminole County emergency management to warn emergency vehicles en route to the Oklahoma City area. (left) A hook echo in base reflectivity from the Oklahoma City WSR-88D at 2107 LT crosses I-40; (right) cursor readout of more-detailed storm attribute information

  • View in gallery

    Counties in OK declared disaster areas by U.S. President Clinton and the Federal Emergency Management Agency for the severe weather that occurred on 3–4 May 1999. The declarations were made because the amount of damage in these counties was beyond the ability of local governments to respond and provide help to individuals without insurance or to repair public infrastructure (i.e., public assistance)

  • View in gallery

    (left) The 3-h rainfall accumulation from the Springfield WSR-88D radar at 0413 LT, indicating more than 3 in. in Ottawa County near Miami; (right) total rainfall accumulations through 0430 LT from the Oklahoma Mesonet. Corresponding flash-flood guidance values for Craig and Ottawa Counties were 2.2 in

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 0 0 0
PDF Downloads 0 0 0

OK-FIRST: An Example of Successful Collaboration between the Meteorological and Emergency Response Communities on 3 May 1999

View More View Less
  • 1 Oklahoma Climatological Survey, The University of Oklahoma, Norman, Oklahoma
  • 2 Department of Emergency Management and Communications, City of Moore, Moore, Oklahoma
© Get Permissions
Full access

Abstract

The Oklahoma Climatological Survey began an initiative known as Oklahoma's First-Response Information System using Telecommunications (OK-FIRST) in 1996 to support local public-safety agencies (fire departments, law enforcement agencies, and emergency management). OK-FIRST provides participant agencies with a wealth of real-time weather information and instruction for these users on the interpretation and application of the data in their operations. The 3 May 1999 tornado outbreak was a critical test for the OK-FIRST system. Users of OK-FIRST prevented the tragedy from becoming worse by making informed decisions that protected life and property.

Corresponding author address: Dale A. Morris, Oklahoma Climatological Survey, 100 E. Boyd, Suite 1210, Norman, OK 73019. Email: dmorris@ou.edu

Abstract

The Oklahoma Climatological Survey began an initiative known as Oklahoma's First-Response Information System using Telecommunications (OK-FIRST) in 1996 to support local public-safety agencies (fire departments, law enforcement agencies, and emergency management). OK-FIRST provides participant agencies with a wealth of real-time weather information and instruction for these users on the interpretation and application of the data in their operations. The 3 May 1999 tornado outbreak was a critical test for the OK-FIRST system. Users of OK-FIRST prevented the tragedy from becoming worse by making informed decisions that protected life and property.

Corresponding author address: Dale A. Morris, Oklahoma Climatological Survey, 100 E. Boyd, Suite 1210, Norman, OK 73019. Email: dmorris@ou.edu

1. Introduction

The Oklahoma's First-Response Information Resource System using Telecommunications (OK-FIRST) program was developed, beginning in October of 1996, as a formal educational outreach program of the Oklahoma Climatological Survey (Morris et al. 2001). The goal of OK-FIRST was to develop a transportable, agency-driven information support system to help public-safety agencies harness the information age. The goal of OK-FIRST continues to be improvement in how public-safety agencies (fire, police, and emergency management) mitigate and respond to weather emergencies.

Today, over four years later, more than 100 public-safety agencies (Fig. 1)—in support of their respective missions—have received formal training in how to access and use many new forms of environmental information via OK-FIRST [e.g., data from the Oklahoma Mesonetwork, volume-scan data from 15 Weather Surveillance Radar-1988 Doppler (WSR-88D) units, and other data from the modernized National Weather Service (NWS)]. By design, most communities served by OK-FIRST represent rural areas of Oklahoma (e.g., 52% have populations of 5000 or less). As a result of having been “modernized” by OK-FIRST, public-safety agencies across Oklahoma have achieved numerous success stories from the application of OK-FIRST data.

The most revealing testimonials about the effectiveness and robustness of OK-FIRST occurred on 3 May 1999—a day of unparalleled killer tornadoes that affected central and northern Oklahoma (Fig. 2). Because the meteorological community of Oklahoma (including the NWS and the broadcast media) performed superbly in handling this major outbreak, the death and injury toll was amazingly limited to 45 fatalities and 600 injuries. This tornado outbreak was responsible for damage to, or destruction of, nearly 10 000 houses and buildings across the state. The impact of the storms on several rural towns was immense. The town of Mulhall lost most of its buildings, including its churches, its only school, and the post office. In a matter of minutes, one tornado eliminated over 50% of the tax revenue for the town of Stroud by destroying the town's three major employers—who decided not to rebuild their facilities.

However, as major media outlets properly focused on the widespread death and destruction across heavily populated central Oklahoma, the OK-FIRST system passed a critical test. OK-FIRST servers shared over 25 000 files of WSR-88D information with public-safety users on 3 May 1999 (Fig. 3). As a result, many significant, life-saving success stories from rural Oklahoma, which did not make the national headlines, provided convincing evidence that OK-FIRST played an important role in saving the lives of many Oklahomans on 3 May 1999. The purpose of this manuscript is to share some of these “untold stories.”

2. Historical context

During the recent past, Oklahomans have contended with lengthy droughts, severe flooding, bouts of damaging thunderstorms, and significant tornado outbreaks. Even in “more normal” periods, “nonsevere” weather regularly affects human activities such as outdoor entertainment events and outside work crews. In addition, responses to wildfires, hazardous materials incidents, and acts of terrorism are impacted by environmental conditions. Before OK-FIRST, Oklahoma was a microcosm of the entire country in how agencies responded to emergencies in that local decision-support systems generally suffered from a near-complete lack of current and relevant environmental information. With OK-FIRST, Oklahomans have made great strides in local responses to weather-impacted emergencies; this paper documents some of this progress.

Over the past four decades, much evidence has accumulated to suggest that the NWS was disconnected from outside agencies because of outmoded dissemination policies. Access to NWS information by local officials nationwide was cumbersome, expensive, nonintuitive, and lacked critical details. In addition, the NWS occasionally did not receive critical storm or flood reports from the local level and thus could not produce appropriate warnings. Morris et al. (2000) documented 30 years of this evidence. In most recent times, this evidence has included the following.

  • Although the Wichita Falls and Oklahoma City NWS offices provided “excellent warning services” during the 10 April 1979 tornado outbreak, over 40 people perished in Wichita Falls. Many of the deaths resulted from people who were caught in automobiles (NOAA 1980). In the Kansas City flash flood of 12–13 September 1977, 17 deaths were car related (NOAA 1977). Even today, weather conditions are rarely transmitted automatically to automobiles although technological advances permit it.
  • A National Research Council report stated that, “for many years, the National Weather Service … operated on the assumption that if they produced a good product, someone would come to get it and use it. … Users are currently left largely to their own devices in determining what is available and how to use it; many are unaware of the information available.” (National Research Council 1980).
  • In June of 1990, a localized area of rain with amounts between 3 and 4 in. occurred in a 45-min period upstream of Shadyside, Ohio. This flood caused 26 fatalities and the destruction of 80 residences. The NWS issued a flash-flood watch for the affected county, but no flash-flood warning was issued because the NWS received no information indicating that a flood was imminent. The NWS did not receive real-time flood information from local officials; they did not find out about the flood until 4 h after the peak of the flood. In addition, the flash-flood watch only reached Shadyside officials through the broadcast media. County officials received notification of the flash-flood watch but did not relay this information to local officials in Shadyside (NOAA 1991).
  • Although local NWS offices issued tornado warnings with adequate lead time, 42 lives (20 at one church) were lost during the 27 March 1994 Palm Sunday tornado outbreak. Though warnings were disseminated using National Oceanic and Atmospheric Administration (NOAA) weather radio, local officials and citizens at risk did not receive notification because of the “limited resources many rural county emergency managers and law enforcement officials had at their disposal for receiving the emergency messages and enacting their response plans” (NOAA 1994).

The $4.5 billion modernization of the NWS during the 1990s exacerbated this data–telecommunications problem by producing vast amounts of high-quality, county-scale information with no viable delivery mechanism to those ultimately responsible for making life-and-death decisions. The NWS has deployed systems such as NOAA weather radio and the Emergency Manager's Weather Information Network (EMWIN); both systems have somewhat limited geographical coverage. Neither system was designed to provide real-time graphical information; they assist in wide dissemination of NWS warning decisions. Other NWS outreach activities coordinated by their warning coordination meteorologists, such as spotter training seminars, help in maintaining dialogue with emergency managers and form an important link in the ability of the NWS to provide quality warnings. Nevertheless, very few systems to disseminate high-quality, real-time, graphical products and designed for the public-safety community have ever been developed. In addition, rural areas, traditionally underserved by telecommunications and technology, have been at especially high risk. As a consequence, local officials made weather-impacted decisions without adequate information (e.g., storm spotters were deployed precariously because coordinators lacked information about storm location, movement, and intensity).

One major component of the NWS modernization was the deployment of an advanced, nationwide Doppler radar network, known as Next-Generation Weather Radar (NEXRAD), or WSR-88D. Local officials around the country expected to reap many of the benefits of the WSR-88D system because the network employed computer algorithms to detect significant weather phenomena, including rotational signatures, hail, and excessive rainfall, automatically. At the same time, the NWS outsourced the exclusive dissemination of this data among four private weather data vendors who provided the data to all entities outside the NWS. Known as the NEXRAD Information Dissemination Service (NIDS; Baer 1991; Klazura and Imy 1993), this arrangement was based upon management and logistical problems inherent in the radar-data distribution system used by the NWS during the pre-NEXRAD era; it was based upon federal policy from the 1980s that encouraged the privatization of government information services. Each of the NIDS vendors licensed the data from the NWS and constructed a variety of dissemination mechanisms that ranged from dial-up access to satellite delivery. The vendors also maintained dedicated dial-up connections to each of 154 radars, provided various levels of customer service, and created “value-added” products such as regional mosaics. As a result, market forces and the cost of doing business made access to NEXRAD data prohibitively expensive for most public-safety and other government users nationwide. Yet, in Oklahoma, as a result of OK-FIRST, rural users statewide have routinely accessed the NIDS data stream through a public–private partnership with one of the NIDS vendors; this situation is a NIDS success story (Crawford et al. 1999).

Meanwhile, the NWS was working to improve its infrastructure to take advantage of modern computer workstations. Despite these improvements, which dramatically improved forecasting and warning capabilities and produced tremendous data resources within the federal government, many of the associated benefits were beyond the financial reach of local public-safety officials. Time after time during the past 15–20 years, the NWS made correct decisions involving the issuance of severe weather warnings based upon timely and modern information. Yet, because adequate dissemination systems designed for local officials did not exist, decisions affecting the protection of citizens frequently were made without the benefit of critical and local information.

The modernization of the NWS did not necessarily fix the problem of telecommunication of data to public-safety officials. A postevent assessment after the 27 May 1997 outbreak of tornadoes in central Texas recommended that the NWS should intensify efforts on the local and state levels “to explore alternative methods of communicating critical weather products to emergency management officials” (NOAA 1998a). A similar recommendation followed the 22–23 February 1998 tornado outbreak in central Florida: the NWS should “work with local emergency management personnel to develop new or enhanced mass communication systems” (NOAA 1998b).

The above references to failures in communications between the NWS and the public-safety community illustrate particular instances in both the pre- and postmodernization eras in which improvements could be made. Of course, not every NWS office suffers from this malady. In fact, the Norman NWS Forecast Office (FO) was commended by the service assessment team (NOAA 1999) assembled after the 3 May 1999 event for “a well-trained and widespread spotter network;” effective dissemination of NWS warnings, forecasts, and statements via several mechanisms including EMWIN and OK-FIRST; strong partnerships with local emergency managers that “eased the communication process during the severe weather outbreak;” and “amateur radio repeaters [that] helped keep the Norman NWSFO in contact with spotters and Emergency Operations Centers throughout the far reaching counties in its county warning area.”

This same report also cited OK-FIRST as a benefit during the event and recommended that the NWS provide radar products to partners and users in real time.

3. The OK-FIRST system

The OK-FIRST “information support system,” introduced by Crawford et al. (1998), was intended to fill a recognized “service void” in Oklahoma's weather-warning system by building information bridges between the modernized NWS and the critical need for information in the hands of nonmeteorologists in rural areas during emergencies. It was built upon successes in implementing the Oklahoma Mesonetwork (Brock et al. 1995) and its Kindergarten–12th grade (K–12) educational outreach program known as “EARTHSTORM” (McPherson and Crawford 1996).

Critical design decisions included evolving OK-FIRST into a “Web-based” decision support system (Morris 1998) built around Internet browsers, custom “plug-in” software (Wolfinbarger 1998a,b), and extensive feedback from front-line users (Morris et al. 1999). The explosion of widespread Internet access during the last half of the 1990s provided a cost-effective and convenient dissemination mechanism. Yet, mere access to modern weather information was not the total solution to the data-void problem for public-safety users; OK-FIRST staff have supported these users through regular instructional workshops that maintain and improve data interpretation and software skills. The NWS in Oklahoma has participated in OK-FIRST in various ways from assisting with the workshops to providing wise counsel on operational and logistical issues. OK-FIRST has helped to improve the “two-way” warning process because both NWS and local personnel view the same imagery as they discuss storm attributes. Morris and Duvall (1999) provided a subjective evaluation, and James et al. (2000) provided an objective and independent evaluation of the impact that OK-FIRST was beginning to have. Several additional evaluations independently recognized OK-FIRST nationally and internationally. The John F. Kennedy School of Government at Harvard University recognized OK-FIRST as a winner in their Innovations in American Government program (Morris et al. 2000). An international jury placed OK-FIRST as a finalist in the 2001 Stockholm Challenge for innovative information technology projects.

OK-FIRST was designed to be one important component in the integrated warning system, described by Doswell et al. (1999), that provides redundant mechanisms for public-safety officials to receive emergency messages. Even so, feedback from OK-FIRST participants has suggested the usefulness of a system that conveniently provides graphical weather information even during situations that do not necessarily demand the attention of the NWS or the media (Morris et al. 2001). One consequence of OK-FIRST has been a reduction in the amount of communication between some local officials and the NWS. Part of the decrease results from a reduced need for verbal descriptions of radar images. A small number of users, however, have misinterpreted certain images (including algorithm output) without consulting with the NWS. To mitigate these concerns, a user certification program, which includes both initial training and ongoing refresher workshops, was implemented as a condition to use the system (Morris et al. 2001). During the workshops, the nonmeteorologist users are reminded that OK-FIRST allows them to anticipate NWS decisions, and that these users should not try to “out-weather-service the weather service.”

4. Local operational readiness achieved through OK-FIRST

Local emergency management (EM) programs typically consist of only one or two paid positions. Other offices are staffed on a volunteer basis. In both cases, the personnel often have a background in one of the other public-safety disciplines (i.e., law enforcement, fire and rescue, or emergency medical service). Although many of these personnel have a personal or professional interest in weather, very few have formal meteorological training, outside perhaps of annual severe weather spotting training provided by local NWS staff. Although these personnel need not necessarily possess a degree in meteorology to be effective in their jobs, many seem to be interested in obtaining lots of meteorological training because of the impact weather has on their jobs. In addition, local public-safety agencies historically have had no real-time weather data access. This lack of access can be tied to the lack of infrastructure to receive data in rural locations, the high cost of the infrastructure and data, and the lack of funding in the EM programs.

One source of weather data for the EM has been commercial television. It has not been uncommon in the past for an EM volunteer in a local emergency operations center during a severe weather episode to be given the task of switching between television channels in hopes of finding a 30-s live radar broadcast. Further, the EMs had to be content with the picture that was being displayed, with no control over the locations being displayed or the zoom level. For example, it is common practice for the Oklahoma City stations to zoom to Oklahoma County when severe weather is in progress, because this is the most-populated county in their viewing area. This does little to assist EMs in other areas.

In a modernized EM office, the manager typically begins the day by examining potential threats to the community for that day. In addition, the manager investigates other factors that might contribute to any incidents that occur. Weather potentially could be both a threat and a contributing factor. A typical self-briefing by an EM might start with the zone forecast, area forecast discussion, and hazardous weather outlooks. The briefing would continue by checking current observations and satellite and radar data to arrive at a threat level. During a day on which weather will not be a direct threat, an EM might monitor the weather on an “as-needed” basis, such as when a nonweather event such as a fire or hazardous materials incident occurs. During a day on which the weather is a potential threat, the modernized EM will more actively monitor observations, particularly wind, moisture, and temperature observations. The manager will also pay close attention to forecast updates, hazardous weather outlook updates, and radar and satellite observations—updating the morning briefing every few hours. Real-time weather monitoring becomes constant when the weather threatens. At that point, the EM is typically glued to radar products, correlating the radar data with information from weather spotters and the NWS.

One of the authors (Kitch) was among the first trained participants of OK-FIRST. He is the emergency management director for the city of Moore. The Moore 911/Emergency Operations Center (EOC) displays radar and/or Oklahoma Mesonet products on an around-the-clock basis. In addition, Kitch monitors products available via OK-FIRST from home. In this case, he received forecasts from the NWS and Storm Prediction Center (SPC) in Norman on Sunday, 2 May 1999, which foretold the possibility of severe weather. Thus, he placed this information in the “potential threat” category to be monitored.

On Monday morning, the thunderstorm outlook issued by NWS Norman and the day-1 severe weather outlook issued by the SPC indicated slight risk of severe thunderstorms. Public-safety officials are taught by OK-FIRST staff to monitor these products. In addition, the local official can analyze the underlying conditions to utilize better other weather data during the day. For example, by 1500 local time (LT), EMs throughout Oklahoma, including Kitch in Moore, were watching Oklahoma Mesonet products indicating that southerly winds were continuing to transport moisture into the eastern two-thirds of Oklahoma. The local emergency managers noticed dewpoint temperatures between 65° and 70°F, surface temperatures between 75° and 85°F, and a dryline in northwest Oklahoma. More details on the evolution of the synoptic-scale and mesoscale environments were provided by Thompson and Edwards (2000, hereinafter TE).

Because of these factors, Kitch and other EMs continued to watch satellite imagery and radar products from the WSR-88Ds located near Oklahoma City and Frederick. Kitch noted the convective development around 1600 LT southwest of Oklahoma City and the SPC issuance of a severe-weather outlook detailing a high risk of severe thunderstorms. Within an hour, the Oklahoma City television stations would be broadcasting live pictures of a tornado.

As part of their preparation-and-response mission, the local EM program typically has primary responsibility for local, community warning of impending emergencies and disasters. In areas prone to severe thunderstorms, community warning can be accomplished via several means, including outdoor warning sirens, cable television interrupt systems, paging and telephone notifications, and media outlets using the Emergency Alert System. The decision to warn is generally based upon information from field spotters and upon official warnings from the NWS. Therefore, it is critical that emergency managers have access to current weather data, particularly radar products, in order to manage properly the field spotters and to add to the intelligence necessary to make critical warning decisions.

On 3 May 1999, the Moore EOC used radar products extensively to track the tornadic storms that approached from the southwest. By 1800 LT, the storms had moved to near Chickasha, about 1 h of tornado travel time southwest of Moore. Based upon the radar products delivered by OK-FIRST and the ground truth provided by the live television broadcasts, the decision was made to mobilize the field spotters for Moore and fully activate the EOC. Information combined from spotters, radar data, television broadcasts, and amateur radio conversations with NWS personnel indicated that it would be reasonable to extrapolate the track of the storm (labeled A in TE), shown on television to be very large, directly into Moore. The decision to warn officially the city of Moore was made, and the warning sirens and cable television interrupt were activated at 1902 LT. The tornado—later rated at F5—would track through the northwestern quarter of Moore at 1925 LT, destroying more than 800 homes and causing five fatalities.

While the ultimate goal of all public-safety agencies is to eliminate fatalities and injuries, it is apparent that the mortality rate in Moore was extremely low. Kitch estimated that 3000 persons were in the direct path of the tornado. That there were only five fatalities is a testament to the effectiveness of the EM and weather-preparedness programs in Moore. OK-FIRST has become an integral part of the EM program in Oklahoma.

5. Testimonials arising from 3 May 1999

As a result of using the OK-FIRST system, many rural public-safety officials have become proactive rather than reactive when handling weather-impacted situations. Stellar examples of this new approach to their duties were revealed through life-saving actions that resulted from decisions made during the killer tornado outbreak of 3 May 1999—significant stories that did not receive widespread media attention. Following are some examples.

  • S. Chapman, emergency management director for the town of Chickasha (location 1 in Fig. 2), used pinpointed information from his OK-FIRST displays (resulting from combined NWS and NEXRAD information; see Fig. 4 and storm A in TE) to discern that the municipal airport was threatened by one of the first tornadoes of the day. He accordingly evacuated the Chickasha airport—a full 15 min before the tornado struck. No fatalities or injuries resulted.
  • Later that evening, when another tornado demolished an outlet shopping mall in Stroud (location 2 in Fig. 2; see Fig. 5 and storm D in TE), all stores had been vacated. B. Springfield, Lincoln County emergency management director, was provided frequent radar updates from OK-FIRST and notified Stroud 30 min in advance.
  • People in their homes in rural areas also were more secure thanks to the actions of emergency managers that day. After the storms had spun their path of destruction across the Oklahoma City area, they continued northeast. Residents in rural areas received minimal attention from the media, as local news focused much of their coverage upon the devastation and recovery operations in and near Oklahoma City. One of Springfield's assistants, who was monitoring the OK-FIRST radar displays, relayed updates every 5 min on radio frequencies received by scanner. Potential victims received the information and took shelter. Springfield later said that many of these people would otherwise not have taken shelter had it not been for the trustworthy information coming across their scanner. Convinced that OK-FIRST made a life-and-death difference, he stated, “I promise you we'd have had deaths in Lincoln County without it. I firmly believe that” (Daily Oklahoman, 11 May 1999).
  • In Kingfisher County (Fig. 6), the town of Dover (location 3 in Fig. 2 and storm E in TE) was hit hard, with two-thirds of the houses either destroyed or damaged. D. Mastalka, then director of Kingfisher County emergency management, ordered emergency vehicles (including those belonging to police, sheriff, and game warden officials) to traverse the streets to warn the town's citizens 10–20 min in advance of the storm. The lone fatality in Dover was an individual who chose not to take immediate action after receiving the warning.
  • Rescue workers themselves were targets of the storms. A tornado completely destroyed the small community of Mulhall in Logan County (location 4 in Fig. 2 and storms B and G in TE). Rescue workers set up a command center to manage the recovery operations. J. Lewis, Logan County emergency manager at the time, saw successive tornadoes following similar paths on his OK-FIRST system (Fig. 7). He alerted the command center to move their operations—twice. As a result, the rescue workers did not themselves become victims of the storms. In his letter of 10 May 1999, Lewis stated:

    When police and rescue crews arrived at the first Logan County damage site near the City of Crescent, one of the first tasks was to open the highway sufficiently to get an ambulance through from Crescent to the hospital in Guthrie. All efforts were to get that ambulance moving with a critically injured tornado victim. About the time they succeeded, a second tornado approached in the dark. The ambulance and the tornado moved on intersecting paths. Emergency management, aware of both events, was able to stop the ambulance until the tornado passed just in front of it. The town of Mulhall, devastated by the initial tornado after it passed Crescent, was warned primarily by two law enforcement units sounding their vehicle sirens in the town. The units had been dispatched there by the Sheriff's Office based upon OK-FIRST data. Both units continued warning residents until they were each hit by debris: one by power lines down across his car, the second by a large tree upon his unit. Both officers were uninjured—and so were all but one Mulhall town resident!

    Practically every structure in Mulhall, a community of 945 citizens, was destroyed, including the town's only water tower, which had stood since the 1920s.
  • In Seminole County, Emergency Management Director Herb Gunter radioed a warning to a caravan of emergency vehicles responding to the Oklahoma City area. Gunter noticed that another tornado was developing (Fig. 8; storm D in TE) and would cross an interstate highway ahead of them. The law enforcement convoy closed the highway (location 5 in Fig. 2) so that neither they nor other vehicles would drive into the storm.
  • In Garfield County (location 6 in Fig. 1), seven chase teams monitored four supercell storms, including one that damaged a farm. Perhaps more important, other rescue teams from Garfield County traveled to the Oklahoma City area to assist with the aftermath of the storm there. M. Honigsberg, director of Garfield County emergency management, provided periodic updates of OK-FIRST information to maneuver these rescue teams successfully and safely around the intervening Dover and Mulhall storms.
  • The damage that resulted from the hazardous weather on 3 May 1999 was not limited to the counties surrounding Oklahoma City (Fig. 9). In far northeast Oklahoma, as attention remained focused on the central Oklahoma tornadoes, heavy thunderstorms with flood-producing rains brought 5–6 in. of rain to Ottawa and surrounding counties on the night of 3 May (Fig. 10). T. Durborow, emergency management director of the city of Miami, used OK-FIRST to “help protect the public in a timely manner.”

These statements provide strong evidence that OK-FIRST played an important role in saving the lives of many Oklahomans on the night of 3 May 1999. In addition, because first responders themselves were also in grave danger, the use of OK-FIRST prevented even greater tragedies from occurring.

6. Summary

As a result of operational successes on 3 May 1999 and during many other situations, OK-FIRST has become a catalyst for change in many local governments. Local officials are now empowered to close bridges during floods, to save property in wildfires, to improve evacuations after hazardous spills, and to protect audiences at outdoor events. Other benefits include more efficient scheduling of public works projects and information for police and fire investigations. An independent evaluator concluded (James et al. 2000) that OK-FIRST changed the behavior of its graduates and their approach to decision making—for the better. In particular, the content of conversations between the EMs and the NWS during severe-weather operations has changed from asking basic questions about “what is going on” to specifically questioning particular data signatures.

The events of 3 May 1999 and the response to these events by local public-safety officials clearly indicated the utility of the OK-FIRST system. These officials have been modernized in their approach to decision making and were guided by high-quality information and training provided by OK-FIRST. Informed decisions were not made in a few isolated situations, but occurred repeatedly during the event. Thus, actions of local emergency personnel supported by OK-FIRST complemented the superb actions by the NWS and the Oklahoma City broadcast media. The result was a remarkable reduction of casualties from this storm.

REFERENCES

  • Baer, V. E., 1991: The transition from the present radar dissemination system to the NEXRAD Information Dissemination Service (NIDS). Bull. Amer. Meteor. Soc., 72 , 2933.

    • Search Google Scholar
    • Export Citation
  • Brock, F. V., , Crawford K. C. , , Elliott R. L. , , Cuperus G. W. , , Stadler S. J. , , Johnson H. L. , , and Eilts M. D. , 1995: The Oklahoma Mesonet: A technical overview. J. Atmos. Oceanic Technol., 12 , 519.

    • Search Google Scholar
    • Export Citation
  • Crawford, K. C., , Morris D. A. , , McPherson R. A. , , Johnson H. L. , , Shafer M. A. , , Wolfinbarger J. M. , , and Hughes T. W. , 1998: OK-FIRST: A decision-support system for public-safety agencies. Preprints, 14th Int. Conf. on Interactive Information and Processing Systems for Meteorology, Oceanography, and Hydrology, Phoenix, AZ, Amer. Meteor. Soc., 5–9.

    • Search Google Scholar
    • Export Citation
  • Crawford, K. C., , and Lewis D. E. , 1999: OK-FIRST: A NIDS success story. Preprints, 15th Int. Conf. on Interactive Information and Processing Systems for Meteorology, Oceanography, and Hydrology, Dallas, TX, Amer. Meteor. Soc., 44–47.

    • Search Google Scholar
    • Export Citation
  • Doswell, C. A,I. I. I., , Moller A. R. , , and Brooks H. E. , 1999: Storm spotting and public awareness since the first tornado forecasts of 1948. Wea. Forecasting, 14 , 544557.

    • Search Google Scholar
    • Export Citation
  • James, T. E., , Long P. O. , , and Shafer M. A. , 2000: An independent evaluation of the OK-FIRST decision-support system. Preprints, Second Symp. on Environmental Applications, Long Beach, CA, Amer. Meteor. Soc., 40–47.

    • Search Google Scholar
    • Export Citation
  • Klazura, G. E., , and Imy D. A. , 1993: A description of the initial set of analysis products available from the NEXRAD WSR-88D system. Bull. Amer. Meteor. Soc., 74 , 12931311.

    • Search Google Scholar
    • Export Citation
  • McPherson, R. A., , and Crawford K. C. , 1996: The EARTHSTORM project: Encouraging the use of real-time data from the Oklahoma Mesonet in K–12 classrooms. Bull. Amer. Meteor. Soc., 77 , 749761.

    • Search Google Scholar
    • Export Citation
  • Morris, D. A., 1998: The design and implementation of a Web-based decision-support system for public-safety agencies. Preprints, 14th Int. Conf. on Interactive Information and Processing Systems for Meteorology, Oceanography, and Hydrology, Phoenix, AZ, Amer. Meteor. Soc., 10–14.

    • Search Google Scholar
    • Export Citation
  • Morris, D. A., , and Duvall C. , 1999: An evaluation of the use of real-time weather data by public-safety agencies,. Preprints, 15th Int. Conf. on Interactive Information and Processing Systems for Meteorology, Oceanography, and Hydrology, Dallas, TX, Amer. Meteor. Soc., 380–383.

    • Search Google Scholar
    • Export Citation
  • Morris, D. A., , Lewis J. , , Rowe L. , , King D. , , Honigsberg M. , , and Conaghan C. , 1999: OK-FIRST in the field: How public-safety agencies use and apply weather data,. Preprints, 15th Int. Conf. on Interactive Information and Processing Systems for Meteorology, Oceanography, and Hydrology, Dallas, TX, Amer. Meteor. Soc., 384–387.

    • Search Google Scholar
    • Export Citation
  • Morris, D. A., , Crawford K. C. , , Johnson H. L. , , McPherson R. A. , , Shafer M. A. , , and Wolfinbarger J. M. , 2000: OK-FIRST: An innovative information-support system for public-safety agencies. Preprints, Second Symp. on Environmental Applications, Long Beach, CA, Amer. Meteor. Soc., 32–37.

    • Search Google Scholar
    • Export Citation
  • Morris, D. A., , Kloesel K. A. , , and Wolfinbarger J. M. , 2001: OK-FIRST: A meteorological information system for public safety. Bull. Amer. Meteor. Soc., 82 , 19111923.

    • Search Google Scholar
    • Export Citation
  • National Research Council, 1980: Technological and scientific opportunities for improved weather and hydrological services in the coming decade. National Academy Press, 87 pp.

    • Search Google Scholar
    • Export Citation
  • NOAA, 1977: Kansas City flash flood of September 12–13, 1977. Natural Disaster Survey Rep. 77-2, 49 pp.

  • NOAA, 1980: Red River valley tornadoes of April 10, 1979. Natural Disaster Survey Rep. 80-1, 60 pp.

  • NOAA, 1991: Shadyside, Ohio, flash floods June 14, 1990. Natural Disaster Survey Rep., 27 pp. [NTIS PB91-204966INZ.].

  • NOAA, 1994: Southeastern United States Palm Sunday tornado outbreak of March 27, 1994. Natural Disaster Survey Rep., 75 pp. [Available from National Weather Service, 1325 East–West Highway, W/OS52, Silver Spring, MD 20910-3283.].

    • Search Google Scholar
    • Export Citation
  • NOAA, 1998a: The central Texas tornadoes of May 27, 1997. Service Assessment, 99 pp. [Available from National Weather Service, 1325 East–West Highway, Silver Spring, MD 20910-3283.].

    • Search Google Scholar
    • Export Citation
  • NOAA, 1998b: Central Florida tornado outbreak of February 22–23, 1998. Service Assessment, 37 pp. [Available from National Weather Service, 1325 East–West Highway, Silver Spring, MD 20910-3283.].

    • Search Google Scholar
    • Export Citation
  • NOAA, 1999: Oklahoma/southern Kansas tornado outbreak of May 3, 1999. Service Assessment, 51 pp. [Available from National Weather Service, 1325 East–West Highway, W/OS52, Silver Spring, MD 20910-3283.].

    • Search Google Scholar
    • Export Citation
  • Thompson, R. L., , and Edwards R. , 2000: An overview of the environmental conditions and forecast implications of the 3 May 1999 tornado outbreak. Wea. Forecasting, 15 , 682699.

    • Search Google Scholar
    • Export Citation
  • Winston, H. A., , and Ruthi L. J. , 1986: Evaluation of RADAP II severe storm detection algorithms. Bull. Amer. Meteor. Soc., 67 , 145150.

    • Search Google Scholar
    • Export Citation
  • Wolfinbarger, J. M., , Young R. A. , , and Stanley T. B. , 1998a: Interactive software for viewing NEXRAD level 3 data on the World Wide Web. Preprints, 14th Int. Conf. on Interactive Information and Processing Systems for Meteorology, Oceanography, and Hydrology, Phoenix, AZ, Amer. Meteor. Soc., 208–212.

    • Search Google Scholar
    • Export Citation
  • Wolfinbarger, J. M., , Young R. A. , , and Stanley T. B. , 1998b: Delivering real-time interactive data from the Oklahoma Mesonet via the World Wide Web. Preprints, 14th Int. Conf. on Interactive Information and Processing Systems for Meteorology, Oceanography, and Hydrology, Phoenix, AZ, Amer. Meteor. Soc., 213–217.

    • Search Google Scholar
    • Export Citation
Fig. 1.
Fig. 1.

Communities served by the OK-FIRST program as of Aug 2001

Citation: Weather and Forecasting 17, 3; 10.1175/1520-0434(2002)017<0567:OFAEOS>2.0.CO;2

Fig. 2.
Fig. 2.

Approximate damage paths and Fujita-damage-scale ratings for several of the tornadoes in central OK on 3 May 1999. Gray circles locate emergency management offices discussed in the text. Numbered circles denote areas impacted by decisions based on OK-FIRST information: 1) Chickasha Municipal Airport, 2) town of Stroud and Tanger Outlet Mall, 3) town of Dover, 4) town of Mulhall, and 5) I-40 highway exit near Shawnee. Adapted from NOAA (1999)

Citation: Weather and Forecasting 17, 3; 10.1175/1520-0434(2002)017<0567:OFAEOS>2.0.CO;2

Fig. 3.
Fig. 3.

Radar product usage by OK-FIRST participants during the period from 1800 UTC 3 May 1999 through 1800 UTC 4 May 1999

Citation: Weather and Forecasting 17, 3; 10.1175/1520-0434(2002)017<0567:OFAEOS>2.0.CO;2

Fig. 4.
Fig. 4.

Base reflectivity from the Oklahoma City (Twin Lakes; KTLX) WSR-88D at 1755 LT, indicating a hook echo near the town of Chickasha. Emergency Management Director S. Chapman used this image to decide to evacuate the airport, which is located a few miles northwest of Chickasha

Citation: Weather and Forecasting 17, 3; 10.1175/1520-0434(2002)017<0567:OFAEOS>2.0.CO;2

Fig. 5.
Fig. 5.

Base radial velocity from the Oklahoma City WSR-88D at 2215 LT. A couplet of inbound and outbound velocities (with respect to the radar location southwest of the couplet) was located between the towns of Chandler and Stroud, approaching the Tanger Outlet Mall. The mall was located at the highway exit on I-44 near Stroud

Citation: Weather and Forecasting 17, 3; 10.1175/1520-0434(2002)017<0567:OFAEOS>2.0.CO;2

Fig. 6.
Fig. 6.

(left) Base reflectivity and (right) vertically integrated liquid from the Oklahoma City WSR-88D at 2057 LT. Emergency Management Director D. Mastalka used these images to note that a tornadic supercell was west of Kingfisher and headed for the town of Dover (between Kingfisher and Hennessey) and to note the hail potential of the storm. The largest values of vertically integrated liquid [a reliable indicator of hail (Winston and Ruthi 1986)] were also located west of Kingfisher, indicating a strong updraft with suspended precipitation particles

Citation: Weather and Forecasting 17, 3; 10.1175/1520-0434(2002)017<0567:OFAEOS>2.0.CO;2

Fig. 7.
Fig. 7.

Storm-relative radial velocity images from the Oklahoma City WSR-88D at 2217, 2312, and 2340 LT. Circled areas were added for clarity in this article to highlight tornadic supercell thunderstorms (i.e., radial velocity couplets) in Logan County (indicated by heavy solid line). Light thin lines denote various highways. The circled storm at 2217 LT first hit the Crescent area and subsequently hit the town of Mulhall. Crescent was sandwiched between two tornadoes at 2312 LT that followed similar tracks. The southwestern storm at 2312 LT (just entering Logan County) was the same storm that was near the Mulhall area at 2342 LT. These images were guidance for the decisions made by Logan County emergency management personnel as discussed in the text

Citation: Weather and Forecasting 17, 3; 10.1175/1520-0434(2002)017<0567:OFAEOS>2.0.CO;2

Fig. 8.
Fig. 8.

OK-FIRST images used by Seminole County emergency management to warn emergency vehicles en route to the Oklahoma City area. (left) A hook echo in base reflectivity from the Oklahoma City WSR-88D at 2107 LT crosses I-40; (right) cursor readout of more-detailed storm attribute information

Citation: Weather and Forecasting 17, 3; 10.1175/1520-0434(2002)017<0567:OFAEOS>2.0.CO;2

Fig. 9.
Fig. 9.

Counties in OK declared disaster areas by U.S. President Clinton and the Federal Emergency Management Agency for the severe weather that occurred on 3–4 May 1999. The declarations were made because the amount of damage in these counties was beyond the ability of local governments to respond and provide help to individuals without insurance or to repair public infrastructure (i.e., public assistance)

Citation: Weather and Forecasting 17, 3; 10.1175/1520-0434(2002)017<0567:OFAEOS>2.0.CO;2

Fig. 10.
Fig. 10.

(left) The 3-h rainfall accumulation from the Springfield WSR-88D radar at 0413 LT, indicating more than 3 in. in Ottawa County near Miami; (right) total rainfall accumulations through 0430 LT from the Oklahoma Mesonet. Corresponding flash-flood guidance values for Craig and Ottawa Counties were 2.2 in

Citation: Weather and Forecasting 17, 3; 10.1175/1520-0434(2002)017<0567:OFAEOS>2.0.CO;2

Save