Weather Forecasting: Radar Availability Requirement Not Being Met
(Chapter Report, 05/31/95, GAO/AIMD-95-132).

Pursuant to a congressional request, GAO reviewed the joint acquisition
of the Next Generation Weather Radar (NEXRAD) by the National Weather
Service (NWS), Federal Aviation Administration (FAA), and Air Force,
focusing on: (1) changes in acquisition and deployment plans; (2) the
feasibility and cost of purchasing additional radars; (3) the Air Force
contribution to the national NEXRAD network; and (4) the availability of
Air Force and NWS NEXRAD.

GAO found that: (1) the three agencies reduced their original deployment
plan from 175 radars to 163 radars because of changes in agency
requirements, funding limitations, and military base closures; (2) the
National Research Council (NRC) is studying the adequacy of proposed
NEXRAD coverage; (3) NWS and the Air Force plan to deploy their radars
by the end of fiscal year (FY) 1996, but FAA plans to delay deploying
five of its radars until at least FY 1997 because of budget constraints;
(4) because the option for purchasing additional radars is unpriced and
subject to negotiation, additional NEXRAD could cost three times more
than units currently under contract; (5) NWS has unrestricted access to
Air Force NEXRAD, which provide essential backup, primary, and
supplemental radar coverage; (6) all three agencies signed an agreement
to operate all NEXRAD to satisfy the integrated needs of all three
agencies; (7) Air Force radar availability data, which may be unreliable
and overstated, indicate that some of its radars are performing below
the tri-agency system availability requirements; and (8) NWS does not
calculate operational availability for each NEXRAD site and it does not
know if sites are meeting availability requirements.

--------------------------- Indexing Terms -----------------------------

 REPORTNUM:  AIMD-95-132
     TITLE:  Weather Forecasting: Radar Availability Requirement Not 
             Being Met
      DATE:  05/31/95
   SUBJECT:  Weather forecasting
             Radar equipment
             Air traffic control systems
             Data transmission operations
             Avionics
             Interagency relations
             Computer networks
             Storms
             Warning systems
             Federal procurement
IDENTIFIER:  NWS Next Generation Weather Radar
             FAA Next-Generation Weather Radar Program
             NWS Automated Surface Observing System
             National Weather Radar Network
             
**************************************************************************
* This file contains an ASCII representation of the text of a GAO        *
* report.  Delineations within the text indicating chapter titles,       *
* headings, and bullets are preserved.  Major divisions and subdivisions *
* of the text, such as Chapters, Sections, and Appendixes, are           *
* identified by double and single lines.  The numbers on the right end   *
* of these lines indicate the position of each of the subsections in the *
* document outline.  These numbers do NOT correspond with the page       *
* numbers of the printed product.                                        *
*                                                                        *
* No attempt has been made to display graphic images, although figure    *
* captions are reproduced. Tables are included, but may not resemble     *
* those in the printed version.                                          *
*                                                                        *
* A printed copy of this report may be obtained from the GAO Document    *
* Distribution Facility by calling (202) 512-6000, by faxing your        *
* request to (301) 258-4066, or by writing to P.O. Box 6015,             *
* Gaithersburg, MD 20884-6015. We are unable to accept electronic orders *
* for printed documents at this time.                                    *
**************************************************************************


Cover
================================================================ COVER


Report to the Chairman and Ranking Minority Member, Committee on
Science, House of Representatives

May 1995

WEATHER FORECASTING - RADAR
AVAILABILITY REQUIREMENT NOT BEING
MET

GAO/AIMD-95-132

Weather Forecasting


Abbreviations
=============================================================== ABBREV

  ASOS - Automated Surface Observing System
  AWIPS - Advanced Weather Interactive Processing System
  C4 - Command, Control, Communications, and Computer
  CONUS - conterminous United States
  DOD - Department of Defense
  FAA - Federal Aviation Administration
  GAO - General Accounting Office
  GOES-Next - Next Generation Geostationary Operational Environmental
     Satellite
  NEXRAD - Next Generation Weather Radar
  NOAA - National Oceanic and Atmospheric Administration
  NRC - National Research Council
  NWS - National Weather Service
  PUP - principal user processor
  RDA - radar data acquisition
  RPG - radar product generator

Letter
=============================================================== LETTER


B-258872

May 31, 1995

The Honorable Robert S.  Walker
Chairman
The Honorable George E.  Brown, Jr.
Ranking Minority Member
Committee on Science
House of Representatives

This report responds to your request that we review the Next
Generation Weather Radar (NEXRAD), a Doppler weather radar that the
Departments of Commerce, Defense, and Transportation are jointly
acquiring.  As agreed, we focused on changes in the number of NEXRADs
being deployed, contract options to purchase additional radar units,
the Air Force's contribution to the national NEXRAD network, and the
accessibility and availability of NEXRADs. 

We are providing copies of this report to the Secretaries of
Commerce, Defense, Transportation, and the Air Force; the Director of
the Office of Management and Budget; and other interested
congressional committees.  Copies will also be made available to
others upon request. 

Please call me at (202) 512-6253 if you or your staff have any
questions concerning this report.  Major contributors are listed in
appendix III. 

Joel C.  Willemssen
Director, Information Resources
 Management/Resources, Community,
 and Economic Development


EXECUTIVE SUMMARY
============================================================ Chapter 0


   PURPOSE
---------------------------------------------------------- Chapter 0:1

One needs to look no further than the evening news to appreciate how
Doppler radar systems\1 are allowing forecasters to better see the
makeup and movement of weather and to quickly advise the public of
severe events, such as tornadoes.  As part of its approximately $4.5
billion systems modernization program, the National Weather Service
(NWS), a component of the National Oceanic and Atmospheric
Administration (NOAA), is collaborating with the Federal Aviation
Administration (FAA) and the Air Force in acquiring Doppler radar
technology.  This $1.4 billion radar project, known as the Next
Generation Weather Radar (NEXRAD), is to establish a constellation of
radars to increase the accuracy, timeliness, and credibility of
hazardous weather warnings. 

Recent changes to the deployment schedule, uncertainties about the
need for additional radars, and questions concerning interagency
cooperation prompted the Chairman and Ranking Minority Member of the
House Committee on Science to request that GAO determine (1) the
NEXRAD units that were dropped from the original deployment plan and
the reasons they were dropped, (2) the feasibility and estimated cost
of extending the NEXRAD contract to purchase additional radars, (3)
the Air Force NEXRADs' contribution to the national NEXRAD network
and the accessibility of the Air Force NEXRAD data to civilian
forecasters, and (4) the availability of the Air Force and NWS
NEXRADs.  Our objectives did not include determining the adequacy of
national radar coverage because the National Research Council (NRC)
is reporting separately on this issue. 


--------------------
\1 Doppler radar is used to determine the speed and direction of rain
or snow particles, cloud droplets, or dust moving toward or away from
the radar.  The radar accomplishes this by sending out a pulse using
a stable frequency and then measuring the changing frequencies as the
distance between the radar and the object changes. 


   BACKGROUND
---------------------------------------------------------- Chapter 0:2

NEXRAD is a Doppler radar system that measures wind velocity in
severe weather, tracks storm movement and intensity, and generates
data and imagery for forecasters and other users.  NWS and the Air
Force report that their experience with the data from over 100
operational NEXRADS has increased the accuracy, timeliness, and
credibility of warnings of severe thunderstorms, tornadoes, flash
floods, turbulence, wind shear, and other types of hazardous weather
events. 

NWS, the Air Force, and FAA currently plan to purchase and deploy
119, 30, and 14 NEXRADs, respectively, for a total of 163.  As of
February 1995,
107 radars had been deployed--81 by NWS, 22 by the Air Force, and 4
by FAA.  Of the 163 planned radars, 144 are to be located at NWS and
Defense sites within the conterminous United States (CONUS).\2 These
CONUS sites are to provide adequate geographic coverage of national
weather events, thereby collectively supporting the three agencies'
respective missions.  The non-CONUS radars are also to support the
agencies respective missions at 19 selected locations in Hawaii,
Alaska, the Caribbean, the Atlantic, the Pacific, and Korea.  NWS
relies on several of these 19 non-CONUS radars to provide information
about approaching off-shore weather.  FAA and the Air Force rely on
many of these 19 radars to ensure safe aviation operations and
resource protection. 


--------------------
\2 The conterminous United States consists of all the states except
Alaska and Hawaii. 


   RESULTS IN BRIEF
---------------------------------------------------------- Chapter 0:3

Originally the three agencies planned to deploy 175 radars.\3 Recent
changes to the NEXRAD deployment plan have decreased the number to
163.  These changes are because of changes to agency requirements,
military base closings, and funding limitations.  Also, FAA and NWS
officials told GAO that FAA will delay deploying five of its radars
for probably more than a year because of budget constraints. 

A NEXRAD contract option exists to acquire up to 20 additional
radars.  These radars could be as much as three times as expensive as
current units because manufacturer production lines have been shut
down, and restarting them would involve considerable expense.  NWS
officials told GAO that the 163 radars will provide coverage equal to
or better than the existing coverage.  The 163 radars are expected to
meet the needs of the three agencies, and the agencies do not plan to
acquire additional NEXRADs.  Therefore, NWS has not reassessed the
cost-effectiveness of acquiring additional radars under the contract
option.  However, the NRC study director stated that NRC expects to
report on weaknesses in national coverage that may require NWS to buy
additional radars, assuming that the benefits of doing so outweigh
the associated cost. 

According to NWS, the Air Force NEXRADs are essential to NWS' ability
to issue quality forecasts and warnings because some provide the sole
radar coverage for certain geographic areas, and all provide backup
coverage in the event an NWS radar goes down.  The Air Force does not
restrict NWS forecasters' access to its radar data; however, Air
Force data show that its radars are not available\4 to the extent
that the three agencies agreed is necessary.  To make matters worse,
the Air Force availability data are unreliable and appear to be
overstated.  Also, NWS does not know if its individual radars are
available to the extent necessary because it does not monitor radar
availability by site. 


--------------------
\3 These 175 radars included 115 for NWS, 44 for the Air Force, and
16 for FAA. 

\4 Available means the time that the system is operating
satisfactorily, expressed as a percentage of total time. 


   PRINCIPAL FINDINGS
---------------------------------------------------------- Chapter 0:4


      GLOBAL NEXRAD DEPLOYMENTS
      HAVE DECREASED
-------------------------------------------------------- Chapter 0:4.1

In 1980, NWS, the Air Force, and FAA agreed to jointly deploy 175
NEXRAD units.  However, they now plan to deploy 163 units.  Most of
these radars will be deployed by the end of fiscal year 1996.  The
reduction of 12, consisting of a net decrease of 13 outside CONUS and
a net increase of one within CONUS, was due to changes to agency
requirements, funding limitations, and military base closures.  Table
2.3 provides a complete list of sites added to and deleted from the
deployment plan and the reasons for these changes. 

Also, FAA plans to place 5 of its 14 radars in storage until at least
fiscal year 1997, and probably longer, because higher priority
funding requirements are preventing FAA from paying the costs
associated with deploying these radars.  While these units are not
located within CONUS, and thus do not affect NWS CONUS weather
coverage, NWS officials said the radars are important to NWS' ability
to issue timely and accurate forecasts and warnings.  For instance,
two of the radars located in the Caribbean would allow NWS to better
track and monitor hurricanes as they approach the United States. 
However, NWS does not have a mission requirement for radar coverage
outside CONUS. 

NWS told GAO that the national radar coverage that is currently
planned is equal to or better than existing coverage.  However, the
Secretary of Commerce, at the request of the House Committee on
Science, Space, and Technology (now the House Committee on Science),
commissioned NRC to study and report on the adequacy of proposed
CONUS coverage compared to the existing coverage.  The NRC study
director said he expects the study to identify weaknesses in coverage
and potential areas where additional radars may be needed.  This
report is to be issued in June 1995. 


      UNIT COST TO ACQUIRE
      ADDITIONAL NEXRADS COULD BE
      MUCH HIGHER
-------------------------------------------------------- Chapter 0:4.2

Included in the contract for the development and acquisition of
NEXRADs is an option to purchase up to 20 additional NEXRADs through
August 1996.  However, this option is not priced, meaning that the
unit cost for each additional radar is subject to negotiation. 
NEXRAD program office officials estimate that, depending on the
number of radars that are needed, each radar could cost as much as
three times the price of current units.  These officials stated that
the higher unit costs would be due to breaks in production. 

Program officials said they agreed to an unpriced option because no
firm requirements for additional radars existed at the time the
contract was negotiated.  The contractor did not price the optional
units because of the uncertainty of future costs (e.g., the costs of
restarting subcontractor production lines). 

NWS has no plans to buy additional radars, and as a result, has not
reassessed the cost-effectiveness of acquiring the more expensive
radars.  However, NRC is expected to report on national radar
coverage that may warrant a reassessment of these plans. 


      AIR FORCE NEXRADS ARE
      INTEGRAL TO NWS MISSION
      PERFORMANCE AND ACCESSIBLE
      TO NWS FORECASTERS
-------------------------------------------------------- Chapter 0:4.3

The Air Force NEXRADs play a critical role in NWS' overall ability to
issue complete and accurate weather forecasts and warnings.  Seven of
the 22 CONUS-based operational Air Force NEXRADs provide primary NWS
radar coverage, according to NEXRAD program office officials.  Also,
the Air Force radars provide important backup coverage in the event
that an NWS radar providing primary coverage for a given geographic
area fails, and they provide supplemental views of severe weather
patterns from different angles that strengthen NWS' watch and warning
capabilities.  The Air Force does not restrict NWS forecasters access
to its radar data. 


      AIR FORCE AND NWS RADARS MAY
      NOT BE AVAILABLE WHEN NEEDED
-------------------------------------------------------- Chapter 0:4.4

The Air Force NEXRADs may not be available when information from them
is needed.  A key NEXRAD requirement is that each unit should be
operationally available 96 percent of the time.  However, 1994 and
1995 data show that only 38 to 90 percent of Air Force radars met
this requirement each month, and in fact for 9 of these months no
more than 70 percent of the sites met the requirement.  Moreover,
this situation could be much worse because the availability
information that the Air Force has been reporting is unreliable.  For
instance, the Air Weather Service, which monitors the Air Force's
NEXRAD operations, reported that the Eglin Air Force Base radar was
available 100 percent of the time from September through December
1994, based on data from Air Force maintenance data collection
systems.  However, base radar officials told GAO that their radar's
availability ranged from 78 to 87 percent per month for these 4
months.  GAO found other examples of radar outages that were not
reflected in Air Force availability data collected by the Air Weather
Service. 

In many cases, the Air Force NEXRAD operators and maintainers were
not aware of the 96 percent availability requirement and, therefore,
had no way of knowing that their performance was subpar. 
Inefficiencies in the Air Force's logistics process for obtaining
spare parts have also made it difficult to meet availability
requirements. 

NWS also does not know if it is meeting the availability requirement
for each of its units because it does not monitor availability on a
site by site basis.  Although NWS records radar downtime by site, it
only uses this information to calculate the average availability of
all sites, and it only monitors radar availability performance on
this basis.  It does not use this information to calculate and
monitor site-specific availability.  While GAO agrees that these
aggregate data are useful in monitoring such things as spare parts
usage and maintenance staffing trends, the data do not disclose
whether each radar meets the required 96- percent availability
requirement. 


   RECOMMENDATIONS
---------------------------------------------------------- Chapter 0:5

GAO recommends that the Secretary of Commerce direct the NOAA
Assistant Administrator for Weather Services to

  not purchase additional radars to address any weaknesses in radar
     coverage that may result from the NRC study until assessing
     FAA's plans for deploying the five radars scheduled for storage
     and NWS' mission requirements for NEXRADs in these areas,

  ensure that any radars bought in response to NRC's national radar
     coverage findings are cost-beneficial, given that their unit
     cost could be substantially higher than those already purchased,
     and

  analyze and monitor system availability data on a site-specific
     basis for operational NEXRADs and correct any shortfalls in
     system availability that this analysis shows. 

GAO also recommends that the Secretary of the Air Force direct the
Air Force Director of Weather to improve the reliability of Air Force
NEXRAD availability data and to correct any shortfalls that these
data show. 


   AGENCY COMMENTS
---------------------------------------------------------- Chapter 0:6

GAO received written comments on a draft of this report from the
Departments of Commerce and Defense, and oral comments from senior
FAA officials, including the NEXRAD program manager.  Commerce's
written comments are in appendix I, and Defense's written comments
are in appendix II. 

The Department of Commerce generally concurred with GAO's findings,
conclusions, and recommendations, and stated that NWS is taking steps
to analyze and monitor system availability on a site-specific basis. 
In a draft of this report, GAO proposed that the NOAA Assistant
Administrator for Weather Services assess the operational impact of
FAA delays in deploying the five radars and, on the basis of this
assessment, take the necessary steps to ensure that NWS' radar
coverage needs are met.  In their comments, the Department of
Commerce and the FAA NEXRAD program manager stated that NWS does not
have a mission requirement for radars outside of CONUS.  GAO has
incorporated this comment in the section describing FAA plans for
storing five radars and has revised the recommendation accordingly. 

Commerce also partially concurred with the recommendation to ensure
that any radars bought in response to NRC's coverage findings are
cost-beneficial.  However, Commerce requested that the recommendation
be modified to reflect only NWS core mission and the Weather Service
Modernization Act requirements.  The NRC study director told GAO that
its study will only address NWS' core mission and the act and GAO,
therefore, did not modify its recommendation. 

The Department of Defense concurred with GAO's recommendation
concerning the Air Force, and stated that it will develop management
actions to improve NEXRAD availability and the reliability of Air
Force data. 


INTRODUCTION
============================================================ Chapter 1

Accurate and timely weather forecasts and warnings are vital to the
protection of life and property.  Hundreds of lives and billions of
dollars in property are lost every year as a result of thunderstorms,
lightning, tornadoes, hurricanes, blizzards, and floods. 

The National Weather Service's (NWS) basic mission is to provide
weather and flood warnings, forecasts, and advisories for the
protection of life and personal property.  NWS operations also
support other federal missions, such as aviation safety, and our
nation's commercial interests, such as the agriculture industry.  The
Federal Aviation Administration (FAA) and the Air Force, besides
being users of NWS data and information, also collect and analyze
certain weather observations to support their respective missions. 
FAA, for example, collects and displays weather radar, cloud ceiling,
and visibility data for its air traffic controllers to use. 

Since the early 1980s, NWS has been modernizing its weather
observing, information processing, and communication systems to
predict the weather more accurately and quickly.  This approximately
$4.5 billion modernization consists of four major system acquisitions
and several smaller system upgrades and developments.  FAA and the
Department of Defense (DOD) are collaborating with NWS on two of
these major acquisitions--the Next Generation Weather Radar (NEXRAD)
and the Automated Surface Observing System (ASOS).\1 In addition to
improved weather predictions, NWS expects the modernization to permit
it to streamline its operations and downsize its organization without
a degradation of service.  For example, it expects to reduce its
number of field offices from about 250 to 118 and to reduce staffing
levels from 4,700 to 3,900. 


--------------------
\1 The other two major system acquisitions are the Advanced Weather
Interactive Processing System (AWIPS) and the Next Generation
Geostationary Operational Environmental Satellite (GOES-Next). 


   NEXRAD:  A BRIEF OVERVIEW
---------------------------------------------------------- Chapter 1:1

NEXRAD is a Doppler radar\2 system that measures wind velocity in
severe weather, tracks storm movement and intensity, and generates
data and imagery for forecasters and other users, such as air traffic
controllers.  NEXRAD is expected to provide improved weather radar
information, thus increasing the accuracy, timeliness, and
credibility for warnings of severe thunderstorms, tornadoes, flash
floods, turbulence, wind shear, and other types of hazardous weather
and related events.  The radars are also expected to be extremely
useful in analyzing a variety of other weather events, including
estimating accumulated rainfall and analyzing large-scale
precipitation systems. 

NEXRAD is being acquired jointly by NWS, the Air Force, and FAA.  The
three agencies currently plan to purchase and deploy 163 NEXRADs--119
for NWS, 30 for the Air Force, and 14 for FAA--at an estimated cost
of just over
$1.4 billion--$860 million from NWS, $264 million from the Air Force,
and $293 million from FAA.\3

Of the 163 NEXRADS, 144 are to be located within the conterminous
United States (CONUS),\4 11 are to be located in Hawaii and Alaska,
and 8 are to be located in the Caribbean, the Atlantic, the Pacific,
and Korea.  Data from these radars are shared among the three
agencies to support their respective missions.  For example, NWS
needs adequate CONUS coverage to issue timely and accurate forecasts
and warnings, and uses data from several Air Force NEXRADs to fill
some gaps in coverage.  Likewise, the Air Force and FAA rely on NWS
radars in addition to their own to support their respective national
defense and aviation missions. 

On the basis of the three agencies' collective mission needs and the
Weather Service Modernization Act, which mandates that the Secretary
of Commerce certify that there will be no degradation in radar
coverage at the 10,000 foot level prior to closing, consolidating,
automating, or relocating any of NWS' field offices, the three
agencies negotiated the radars' locations to meet tri-agency radar
coverage requirements.\5 The locations of all CONUS radars are shown
in figure 1.1. 

   Figure 1.1:  Locations of
   NEXRADs Within CONUS

   (See figure in printed
   edition.)

   .

   (See figure in printed
   edition.)



--------------------
\2 Doppler radar is used to determine the speed and direction of rain
or snow particles, cloud droplets, or dust moving toward or away from
the radar.  The radar accomplishes this by sending out a pulse using
a stable frequency and then measuring the changing frequencies as the
distances between the radar and the object changes. 

\3 The Air Force and FAA totals cover their respective NEXRADs and
2.7 percent and 20 percent, respectively, of NWS' NEXRADs due to a
cost sharing arrangement agreed to by the three agencies. 

\4 CONUS consists of all the States except Alaska and Hawaii. 

\5 The 10,000 foot level is significant because this is the elevation
at which the coverage range of an individual NEXRAD is measured.  The
ascending radar beam loses its reliability about 125 miles from the
radar.  At this distance the lowest part of the beam is approximately
10,000 feet off the ground.  Therefore, each radar has a coverage
diameter of 250 miles.  The 250 mile cylinders were the basis for
siting NEXRADs to ensure adequate CONUS coverage. 


      NEXRAD HISTORY AND STATUS
-------------------------------------------------------- Chapter 1:1.1

In 1980, NWS, the Air Force, and FAA agreed to jointly develop and
acquire 175 NEXRADs.  By 1987, production of a limited number of
NEXRAD units had been approved.  By 1990, the contractor, Unisys,
began experiencing development problems.  Unisys was behind schedule,
cost estimates were overrun, and specified performance requirements
were not being met.  Commerce's National Oceanic and Atmospheric
Administration (NOAA), of which NWS is a part, raised serious
concerns about the contractor's ability to complete the contract. 
NOAA's concerns focused on cost, schedule, and performance issues,
and on Unisys' financial condition.  According to a program manager
at that time, Unisys underbid the contract.  These issues led to NOAA
suspending the radar deliveries and considering contract termination. 

To address this dilemma, NOAA evaluated the pros and cons of (1)
reaching a comprehensive settlement with Unisys to deliver radar
systems or (2) terminating the existing contract and contracting with
another vendor.  After analyzing both choices, NOAA and Unisys signed
a comprehensive settlement of contractual issues in August 1991,
renegotiated the contract, and the production of radars resumed.  The
renegotiated contract included a $182 million increase in the
contract cost.  The associated increase in the unit cost of the
NEXRADs forced the Air Force to drop 13 units to remain within its
program funding limits. 

As of February 1995, 107 radars had been deployed--81 by NWS, 22 by
the Air Force, and 4 by FAA.  The final NEXRAD deployment is
scheduled for June 1996.  By September 1995, the three agencies are
expected to have collectively spent $1.2 billion. 


      NEXRADS CONTAIN THREE MAJOR
      COMPONENTS
-------------------------------------------------------- Chapter 1:1.2

Each NEXRAD consists of three major subsystems--the radar data
acquisition (RDA) subsystem, the radar product generator (RPG)
subsystem, and the principal user processor (PUP) subsystem--and
associated communications among these subsystems.  Each NEXRAD
includes about 400,000 lines of code for operating the radar,
processing radar signals, generating and transmitting data, and
displaying data products. 

The RDA consists of a 10 centimeter wavelength Doppler weather radar
that collects the raw data to, among other things, (1) measure wind
velocity in severe weather, (2) provide improved estimates of
precipitation amounts, and (3) track storm movement and intensity. 
The technology needed to perform this function includes an antenna,
pedestal, radome (a dome-shaped covering to protect the antenna),
transmitter, and receiver.  Included in the RDA unit is hardware and
software necessary for a variety of control functions, including
signal processing, monitoring, and error detection, as well as
archiving the radar data.  A computer processes the radar signals to
create digital data that can be further processed by the RPG. 

The RPG includes all hardware and software necessary for turning the
data into displayable data products.  Specifically, the RPG provides
real-time generation, storage, and distribution of products for
users.  It includes hardware and software required for system
control; status monitoring; and error detection, archiving, and data
processing. 

The PUP is a workstation that consists of the hardware and software
required for the request, display, local storage and annotation, and
distribution of products by forecasters.  It also includes the
hardware and software required for local control, status monitoring,
archiving, and communicating with other users.  The PUP maintains a
dedicated communication link to the RPG located on-site, and it
routinely receives NEXRAD products.  The PUP also has the capability
to access data from RPGs at other NEXRAD sites.  In addition, under
an NWS administered NEXRAD information dissemination service, NWS has
set aside four communications ports to allow access by commercial
companies that provide data to other government agencies and the
public.  Figure 1.2 shows the key NEXRAD subsystems for a typical NWS
weather forecast office. 

   Figure 1.2:  Key NEXRAD
   Subsystems

   (See figure in printed
   edition.)



      PROGRAM MANAGEMENT AND
      ORGANIZATION
-------------------------------------------------------- Chapter 1:1.3

The NEXRAD Joint Systems Program Office, hereafter referred to as the
program office, organizationally resides within NOAA and is
responsible for managing the acquisition of the radars.  The program
office is staffed and funded by the three participating agencies. 
The radars are to be purchased, operated, and maintained by the
respective acquiring agency, but information from each radar is to be
shared among all three. 

NWS' Office of Systems Operations will assume program management
responsibility from the program office once all NEXRADs have been
delivered.  In addition, the Operational Support Facility provides
technical support for operating and maintaining radar equipment.  To
ensure that the needs of all three agencies are met, both the program
office and the Operational Support Facility are jointly staffed and
funded by NWS, the Air Force, and FAA.  NWS' National Logistics
Supply Center in Kansas City, Missouri, will be the centralized
NEXRAD depot and repair center for all three agencies' radars. 

The Air Force owns all of DOD's NEXRADs, including four radars that
are located at Army locations.\6 The Director of Weather, Office of
the Deputy Chief of Staff, Plans, and Operations, within the Air
Force, is responsible for planning, programming, and budgeting for
weather support.  The Director of Weather's responsibilities include
publishing weather policy\7 and standardized procedures, and
assessing the technical performance and effectiveness of Air Force
weather support, including those associated with NEXRAD.  The
Director of Weather disseminates weather policy through the Air Force
major commands,\8 which in turn distribute it to the Air Force bases
responsible for the individual NEXRAD units.  The Air Force is
responsible for operating and maintaining DOD NEXRADs.  The Air
Force's Air Weather Service is the lead organization for oversight of
all Air Force NEXRADs. 


--------------------
\6 The Air Force owns the RDAs and RPGs associated with each DOD
NEXRAD.  The PUPs are owned by the Air Force, Navy, and Army since
all services are users of the weather data. 

\7 Weather support policy is established in the Office of the Chief
of Staff of the Air Force. 

\8 These commands include (1) Air Combat Command, (2) Air Education
and Training Command, (3) Air Mobility Command, (4) Air Force
Material Command, (5) Space Command, and (6) Pacific Air Forces. 


   OBJECTIVES, SCOPE, AND
   METHODOLOGY
---------------------------------------------------------- Chapter 1:2

The objectives of our review were to determine (1) the NEXRAD units
that were dropped from the original deployment plan and the reasons
they were dropped, (2) the feasibility and estimated cost of
extending the NEXRAD contract to purchase additional radars, (3) the
Air Force NEXRADs' contribution to the national NEXRAD network and
the accessibility of the Air Force NEXRAD data to civilian
forecasters, and (4) the availability of the Air Force and NWS
radars.  Generally, our methodology was guided by those sections of
GAO's System Assessment Framework pertaining to operational systems. 
A detailed description of our methodology follows. 

To determine which units have been dropped from the original
deployment schedule and why they were dropped, we met with NEXRAD
program officials to obtain the original and current deployment
schedules, discuss which units were dropped or added to the
deployment schedule, and identify the reasons why.  In addition, we
reviewed documentation on the 1991 comprehensive settlement, since
this settlement led to the majority of the deployment changes. 
Finally, we verified our analysis of the units affected and the
reasons why with NEXRAD program officials.  We did not identify the
impact of these changes because the National Research Council (NRC)
is currently reviewing the adequacy of proposed NEXRAD CONUS coverage
in terms of the "no degradation of service" requirement of the
Weather Service Modernization Act.\9

To determine the feasibility and estimated cost of extending the
NEXRAD contract to purchase additional radars, we reviewed the
current contract option for additional radars.  Since this option is
unpriced, we obtained the program office's per unit cost estimate of
acquiring radars if this option was exercised and compared this
estimate to an oral estimate that the contractor provided to the
program office. 

To determine the Air Force NEXRADs' contribution to the national
network, we reviewed the Federal Meteorological Handbook Number 11,
published by the Office of the Federal Coordinator for Meteorology,
to identify the types of data Air Force radars provide to NWS.  In
addition, we interviewed NWS and program officials to determine how
NWS accesses and uses the Air Force radar data, and the impact of NWS
not having the Air Force radar data.  To determine the accessibility
of the Air Force NEXRAD data to civilian forecasters, we interviewed
program office, NWS, and Air Force officials about potential data
restrictions. 

To determine the availability of the Air Force radars, we collected
and reviewed availability data from the Air Force and NWS operational
NEXRAD units and compared these data to the availability requirement
specified in the NEXRAD Joint Operational Requirements document.  We
also interviewed officials from the Air Force's Air Weather Service
at Scott Air Force Base near St.  Louis, Missouri; the Air Force's
Air Combat Command in Hampton, Virginia; NWS' Operational Support
Facility and Weather Forecast Office in Norman, Oklahoma; and seven
DOD bases that operate and maintain NEXRADs. 

We performed our work primarily at the NEXRAD program office, and
NOAA and NWS headquarters in Silver Spring, Maryland.  Our work was
performed from October 1994 to May 1995, in accordance with generally
accepted government auditing standards. 

As requested, the Departments of Commerce and Defense provided
written comments on a draft of this report.  These comments are in
appendixes I and II.  We obtained oral comments from senior FAA
officials, including the NEXRAD program manager.  The comments from
Commerce, Defense, and FAA are presented and evaluated throughout the
report. 


--------------------
\9 At the request of the House Committee on Science, Space, and
Technology (now the House Committee On Science), the Secretary of
Commerce commissioned this review by NRC's NWS Modernization
Committee to ensure that NWS complies with the Weather Service
Modernization Act (Public Law 102-567), which requires, among other
things, that the Secretary of Commerce certify that there is no
degradation of service resulting from office closures associated with
the modernization. 


NEXRAD DEPLOYMENTS HAVE BEEN
REDUCED AND DELAYED BECAUSE OF
REQUIREMENTS CHANGES, BUDGET
CONSTRAINTS, AND BASE CLOSURES
============================================================ Chapter 2

NWS, the Air Force, and FAA plan to meet their needs by deploying 163
NEXRADs worldwide, 12 less than the 175 originally planned.  This
decrease is due to changes in agency requirements, funding
limitations, and military base closures.  It includes a net decrease
in deployments outside of CONUS of 13 and a net increase in CONUS
deployments of 1.  NRC is currently studying the adequacy of proposed
NEXRAD CONUS coverage.  The NRC study director said he expects the
study to identify weaknesses in coverage and potential areas where
additional radars may be needed. 

Most of the 163 radars are to be deployed by the end of fiscal year
1996.  However, FAA plans to delay deploying five of its radars until
at least fiscal year 1997 because of budget constraints.  While these
radars are outside CONUS, NWS officials said they are important to
NWS' ability to track and forecast severe weather. 


   GLOBAL NEXRAD DEPLOYMENTS
   DECREASED BY 12
---------------------------------------------------------- Chapter 2:1

In November 1991, plans for deploying 13 Air Force and 3 FAA radars
were canceled as part of the comprehensive settlement with the NEXRAD
contractor.  The 16 radars were disposed of by selling 1 radar back
to Unisys; using 2 radars for new DOD and FAA requirements; using
portions of 3 for NWS training, research, and maintenance activities;
using portions of 6 systems to provide redundant hardware at remote
locations in order to have backups on hand should equipment fail; and
converting the remaining 4 systems to spares. 

Since the comprehensive settlement, NWS has added three radars and
FAA has added one.  In addition, the Air Force has added a
requirement for one radar, deleted the requirement for another, and
transferred ownership of one of its radars to NWS.  The net result is
the reduction of 12 NEXRADs.  According to program officials, the 163
remaining radars will still satisfy the three agencies' collective
requirements and provide radar coverage equal to or better than the
existing service.  Table 2.1 summarizes the three agencies'
respective changes to the deployment plan. 



                          Table 2.1
           
             Changes to Worldwide NEXRAD Plan by
                            Agency

                          Comprehens
                Original         ive                 Current
              deployment  settlement  Additional  deployment
Agency              plan     changes     changes      plan\a
------------  ----------  ----------  ----------  ----------
NWS                  115           0       +4 \b         119
Air Force             44         -13       -1 \c          30
FAA                   16          -3       +1 \d          14
============================================================
Total                175         -16          +4         163
------------------------------------------------------------
\a These figures include three NWS and three Air Force systems used
for training, research, and logistics purposes. 

\b NWS added NEXRADs at (1) Greer, South Carolina, (2) Jackson,
Kentucky, and (3) the NOAA National Severe Storms Laboratory in
Norman, Oklahoma.  Also, NWS assumed ownership of an Air Force NEXRAD
because Loring Air Force Base, Maine, is closing. 

\c The Air Force canceled the NEXRAD scheduled for Nellis Air Force
Base, Nevada.  It added a NEXRAD at Fort Polk, Louisiana.  It also
transferred ownership of the NEXRAD at Loring Air Force Base, Maine,
to NWS. 

\d FAA added a NEXRAD at South Shore, Hawaii. 


   CONUS NEXRAD DEPLOYMENTS
   INCREASED BY ONE
---------------------------------------------------------- Chapter 2:2

NEXRAD deployments within CONUS originally totaled 143.  These
deployments now total 144.  The net increase of one radar is the
result of an assortment of Air Force and NWS deployment changes over
the last
4 years that reduced Air Force NEXRADs by three and increased NWS
radars by four.  Specifically, the Air Force eliminated the
requirement for two radars as part of the November 1991 settlement
with the contractor.  Later, NWS added four radars, and the Air Force
dropped an additional radar.  FAA never planned to deploy any of its
radars within CONUS.  Table 2.2 summarizes the agencies' respective
changes to the original CONUS deployment plan. 



                          Table 2.2
           
            Changes to NEXRAD Plan Within CONUS by
                            Agency

                          Comprehens
                Original         ive                 Current
              deployment  settlement  Additional  deployment
Agency              plan     changes     changes      plan\a
------------  ----------  ----------  ----------  ----------
NWS                  115           0       +4 \b         119
Air Force             28       -2 \c       -1 \d          25
FAA                    0           0           0           0
============================================================
Total                143          -2          +3         144
------------------------------------------------------------
\a These figures include three NWS and three Air Force systems used
for training, research, and logistics purposes.  NWS and the Air
Force will deploy 116 and 22 CONUS-based operational systems
respectively. 

\b NWS added NEXRADs at (1) Greer, South Carolina, (2) Jackson,
Kentucky, and (3) the NOAA National Severe Storms Laboratory in
Norman, Oklahoma.  Also, NWS assumed ownership of an Air Force NEXRAD
because Loring Air Force Base, Maine, is closing. 

\c The Air Force canceled NEXRADs scheduled for Grissom Air Force
Base, Indiana, and England Air Force Base, Louisiana, due to
projected base closures. 

\d The Air Force canceled the NEXRAD scheduled for Nellis Air Force
Base, Nevada.  It also added a NEXRAD at Fort Polk, Louisiana, and
transferred ownership of the NEXRAD at Loring Air Force Base, Maine,
to NWS. 


   NEXRAD DEPLOYMENT CHANGES DUE
   TO BUDGET CONSTRAINTS,
   REQUIREMENT CHANGES, AND BASE
   CLOSURES
---------------------------------------------------------- Chapter 2:3

In total, 23 changes have been made to the original NEXRAD deployment
schedule.  Of these 23, 9 changes were made because budget
constraints prevented the purchase of planned radars;\1 9 were
because the requirement for a radar at a given site was deleted or a
new requirement was added;\2 4 were because the military base that
was to receive the radar was identified for closure; and 1 was
because the frequency over which the radar's signal is transmitted
was unavailable at the planned overseas deployment location. 

Of the 23 changes, 3 related to NWS radars, 16 related to Air Force
radars (11 overseas), and 4 related to FAA radars.  Table 2.3
identifies the changes by agency, sites affected, type, and reason
for each of the 23 changes. 



                          Table 2.3
           
           Locations of and Reasons for Changes in
                  NEXRAD Deployment Schedule

Agency    Site                 Change    Reason
--------  -------------------  --------  -------------------
Air       Central Germany      Deletion  Budget constraint
Force

Air       Eastern Germany      Deletion  Budget constraint
Force

Air       Western Germany      Deletion  Budget constraint
Force

Air       Aviano Air Base,     Deletion  Budget constraint
Force     Italy

Air       Crotone Air Base,    Deletion  Budget constraint
Force     Italy

Air       Camp New Amsterdam,  Deletion  Budget constraint
Force     Netherlands

Air       Zaragoza Air Base,   Deletion  Budget constraint
Force     Spain

Air       East United Kingdom  Deletion  Budget constraint
Force

Air       West United Kingdom  Deletion  Budget constraint
Force

Air       Clark Air Base,      Deletion  Base closure
Force     Philippines

Air       Yokota Air Base,     Deletion  Frequency
Force     Japan                          unavailable

Air       England Air Force    Deletion  Base closure
Force     Base, Louisiana

Air       Grissom Air Force    Deletion  Base closure
Force     Base, Indiana

Air       Nellis Air Force     Deletion  Requirement deleted
Force     Base, Nevada

Air       Fort Polk,           Addition  New requirement
Force     Louisiana

Air       Loring Air Force     Transfer  Base closure
Force     Base, Maine

FAA       McGrath, Alaska      Deletion  Requirement deleted

FAA       Bering Sea, Alaska   Deletion  Requirement deleted

FAA       Site to be           Deletion  Requirement deleted
          determined

FAA       South Shore, Hawaii  Addition  New requirement

NWS       Greer, South         Addition  New requirement
          Carolina

NWS       Jackson, Kentucky    Addition  New requirement

NWS       National Severe      Addition  New requirement
          Storms Laboratory,
          Norman, Oklahoma
------------------------------------------------------------

--------------------
\1 These budget constraints arose because radar unit cost increases
associated with the comprehensive settlement prevented the Air Force
from buying as many radars as originally planned while still staying
within its program budget. 

\2 Program office officials stated that FAA and the Air Force deleted
requirements for three and one radars, respectively, because
justification for the original requirements was later invalidated. 


   NRC IS STUDYING THE ADEQUACY OF
   PLANNED NEXRAD COVERAGE
---------------------------------------------------------- Chapter 2:4

According to NWS, the current tri-agency plan for deploying 138
operational NEXRADs within CONUS will provide radar coverage equal to
or better than existing coverage.  However, at the request of the
House Committee on Science, Space, and Technology (now the House
Committee on Science), the Secretary of Commerce commissioned an
independent review of NEXRAD CONUS coverage.  The Committee did this
to ensure that NWS complies with the Weather Service Modernization
Act, which requires, among other things, that the Secretary of
Commerce certify that the modernization result in no degradation in
service. 

The Secretary of Commerce asked the NRC Committee on National Weather
Service Modernization to perform this study.  The NRC study director
said he expects the study, which NRC plans to issue in June 1995, to
identify potential areas where coverage is degraded and where
additional radars may be needed.  Because of NRC's study, we did not
address the impact of the reduced number of radars on the three
agencies' radar coverage objectives. 


   FIVE FAA NEXRADS TO BE PLACED
   IN TEMPORARY STORAGE
---------------------------------------------------------- Chapter 2:5

FAA currently plans to place 5 of its 14 NEXRADs in storage for
probably more than a year because deploying the radars is not a
funding priority.\3 The five FAA NEXRADs are paid for and scheduled
for delivery around June 1996.  However, FAA's fiscal year 1996
budget request does not include the $18 million needed to deploy
them.  According to the FAA program manager for NEXRAD, the earliest
that FAA may request funds is fiscal year 1997; however, the program
manager does not expect funding approval at that time.  The program
manager attributed the funding shortfall to deployment costs that
were higher than expected. 

NWS did not plan to place NEXRADs in these locations because NWS does
not have a mission requirement for radar coverage outside CONUS. 
However, according to NWS officials, data from planned NEXRADs in
Alaska, Hawaii, and the Caribbean will be used by NWS to enhance its
ability to provide timely and accurate forecasts and warnings.  For
instance, according to the NOAA Assistant Administrator for Weather
Services, two of these radars in the Caribbean would allow NWS to
better track and monitor hurricanes approaching the United States. 
Despite NWS' desire to have radars in these areas, the NOAA Assistant
Administrator for Weather Services stated that should FAA decide not
to deploy these radars, he is not sure whether NWS would choose to do
so. 


--------------------
\3 The five sites are (1) Georgetown, Bahamas, (2) Grand Turk,
British West Indies, (3) South Shore, Hawaii, (4) Kohala, Hawaii, and
(5) Nome, Alaska. 


   CONCLUSIONS
---------------------------------------------------------- Chapter 2:6

NRC is expected to identify weaknesses in NEXRAD's national coverage
that may suggest that NWS buy additional radars.  However, five FAA
radars planned for Alaska, Hawaii, and the Caribbean are paid for and
will be warehoused indefinitely.  While NWS does not have a mission
requirement for radars in these locations, NWS officials stated that
radars in these geographic areas would enhance its ability to provide
forecasts and warnings.  Consequently, NWS is uncertain whether it
would choose to deploy radars in these areas should FAA decide not
to. 


   RECOMMENDATION
---------------------------------------------------------- Chapter 2:7

We recommend that the Secretary of Commerce direct the NOAA Assistant
Administrator for Weather Services to not purchase additional radars
to address any weaknesses in radar coverage that may result from the
NRC study until assessing FAA's plans for deploying the five radars
scheduled for storage and NWS' mission requirements for NEXRADs in
these areas. 


   AGENCY COMMENTS AND OUR
   EVALUATION
---------------------------------------------------------- Chapter 2:8

In a draft of this report, we proposed that the NOAA Assistant
Administrator for Weather Services assess the operational impact of
FAA's delays in deploying the five radars and, on the basis of this
assessment, take the necessary steps to ensure that NWS radar
coverage needs are met.  Both the Department of Commerce and the FAA
NEXRAD program manager stated that NWS does not have a mission
requirement for radars outside CONUS.  We have incorporated these
comments in the section describing FAA's plans for storing five
radars and subsequently revised our recommendation in this chapter. 

The Department of Commerce also stated that we did not accurately
characterize the NRC study.  Commerce stated the NRC study is of
proposed NEXRAD radar coverage and consolidation of field offices to
ensure the "no degradation of service" requirement of the Weather
Service Modernization Act.  We have clarified references to the NRC
study in our report to state that the study is of proposed NEXRAD
coverage as compared to premodernization radar coverage. 

Commerce also noted that NOAA believes that the currently planned
NEXRAD network will provide radar coverage equal to or better than
the existing service, and that NOAA is aware of gaps in modernized
radar coverage.  Our report has been modified to reflect this. 

Commerce also stated that the number of systems to be purchased and
deployed still stands at 175, rather than the 163 we reported.  It
explained the disposition of the radars resulting from the
comprehensive settlement.  We agree with this explanation and have
added a clarifying statement in this chapter.  However, the number of
systems to be deployed is 163.  This number, which is based on our
review of the NEXRAD deployment schedule, is consistent with program
office documentation and with the total presented by the Under
Secretary for Oceans and Atmosphere, before the Senate Subcommittee
on Science, Technology, and Space, Committee on Commerce, Science,
and Transportation in his January 31, 1995, testimony. 


UNIT COST OF ANY ADDITIONAL
NEXRADS MAY BE SUBSTANTIALLY
HIGHER
============================================================ Chapter 3

The program office has the option to buy up to 20 additional NEXRADs
through August 1996 or the date that the last radar is delivered,
whichever occurs first.  However, this option is not priced, meaning
that it is subject to negotiation.  Program officials estimate that
if the option was exercised, the unit cost could be as much as three
times the cost of those currently under contract.\1 Because the
program office has no current plans for buying additional radars, it
has not reassessed the cost-effectiveness of the more expensive
radars.  However, as mentioned in chapter 2, NRC's study is expected
to report on gaps in NEXRAD national coverage that could suggest the
purchase of more radars via this contract option. 


--------------------
\1 The actual unit cost estimate is not disclosed in this report
because the information may be acquisition sensitive. 


   ADDITIONAL NEXRADS MAY COST AS
   MUCH AS THREE TIMES CURRENT
   COSTS
---------------------------------------------------------- Chapter 3:1

The program office agreed to an unpriced option because no firm
requirements for additional radars existed at the time the current
contract option was negotiated.  Further, the contractor opposed
pricing the option because of the uncertainty of future costs, such
as restarting production lines.  According to program officials, the
contractor would not agree to a priced option without knowing how
many or when additional radars would be ordered. 

While the price of the additional NEXRADs, should they be required,
will ultimately be subject to negotiation, the program office
estimates that the hardware and software costs for each radar could
be as much as three times that of the mean cost of radars currently
under contract.  The program office based this estimate on a recently
expired, priced contract option that had a not-to-exceed price.  It
then factored in additional costs due to breaks in production.  These
disruptions increase costs because the longer the government waits to
exercise the option, the greater the chances that the contractor and
its subcontractors will have shut down part or all of their
production lines and started work for other clients.  According to
program officials, this has already occurred as the subcontractor
responsible for the transmitter has closed its production line. 

The program office estimate is consistent with a nonbinding, verbal
estimate that the contractor provided to the program office.  This
contractor estimate, however, assumed that the option would be
exercised in the first quarter of fiscal year 1995.  Exercising the
option later, according to program officials, would result in a
higher unit cost.  In addition, the program office based its estimate
on the purchase of 12 radars.  Purchasing fewer radars would also
increase the unit cost because manufacturing start-up costs for a
smaller order would be allocated over fewer units. 

Office of Management and Budget Circular A-11\2 requires that
agencies use benefit-cost analyses to evaluate contemplated
investments in information technology.  The purpose of these analyses
is to maximize an agency's return on its information technology
dollar.  In addition, these analyses are not to be one-time exercises
performed at the beginning of a project.  Instead, it is fiscally
prudent to redo these analyses whenever expected benefits or
estimated costs change significantly.  Without reassessing a system's
payback in the event of sizeable cost growth, poor investment
decisions can result. 

As mentioned in chapter 2, weaknesses in NEXRAD's national coverage
that are identified by NRC may suggest that NWS buy additional
radars.  Should this occur, the existing contract option would be an
available vehicle for doing so, subject to applicable procurement
regulations.  The program office has not assessed the
cost-effectiveness of purchasing the additional units because
currently no requirements exist for additional radars.  If this
changes, program officials agreed that it would be wise in deciding
whether or not to exercise the contract option, to reassess the
benefits to be derived from the additional radars against their
higher cost. 


--------------------
\2 Preparation and Submission of Budget Estimates, July 1994. 


   CONCLUSION
---------------------------------------------------------- Chapter 3:2

Additional NEXRADs could cost as much as three times more than units
currently under contract.  At this price, the radars' benefits may
not exceed their cost, and thus buying more may not be worth the
investment.  While NWS officials acknowledged the value in
reassessing the benefits versus the costs of the more expensive
radars before exercising the NEXRAD contract option, they did not
commit to doing so since requirements for additional radars currently
do not exist. 


   RECOMMENDATION
---------------------------------------------------------- Chapter 3:3

We recommend that the Secretary of Commerce direct the NOAA Assistant
Administrator for Weather Services to ensure that any radars bought
in response to NRC's national radar coverage findings are
cost-beneficial, given that their unit cost could be substantially
higher than those already purchased. 


   AGENCY COMMENTS AND OUR
   EVALUATION
---------------------------------------------------------- Chapter 3:4

In its comments, the Department of Commerce partially concurred with
our recommendation to ensure that any radars bought in response to
NRC's coverage findings are cost-beneficial.  Commerce agreed that
the requirements to procure additional radars need to be justified;
however, it requested that the recommendation be modified to reflect
only NWS' core mission and the Weather Service Modernization Act
requirements.  The NRC study director told us that the act was used
as criteria in its evaluation.  We, therefore, did not modify our
recommendation. 

In addition, Commerce stated that while the existing contract option
is an available vehicle for buying additional radars should the need
arise, it is by no means obvious that this vehicle would be
exercised.  Commerce added that under federal acquisition
regulations, the government would have to determine the best approach
for acquiring additional systems.  We have modified the report to
address these concerns. 


AIR FORCE NEXRADS ARE INTEGRAL TO
NWS MISSION PERFORMANCE AND
ACCESSIBLE TO NWS FORECASTERS
============================================================ Chapter 4

The Air Force's NEXRADs play an integral role in NWS' ability to
issue accurate and timely weather forecasts and warnings.  These
radars, which provide essential primary, backup, and supplemental
coverage, are vital in supporting the NWS CONUS-based network.  Seven
of the 22 CONUS-based operational Air Force NEXRADs provide primary
NWS radar coverage.  The Air Force NEXRADs also provide backup
coverage in the event an NWS NEXRAD is not operating, according to
NWS officials.  Further, the Air Force NEXRADs augment other NEXRADs
by covering severe weather events from different angles, thus
strengthening NWS' watch and warning capabilities by providing
additional insights into the event's behavior.  To illustrate, the
Altus Air Force Base NEXRAD in Frederick, Oklahoma, is critical to
coverage in parts of Oklahoma and Texas because adjacent NWS radars
reach only marginally into these areas. 

NWS has unrestricted access to all Air Force, CONUS-based NEXRAD
products.  According to a tri-agency agreement, all NEXRADs "shall be
operated to satisfy the integrated needs of all three agencies."
Further, each agency is to "support, to the maximum extent possible,
the data, products, and operational requirements of the others,
consistent with the capabilities and mission priorities of that
agency." All CONUS Air Force sites are specifically required to
"provide assistance to NWS offices by providing access to weather
radar data for gaps in the National Weather Radar Network."

NWS' access to the Air Force's CONUS-based NEXRADs is accomplished
via dedicated and dial-up communication lines.  Currently, 13 weather
offices have dedicated lines to Air Force NEXRADs.  The 13 offices
are generally the closest ones geographically to the Air Force
radars.  These dedicated lines operate at 9.6 kilobits per second. 
Each Air Force NEXRAD also provides three to four dial-up
communication ports for use by other NWS field offices.  These lines
are also 9.6 kilobits per second and are reserved for NWS use. 


   AGENCY COMMENTS AND OUR
   EVALUATION
---------------------------------------------------------- Chapter 4:1

In its comments, the Department of Commerce agreed that the Air Force
NEXRADs are essential to NWS' ability to issue quality forecasts and
warnings, and that our report adequately describes most of the
meteorological aspects of this need.  However, Commerce stated that
the Air Force's radars need to be reconfigured to accept input from
real-time rain gauges to meet the needs of the River Forecast
Centers.  Although we understand NWS' desire for all Air Force
CONUS-based NEXRADs to accept input from rain gauges, the tri-agency
agreement does not establish a requirement for Air Force NEXRADs to
perform this function.  The agreement states that a CONUS-based Air
Force NEXRAD "shall not execute the rain gauge data acquisition
function." We confirmed with NWS officials that the rain gauge data
acquisition function is still not an Air Force NEXRAD requirement. 

In its comments, the Department of Defense disagreed with our
conclusion on the role Air Force radars play in the National Weather
Radar Network.  Specifically, Defense disagreed with our use of the
term backup when associated with the Air Force's radars, because it
implies that the Air Force's radars are integral parts of the
national network.  Defense stated that the tri-agency documentation
defines NWS' radars as network sites and the Air Force's as
supplemental sites, and that therefore we should not refer to the Air
Force CONUS-based NEXRADs as backup systems, but rather as
supplemental sites. 

We have decided not to use the term supplemental when referring to
these Air Force radars because they provide primary, backup, and
supplemental coverage.  For example, the tri-agency documentation
specifically states that CONUS-based supplemental Air Force sites are
to "provide assistance to NWS offices by providing access to weather
radar data for gaps in the National Weather Radar Network." Today, 7
of the 22 CONUS-based operational Air Force NEXRADs provide the sole
radar coverage for certain geographic areas.  NWS officials also told
us that NWS uses data from Air Force NEXRADs to provide backup
coverage and to supplement data from their NEXRADs, as well as to
provide primary coverage.  On the basis of this combination of
primary, backup, and supplemental coverage, the Air Force CONUS-based
NEXRADs contribute considerably to the national NEXRAD network and
are indeed integral parts of the national network. 


AIR FORCE AND NWS NEXRADS MAY NOT
BE AVAILABLE WHEN NEEDED
============================================================ Chapter 5

To effectively support NWS' requirements for CONUS radar coverage, it
is important that the Air Force's NEXRADs meet all specified
requirements.  However, Air Force data show that some of its radars
are performing below the tri-agency system availability requirement. 
Moreover, because Air Force availability data are unreliable and
appear to be overstated, it is uncertain whether any of its radars
are meeting availability requirements.  Opportunities exist to
improve Air Force NEXRAD availability performance by increasing radar
operator awareness of the availability requirement and streamlining
the spare parts logistic process.  In addition, because NWS currently
does not calculate and monitor availability by site, there is no way
to determine whether each NWS NEXRAD is meeting the availability
requirement. 


   AIR FORCE NEXRADS ARE NOT
   MEETING SYSTEM AVAILABILITY
   REQUIREMENTS
---------------------------------------------------------- Chapter 5:1

NWS, the Air Force, and FAA have specified that each NEXRAD unit must
be operationally available 96 percent of the time.\1 According to NWS
officials, the 96 percent requirement is based on an analysis that
considered factors such as equipment reliability, staff costs, and
spare parts costs.  These officials stated that the additional costs
(for example, redundant systems, spare parts, and additional
maintenance technicians) associated with achieving availability above
96 percent were not worth the added benefits. 

Many of the Air Force's NEXRADs are not meeting the 96-percent
availability requirement.  Since January 1994, the reported percent
of operational Air Force NEXRADs meeting this requirement each month
has ranged from 38 to 90 percent (see figure 5.1). 

   Figure 5.1:  Number of Air
   Force NEXRADs Meeting
   96-Percent Availability
   Requirement

   (See figure in printed
   edition.)

Note:  Numbers in parentheses indicate the number of Air Force
NEXRADs in operation. 

However, the data upon which these availability statistics are based
are unreliable and appear to be overstated.  For example, the Air
Force data provided by the Air Weather Service show that nine NEXRADs
have been operationally available 100 percent of the time for 4 or
more consecutive months.  This is highly unlikely considering that,
according to NWS' Chief Logistician, the radars are likely to fail an
average of 52 times a year, or about 4 times per month.  We contacted
six of the nine sites reporting 4 or more consecutive months of 100
percent availability and found that three of the sites had
significant outages during this time.  For example, data for Eglin
Air Force Base show 100 percent availability for September through
December 1994, but Eglin radar officials stated that their radar was
available for these 4 months only 87, 78, 79, and 87 percent of the
time, respectively.  Similarly, data for Dyess Air Force Base show
100 percent availability for February 1995, but Dyess officials
stated that their radar was available only 81 percent of the time in
February 1995.  Also, data for Robins Air Force Base show 100 percent
availability for September 1994, but Robins officials stated that
their radar was unavailable for 12 days in September 1994 while they
were waiting for a replacement part. 

According to an internal Air Force report dated October 1994,\2 the
availability data inconsistencies are not unique to NEXRAD, but
rather extend to all Air Force command, control, communications, and
computer (C4) systems.  The report stated that not all system
performance data are being collected and reported on C4 systems. 
This occurs, according to the report, because C4 personnel do not
adequately understand and are not sufficiently trained in the
maintenance data collection process and because the systems
collecting and reporting the performance data have software problems
that have gone unchecked.  The report recommends that the Air Force
establish a team to address these problems. 


--------------------
\1 The NEXRAD Joint Operational Requirements define availability as
the time that the system is operating satisfactorily, expressed as a
percentage of total time (the time the system is operating
satisfactorily plus the time the system is down).  Downtime includes
corrective and preventive maintenance time and delays encountered due
to the delivery of needed spare parts.  Most definitions of
availability exclude scheduled downtime, such as preventive
maintenance. 

\2 Command, Control, Communications, and Computer Data Reporting
Problem Process Action Team Summary Report, October 1994. 


   AIR FORCE NEXRAD OPERATORS AND
   MAINTAINERS ARE UNAWARE OF THE
   TRI-AGENCY AVAILABILITY
   REQUIREMENT
---------------------------------------------------------- Chapter 5:2

Air Force operators and maintainers are unaware of the 96-percent
availability requirement and, therefore, have no way of knowing if
actual radar performance is satisfactory.  We contacted operations
and maintenance technicians at 7 of the 20 Air Force sites that have
an operational NEXRAD, and found that none of the technicians were
aware of the tri-agency availability requirement.  We then examined
the availability data for these seven sites, and found that six of
the seven were not meeting the requirement 13 to 44 percent of the
time.  Although the seventh site's data show it to be meeting the
requirement, these data show 8 consecutive months of 100-percent
availability, and thus as discussed earlier, are clearly unreliable
and appear to be overstated. 


   AIR FORCE PROCESS FOR OBTAINING
   SPARES IS INEFFICIENT, BUT
   IMPROVEMENTS ARE UNDERWAY
---------------------------------------------------------- Chapter 5:3

The supply and logistics process that Air Force sites follow to
obtain NEXRAD spare parts is inefficient and more time-consuming than
NWS' process.  Specifically, when an NWS radar needs a replacement
part that is not on hand, technicians request the part directly from
NWS' National Logistics Supply Center in Kansas City, Missouri.  The
Center, in turn, sends the part directly to the requesting site.  In
contrast, Air Force technicians must request the part from the Air
Force's Sacramento Air Logistics Center, which records the
transaction and electronically passes the request on to NWS' Office
of Systems Operations for processing.\3 At this point the order is
then processed in the same manner as an NWS order. 

This problem is even more severe for the four CONUS Air Force NEXRADs
located near Army installations, where parts that are ordered are
first sent to the nearest Air Force base, which records the
transaction and then ships the parts to the Army base.  The distance
between the Air Force and Army locations varies from 90 to 330 miles,
causing an additional delay of at least 4 to 12 hours.  Air Force
officials told us that they strive to keep their NEXRADs operational
100 percent of the time; however, the current system component
failure rates and the logistics process Air Force NEXRAD sites must
follow to obtain spare parts from NWS' National Logistics Supply
Center make it difficult to achieve the 96-percent availability
requirement.  Figure 5.2 shows the additional steps required for the
Air Force and Army NEXRAD sites to obtain needed spare parts. 

   Figure 5.2:  NWS and Air Force
   NEXRAD Logistics Processes

   (See figure in printed
   edition.)

NWS' Chief Logistician also identified several other logistics
inefficiencies that contributed to the limited availability of the
Air Force's NEXRADs.  For example, until October 1994, the Air
Logistics Center did not operate on evenings, nights, and weekends. 
Also, NWS did not have accurate and complete addresses for all Air
Force NEXRAD sites, and thus parts would sometimes arrive at the
base, but would not be delivered to the proper maintenance location. 
NWS' Chief Logistician stated that these problems have added days to
the parts delivery process.  Also contributing to the Air Force
radars' availability shortfalls, according to the Chief Logistician,
has been a limited supply of parts at Air Force NEXRAD sites. 

To address these logistics concerns, NWS and Air Force officials
established a logistics working group in September 1993.  Members of
this group stated that a number of the procedural problems have been
eliminated.  For example, they said that the group has automated the
parts request process through the Air Logistics Center to the Office
of Systems Operations so that parts orders can be placed 24 hours a
day for emergency requisitions.  They also said that the group
provided NWS with complete addresses for all base supply
organizations servicing Air Force NEXRAD sites.  Also, they said that
steps are underway to improve the stocking of on-site spares.  In
addition to these initiatives, they cited steps underway to provide
better service to remote Army NEXRAD sites.  For example, the Air
Force is examining whether a common carrier can deliver parts to Army
NEXRAD sites more quickly. 


--------------------
\3 The requisitioning process through the Sacramento Air Logistics
Center is required for financial tracking purposes. 


   NWS DOES NOT MONITOR
   AVAILABILITY DATA BY SITE
---------------------------------------------------------- Chapter 5:4

NWS does not calculate operational availability for each NEXRAD site,
and thus does not know whether each radar site is meeting the
availability requirement.  Although NWS records radar downtime by
site, it only uses this information to calculate the average
availability of all sites, and it only monitors radar availability
performance on this basis.  It does not use this information to
calculate and monitor site-specific availability.  NWS officials
stated that monitoring overall availability, rather than individual
availability, is useful in determining such things as spare parts
usage and maintenance staffing trends. 

While we agree that such data on availability are useful, aggregate
data do not allow NWS to determine whether each radar meets the
required 96-percent availability requirement.  The Office of Systems
Operations Director agreed that it was important to track
availability by site, and said that NWS would monitor availability on
a site-by-site basis in the future.  However, this change has yet to
occur, and a time frame for doing so has not been established. 


   CONCLUSIONS
---------------------------------------------------------- Chapter 5:5

Air Force NEXRADs fall short of system availability requirements,
thus increasing the risk that NWS will not have the data it needs to
accurately and quickly predict severe weather.  This risk is more
severe than the data show because the Air Force availability data are
unreliable and because NWS is not even analyzing and monitoring each
radar's availability.  While we believe that the steps described by
Air Force and NWS officials to improve NEXRAD availability are
reasonable, we did not verify that they have occurred.  Further, the
actual impact that each will have on system availability still
remains to be seen, and the availability of NWS' sites also remains
unknown.  Until NWS and the Air Force meet stated radar availability
requirements, lives and property are at greater risk. 


   RECOMMENDATIONS
---------------------------------------------------------- Chapter 5:6

We recommend that the Secretary of the Air Force direct the Air Force
Director of Weather to improve the reliability of the Air Force
NEXRAD availability data and to correct any shortfalls that these
data show. 

We recommend that the Secretary of Commerce direct the NOAA Assistant
Administrator for Weather Services to begin analyzing and monitoring
system availability data on a site-specific basis for its operational
NEXRADs and correct any shortfalls in system availability that this
analysis shows. 


   AGENCY COMMENTS AND OUR
   EVALUATION
---------------------------------------------------------- Chapter 5:7

In its comments, the Department of Defense concurred with our
recommendation concerning the Air Force and stated that it will
develop management actions by mid-fiscal year 1996 to improve NEXRAD
availability and the reliability of DOD data. 

The Department of Commerce also concurred with our recommendation
concerning the analysis and monitoring of system availability data on
a site-specific basis and stated that NWS is taking steps to allowing
it to analyze and monitor system availability on a site-specific
basis. 




(See figure in printed edition.)Appendix I
COMMENTS FROM THE DEPARTMENT OF
COMMERCE
============================================================ Chapter 5



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)


Note:  The Department of Commerce also included in its comments a
statement that the existing contract option is an available vehicle
for buying additional radars, and, thus, any references to the
estimated costs of exercising this option are potentially acquisition
sensitive.  We have deleted this statement from the Department's
comments because it identified the estimated unit cost. 




(See figure in printed edition.)Appendix II
COMMENTS FROM THE DEPARTMENT OF
DEFENSE
============================================================ Chapter 5


MAJOR CONTRIBUTORS TO THIS REPORT
========================================================= Appendix III


   ACCOUNTING AND INFORMATION
   MANAGEMENT DIVISION,
   WASHINGTON, D.C. 
------------------------------------------------------- Appendix III:1

Rona B.  Stillman, Chief Scientist for Computers and Communications
Randolph C.  Hite, Assistant Director
Keith A.  Rhodes, Technical Assistant Director
David A.  Powner, Evaluator-in-Charge
Shane D.  Hartzler, Communications Analyst


   DENVER REGIONAL OFFICE
------------------------------------------------------- Appendix III:2

Joseph P.  Sikich, Staff Evaluator
