2
Pre-Modernization Environment and Planning
This chapter focuses on the state of the National Weather Service (NWS) in the 1980s, prior to the official start of the Modernization and Associated Restructuring (MAR) in 1989. During the period preceding the MAR, improved radar and other observation systems were already under development, the numerical weather prediction operations at the National Meteorological Center (NMC) were improving steadily, and the operational application of data and information from both polar orbiting and geostationary satellites had become a critical component of atmospheric observation and improved forecasting capability. However, the NWS could not fully realize the benefits of these rapidly evolving technological improvements within their existing organizational structure, staffing, and physical infrastructure. The MAR execution objectives were to address this problem, yielding several promised benefits.
PRE-MODERNIZATION WEATHER SERVICE
In the 1980s, surface observations were being made manually, and were often inconsistent between observers and locations. Forecaster workstations, themselves a fairly recent innovation, operated across multiple computing systems, all with limited computational capability. The NWS radar network was composed of three different types of radars that could determine echo structure and intensity, important for tornado detection and forecasts, but had no capability to measure wind speeds; there were significant gaps in coverage, particularly in the West. The field office structure with approximately only one WSFO per state limited relationships between forecasters and local communities, especially in states with large populations and multiple media markets.
Technology
Surface Observations
Prior to the MAR, NWS, Federal Aviation Administration (FAA), and Department of Defense (DOD) staff manually made surface observations. Methods of weather observation had changed very little in the 100 years preceding the MAR (McNulty et al., 1990), and studies had found large variations in manual observations from individual to individual, and from site to site (Chisholm and Kruse, 1974; Woodall, 1966). In addition, the growing aviation industry increased the demand for surface observations. The desire to better address mesoscale weather events (e.g., severe thunderstorms, hail, and tornadoes) required a denser network of observing stations taking frequent and continuous observations.
The NWS and FAA teamed with the DOD (i.e., Air Force and Navy) to begin the process of replacing manual surface observations at approximately 250 airports, which were not always recorded around the clock, with the Automated Surface Observing System (ASOS). The three agencies designed ASOS to improve upon the manual surface observation practices and standards, operate 24 hours a day, seven days a
week, and increase the spatial resolution of surface observations by expanding from 250 to almost 1,000 airports around the country. The network was intended to automate the observation and dissemination of temperature, dew point, visibility, wind direction, wind speed, barometric pressure, cloud height and amount, and the type and amount of precipitation. The goal was acquisition of spatially and temporally uniform measurements, continuous observation and reporting, and more observing sites nationwide.
Radar
The NWS weather radar system in the 1980s comprised some fifty-odd WSR-57 and WSR-74S (Weather Surveillance Radar) S-band “network” radars and nearly seventy WSR-74C C-band “local warning” radars. These radars displayed the storm echo patterns and measured radar reflectivity, related to storm intensity, in a semi-quantitative manner. Coverage at mid-levels for the atmosphere was fairly broad east of the Rockies, but only spotty farther west. The WSR-57s in particular were aging and becoming difficult and expensive to maintain. Thus the need for a replacement system in the not too distant future was becoming pronounced.
Fortunately, the development of the Next Generation Weather Radar (NEXRAD) was well under way long before the nominal beginning of the MAR. Early work using 3.2 cm (X-band) wavelength short-range continuous-wave (CW) Doppler radar technology had demonstrated capability to detect tornadic wind speeds (Smith and Holmes, 1961) in addition to measuring reflectivity. However, that system was limited by inability to determine range to the target and by problems with loss of signal intensity in conditions involving precipitation. For routine operational applications, the development of pulse-Doppler technology for long-range weather radar (at longer wavelengths less subject to attenuation) was needed to furnish both range and velocity information (Whiton et al., 1998). Improvements in data processing and display technology were also needed to present the information in usable formats.
Work on the pulse-Doppler technology also began around the late-1950s (Rogers, 1990), first under U.S. Air Force (USAF) auspices and later at the National Severe Storms Laboratory (NSSL). By the late 1960s it was evident that the technology could reveal storm signatures of potential value in forecast and warning applications (Donaldson et al., 1969); a tornado vortex signature was identified in the echoes from a 1973 Oklahoma storm (Burgess et al., 1975). However, it took the introduction of real-time computing and the development of color display technology in the early 1970s to provide a means for bringing the data from a single Doppler radar to meteorologists in a conveniently usable fashion.
In the mid-1970s the NWS jointly teamed with the DOD and the Department of Transportation (DOT) in anticipation of the need to replace the WSR-57, WSR-74, and FPS-77 radars deployed over the preceding 20 years, to form the Joint Doppler Operational Project (JDOP; Whiton et al., 1998). The experiments and tests performed at NSSL and by the NWS and USAF Air Weather Service in 1976 and 1977 showed that Doppler radar provided much earlier detection of severe and tornadic storms, and could also detect gust fronts that might present a hazard to flight operations at airports.
On the basis of the successful JDOP demonstration of the potential value of Doppler radar to the missions of the NWS, the USAF, and the FAA, development of the NEXRAD system got under way in earnest in 1979: the Office of the Federal Coordinator for Meteorological Services and Supporting Research (OFCM) approved a NEXRAD concept document and established a tri-agency NEXRAD Program Council (NPC); the NPC approved formation of a Radar Test and Development Branch (later to become the Interim Operational Test Facility, then the Operational Support Facility, and eventually the Radar Operations Center); and the Office of Management and Budget (OMB) directed the OFCM to conduct a tri-agency cross-cut study for NEXRAD. Finally, NOAA approved establishment of a NEXRAD Joint System Program Office (JSPO) to move forward with the development, contract award, and deployment of a NEXRAD network. An NRC report (NRC, 1980) added momentum to the effort to implement an operational Doppler weather radar capability. The NPC formed a NEXRAD Technical Advisory Committee in 1980 to provide recommendations on newly-developed capabilities that are ready for implementation as well as engineering and scien-
tific developments needed to improve the NEXRAD capabilities. Thus the NEXRAD development process was under way well before the nominal beginning of the MAR. In fact, the NEXRAD system was eventually designated officially as the WSR-88D, the “88” signifying the year when the basic design was finalized, the year before the MAR officially began.
Congress appropriated the first funding for NEXRAD in the fall of 1980. The JSPO issued Joint Operational Requirements and NEXRAD Technical Requirements (NTR) documents in 1981 to initiate the process of system development and procurement (Whiton et al., 1998). Work by the three System Definition Phase contractors indicated that modifications to the NTR would be needed to define an affordable system. With those revisions accomplished two Validation Phase contractors began work in 1983; this phase, including Initial Operational Test and Evaluation (Part 1), was completed in 1987 and led to the selection of the Unisys design for the Limited and Full-Scale Production phases. During that period a different vendor promoted the idea of using C-band radars as a less expensive alternative to the S-band design, but a 1985 “Blue Ribbon Panel” headed by Raymond Kammer reviewed the revised NEXRAD requirements and found them to be “on target” and directly related to weather and public safety needs (ROC, 2011; U.S. Congress, 1985). The Unisys prototype arrived at the Operational Support Facility (OSF) in late 1988 for further operational test and evaluation, with production readiness established at the end of 1989—by which time the official MAR was under way.
Meanwhile, the site-survey contractor had begun work in 1983 to identify prospective sites for the NEXRAD network. A NEXRAD Siting Handbook issued in 1983 (JSPO, 1983) outlined the planned approach for deploying the radars. Insofar as possible, existing radar sites or other user facilities were to be used, simplifying problems of land acquisition, site access, and utilities. Guidance in the Siting Handbook indicated that radar coverage was to be the primary requirement. After preliminary surveys, in-depth surveys were conducted of promising candidate sites. A detailed report was prepared for each survey, focusing on coverage and cost issues (including particularly the cost of wideband communication between the radar site and the location of the principal users of the data). With the costs of wideband communication links at the time, the principal users had to be located not far from the radar site proper. In some cases the radar site was to be moved from city locations (which suffered from extensive ground clutter, a “cone of silence” or coverage gap, and radio frequency interference [RFI] problems) to more rural locations. While new or modified operational offices or centers were specifically not part of the NEXRAD system at this stage (though the costs for such things were later included in the estimated cost of the NEXRAD system; GAO, 1991a), under the restructuring some of those locations also became preferred locations for the new WFOs.
Satellites
The National Environmental Satellite, Data, and Information Service (NESDIS) is the National Oceanic and Atmospheric Administration (NOAA) line office responsible for satellites and in this capacity was a major contributor to the MAR. Only a combination of geostationary and polar-orbiting satellites can provide the spatial and temporal coverage and resolution required to measure the atmosphere and Earth system for weather and climate information. As early as the late 1980s and early 1990s there was an understanding that modernization of the observing satellite systems was expected to lead to improvements in Numerical Weather Prediction (NWP). NWP models use input data describing temperature, moisture, and wind parameters in the atmosphere. These data are obtained via various observation technologies; however, none are as globally complete and areally consistent as those from satellite data. Upgrades to the sounders, including microwave sounders, were of particular interest to NWP.
Geostationary satellites, consistently stationed above the same point on Earth, are important for near-continuous monitoring of the tropics and mid-latitudes within a hemispheric view, but do not capture the polar regions as well. A set of polar orbiting satellites, each crossing above the equator at a different local time, work together to provide coverage of the entire Earth, including the poles. Each polar satellite observes a given point on Earth’s surface and the atmosphere above it only twice a day. Although the polar system observations have lower temporal resolution in comparison
to those from the geostationary system, they have the advantage of being at a higher spatial resolution due to the much lower orbital altitude. In addition, the temperature and vapor soundings derived from polar orbiters have better vertical resolution. The complete global coverage that the sounder data provides is used for initiation of global NWP models. In addition, the polar-orbiting satellites provide better all-weather performance.
The launch of the Television Infrared Observation Satellite (TIROS-1) in 1960 began significant strides forward in synoptic scale weather interpretation with routine global cloud observations from the system of polar orbiting satellites (NRC, 1999b). The images proved valuable in data-sparse areas, particularly in detecting and tracking tropical storms over the oceans (NRC, 1997b).
Beginning with the launch of the Applications Technology Satellite (ATS-1) in geostationary orbit in 1966, meteorologists obtained full disk images of Earth and its cloud cover every 20 minutes. The spin scan cloud camera implemented on the ATS-1 geostationary platform enabled observations of weather systems in motion during daytime (Purdom, 1996). Since then, each new series of geostationary satellites has incorporated improvements in both instruments and data provision. Improvements in the instruments included addition of infrared and microwave channels to the visible channels on the imager, allowing nighttime observations, and addition of a sounder capability to observe the vertical structure of the atmosphere. Since its first launch in 1975, the Geostationary Operational Environmental Satellite (GOES) data has been a critical part of NWS operations by providing cloud and water vapor imagery to the National Centers through direct receipt. The GOES series of satellites also began to assist in provision and transmission of additional data. For example, starting in the mid-1970s the GOES Data Collection System (DCS) was implemented, allowing for the relay of data from remote, ground-based data collection platforms through the satellite to a central processing facility.
National Centers Computing Capacity
The need to modernize computational capacity at NWS national centers was well recognized at the time of the MAR and was one of the major components of the modernization. Kalnay et al. (1998) document the evolution of numerical weather prediction techniques within the NWS against the backdrop of evolving computing capacity from the 1950s through the mid-1990s. Computing capacity increased approximately six orders of magnitude (in terms of “flops” or “floating point operations per second”) since the NWS undertook NWP activities in the late 1950s. Two emerging capabilities helped define and drive the MAR objectives for more uniform and scientifically-based forecast products: the power to generate timely and accurate information content and the uniformity of nationally distributable forecast products afforded by the growing computational capacity. Managing, disseminating, and interpreting this expanding volume of information content required changes in many areas. The downscaling of numerical prediction results to specific guidance information that forecasters could utilize for their specific location was another important development.
Forecaster Workstations
Before the deployment in the late 1970s and early 1980s of the Automation of Field Operations and Services (AFOS), a computer-based forecaster workstation technology, the communication infrastructure of the NWS consisted of teletypewriter and facsimile circuits. AFOS consisted of a set of mini-computers and telephone communication systems organized as “regional loops” supported by hub-and-spoke networks that interconnected each Weather Service Forecast Office and its Weather Service Offices. The communications system was vulnerable to failure, especially in severe weather conditions (high winds, ice storms, etc.). In the late-1980s, the AFOS system became increasingly technologically obsolete and not worth modification or upgrading (NBS, 1988). Major advances in meteorological instrumentation and measurement techniques were providing new data and information, contributing to improved weather forecasting and warning. The Advanced Weather Interactive Processing System (AWIPS) project addressed the AFOS problem and was intended to harness the rapidly advancing technologies. AWIPS later served as the backbone of the MAR, providing forecasters with a system to use all available NWS sources of data. The first release
of AWIPS was not a true “modern architecture” but a lengthy set of codes operating on updated, higher throughput, hardware. The software was later rewritten to become the modern, modular, open architecture it is today that can accommodate upgrades and improvements such as AWIPS-II, presently being staged for operational deployment.
Operations
The NWS had a two-tiered office structure prior to the MAR. The first tier of 52 Weather Service Forecast Offices (WSFOs), about one per state, had a core component of professional meteorologists. The WSFOs prepared general forecasts for their assigned region of responsibility and provided severe weather warnings for their immediate local area covered by the station radar. They also recorded local observations and often had upper-air radiosonde observing responsibility. The second tier of 204 Weather Service Offices (WSOs) was staffed with observers and meteorological technicians. Some WSOs had local weather radars and had local responsibility for issuing severe weather warnings. All WSOs had surface observing responsibility and some performed upper-air observations. Some WSOs were open only part time.
It is difficult to obtain comprehensive data regarding the skill level, or performance metrics, of the NWS general weather forecasting prior to and during the MAR. Forecast verification data is collected centrally, and is made available to NOAA employees, and to other government employees and researchers on a case-by-case basis. However, some data are available for tornado and flash flood warnings (see Figure 4.3). For example, in the late 1980s, about 40 percent of tornado occurrences were detected, with an average warning lead time of five minutes and a false alarm rate of about 80 percent. There was a similar detection rate of about 40 percent for flash floods, with a warning lead time of near 10 minutes, and a false alarm ratio of about 60 percent.
In November 1988, via Public Law 100-685, Congress instructed the Secretary of Commerce to prepare a 10-year strategic plan for the comprehensive modernization of the NWS (U.S. Congress, 1988).1 The strategic plan would set forth the basic service improvement objectives of the modernization. It would describe the critical new technology components as well as the associated staff and operational changes necessary to fulfill the objectives of weather and flood forecasting and warning service improvements.
In response to the Congressional request, the NWS prepared, in March 1989, the Strategic Plan for the Modernization and Associated Restructuring of the National Weather Service. The Strategic Plan stated the objective of the MAR as follows:
[t]o modernize the NWS through the deployment of proven observational, information processing and communications technologies, and to establish an associated cost effective operational structure. The modernization and associated restructuring of NWS shall assure that the major advances which have been made in our ability to observe and understand the atmosphere are applied to the practical problems of providing weather and hydrologic services to the Nation (NWS, 1989).
The Strategic Plan emphasized that the MAR would be dependent on the development and implementation of several major technologies including
• Automated Surface Observing System (ASOS): an automated electronic sensor instrument system to replace manual weather observations at all NWS (and many other) surface observing locations, and increase the number of observing locations;
• Next Generation Weather Radar (NEXRAD): a network of advanced Doppler radars to measure the motions of the atmosphere responsible for severe weather such as tornadoes, to detect heavy rainfall and hail, and to increase lead times for prediction and warning of severe weather events and flash floods;
• Satellite Upgrades: a new series of geostationary meteorological satellites to provide higher spatial and temporal resolution imagery and data to aid shorter-range forecasts and warnings, and a new series of polar orbiting meteorological satellites to provide improved,
__________________
1Public Law 100-685 was later replaced by Public Law 102-567, which included the same requirements for a Strategic Plan and National Implementation Plan as well as more detailed guidance for the execution of the MAR.
all-weather, atmospheric data to assist in longer term forecasting;
• National Centers Advanced Computer Systems: a ten-fold increase in computing power to support the National Centers. Along with numerical weather prediction model improvements, this improved national guidance for forecasts and warnings; and
• Advanced Weather Interactive Processing System (AWIPS): an advanced computer and communications system to help forecasters integrate all sources of weather data. The system allowed communication between each weather forecast office and distribution of centrally collected data and centrally produced analysis and guidance products, as well as satellite data and imagery (NWS, 1989).
In Public Law 100-685, Congress also requested that one year after submission of the Strategic Plan, the NWS prepare and submit an initial implementation plan with annual revisions. The NWS published in March 1990 The National Implementation Plan for the Modernization and Associated Restructuring of the National Weather Service (NIP). The NIP planned a transition to the modernized NWS that would be driven by service requirements and accomplished in two distinct stages. This staging was associated with the period of time between the deployment of new observational systems such as ASOS and NEXRAD, and that of the new information processing system, AWIPS. The staging would provide a stabilization period to allow field offices to adjust to, and gain familiarity with, the new Doppler radar system and data, and high resolution surface observation data (NWS, 1990).
Stage 1 would be characterized by an improvement in severe weather detection capability. This would result from meteorological interpretation of the new and enhanced observational data made available by the deployment of ASOS and NEXRAD (NWS, 1990). Stage 2 would be characterized by operation of a reliable predictive warning program. Forecasters using AWIPS would have the necessary tools to integrate, analyze, and interpret all the various data and information, and rapidly disseminate products (NWS, 1990).
Congress required that no WSFO or WSO be closed, consolidated, automated, or relocated unless the Secretary of Commerce certified to the appropriate Congressional committees that “such action would not result in any degradation of weather services provided to the affected area” (U.S. Congress, 1992). An independent advisory committee, the Modernization Transition Committee (MTC), was established to provide a review of each certification and advise the Secretary (U.S. Congress, 1992).
The overall objective of the MAR was to improve weather services while simultaneously establishing a more cost efficient organization. The specific benefits the NWS hoped to achieve with the MAR included
• more uniform weather services across the Nation;
• improved forecasts;
• more reliable detection and prediction of severe weather and flooding;
• more cost effective NWS; and
• higher productivity for NWS employees (NWS, 1989).
The NIP, while still stating the overall objectives of the MAR as stated in the Strategic Plan, expanded and clarified the list of specific goals to include
• operational realization of a predictive warning program focusing on mesoscale meteorology and hydrology;
• advancement of the science of meteorology and hydrology;
• development of NWS human resources to achieve maximum benefit from recent scientific and technical advances;
• user acceptance and support of NWS modernization and associated restructuring service improvement objectives;
• strengthening cooperation with the mass media, universities, the research community, and the private hydrometeorological sector to collectively fulfill the Nation’s weather information needs from provision of severe weather warnings and general forecasts for the public as a whole, which is a Government responsibility; to provision of detailed and customer specific weather information, which is a private sector responsibility;
• achievement of productivity gains through automation and replacement of obsolete technological systems; and
• operation of the optimum NWS warning and forecast system consistent with service requirements, user acceptability, and affordability (NWS, 1990).
By the end of Stage 2 of the implementation of the MAR, the NWS would have obtained the capability to forecast and warn of severe weather events with lead times of tens of minutes and with increased geographic specificity.
This page intentionally left blank.
