National Weather Service: Budget Events and Continuing Risks of Systems
Modernization (Testimony, 03/04/98, GAO/T-AIMD-98-97).

Pursuant to a congressional request, GAO discussed its work regarding
the events surrounding the fiscal year (FY) 1997 budget of the National
Weather Service (NWS) and continuing risks of systems modernization.

GAO noted that: (1) NWS was able to operate within its appropriated FY
1997 budget level, but had to do so with fewer funds than in FY 1996;
NWS referred to this difference in available funds as a budget
shortfall; (2) the shortfall was reported in different amounts at
different points in time, such that some members of Congress experienced
confusion as to the actual amount; (3) two events associated with the
shortfall and potential ways to accommodate it raised concerns among
department officials; (4) the first centered on a NWS request that the
National Oceanic and Atmospheric Administration (NOAA) reprogram
funds--and NWS' intention to fill critical field office vacancies before
approval of that request; (5) NOAA informed NWS officials that the
vacancies could not be filled because the reprogramming request had not
been approved; (6) the second event involved NWS' attempt to obtain NOAA
certification approval to consolidate, automate, or close weather
service offices; (7) the NWS modernization program--vital to improved
operations--is one of the largest in the federal government; (8) NWS
systems modernization includes 4 major programs; (9) ongoing problems,
both developmental and operational, have, however, surrounded the
modernization efforts; (10) as part of its streamlining efforts, NWS has
identified various weather offices nationwide, including the Erie area
and Northwestern Pennsylvania, for closure, with responsibilities to be
taken over by neighboring offices; (11) however, there were concerns
that there would be a gap in lake-effect snowcoverage; and (12) because
of this concern, the Director of NWS' Office of Meterology recommended
that a new radar be installed for the Erie area.

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

 REPORTNUM:  T-AIMD-98-97
     TITLE:  National Weather Service: Budget Events and Continuing 
             Risks of Systems Modernization
      DATE:  03/04/98
   SUBJECT:  Weather forecasting
             Reprogramming of appropriated funds
             Systems conversions
             Computer software verification and validation
             Radar equipment
             Warning systems
             Budget administration
             Personnel recruiting
             Federal downsizing
IDENTIFIER:  NWS Advanced Weather Interactive Processing System
             NWS Next Generation Weather Radar
             NWS Automated Surface Observing System
             NOAA Geostationary Operational Environmental Satellite
             
******************************************************************
** 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.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
** A printed copy of this report may be obtained from the GAO   **
** Document Distribution Center.  For further details, please   **
** send an e-mail message to:                                   **
**                                                              **
**                                            **
**                                                              **
** with the message 'info' in the body.                         **
******************************************************************


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


Before the Subcommittee on Energy and Environment, Committee on
Science, House of Representatives

For Release on Delivery
Expected at
10 a.m.
Wednesday,
March 4, 1998

NATIONAL WEATHER SERVICE - BUDGET
EVENTS AND CONTINUING RISKS OF
SYSTEMS MODERNIZATION

Statement of Joel C.  Willemssen
Director, Civil Agencies Information Systems
Accounting and Information Management Division

GAO/T-AIMD-98-97

GAO/AIMD-98-97T


(511447)


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

  AWIPS - Advanced Weather Interactive Processing System
  FSL - Forecast Systems Laboratory
  GOES - Geostationary Operational Environment Satellite
  NEXRAD - Next Generation Weather Radar
  NOAA - National Oceanic and Atmospheric Administration
  NWS - National Weather Service

============================================================ Chapter 0

Mr.  Chairman and Members of the Subcommittee: 

We are pleased to be here today to discuss our work regarding events
surrounding the fiscal year 1997 budget of the National Weather
Service (NWS), a component of the National Oceanic and Atmospheric
Administration (NOAA) within the Department of Commerce.  Our report
on the matter is being released at this hearing.\1 We will also be
reporting on the status of information first presented to this
Subcommittee in 1995 on the Advanced Weather Interactive Processing
System (AWIPS), a major component of the Weather Service's systems
modernization program.\2 Finally, we will briefly update another
report on the Weather Service issued last year, dealing with radar
and weather service coverage to northwestern Pennsylvania with the
closure of the Erie weather service office.\3


--------------------
\1 National Weather Service:  Events Surrounding Fiscal Year 1997
Budget (GAO/AIMD-98-69, Mar.  4, 1998). 

\2 Weather Service Modernization:  Despite Progress, Significant
Problems and Risks Remain (GAO/T-AIMD-95-87, Feb.  21, 1995). 

\3 National Weather Service:  Modernization Activities Affecting
Northwestern Pennsylvania (GAO/AIMD-97-156, Sept.  26, 1997). 


   NWS FISCAL YEAR 1997 BUDGET
   "SHORTFALL"
---------------------------------------------------------- Chapter 0:1

The Weather Service was able to operate within its appropriated
fiscal year 1997 budget level, but had to do so with fewer funds than
in fiscal year 1996; NWS referred to this difference in available
funds as a budget "shortfall." The "shortfall" was reported in
different amounts at different points in time, such that some members
of Congress experienced confusion as to the actual amount.  Erroneous
assumptions and miscommunication between the Weather Service and NOAA
further clouded the issue. 

The differing "shortfall" amounts reported to Congress depended upon
definition as well as time.  Congressional staff briefed in February
1997 were told that the "shortfall" amount was $27.5 million; 2
months later, the figure had risen to $42.2 million.  Finally, at a
hearing last May, the amount given was $47.4 million.  NOAA and NWS
officials explained the discrepancy by saying that the varying
amounts responded to specific questions at particular points in time,
and did not necessarily include all known elements of the
"shortfall."

Two events associated with the "shortfall" and potential ways to
accommodate it raised concerns among department officials.  The first
centered on a Weather Service request that NOAA reprogram funds--and
NWS' intention to fill critical field office vacancies before
approval of that request.  While NWS counted on such approval and the
availability of funds for these positions, NOAA, however, informed
Weather Service officials that the vacancies could not be filled
because the reprogramming request had not yet been approved. 

The second event involved NWS' attempt to obtain NOAA certification
approval to consolidate, automate, and/or close weather service
offices.  Before any NWS office can be closed, the Secretary of
Commerce must certify that the affected geographic areas will not
experience a degradation of weather service.  Upon learning that it
would not be able to fill critical field vacancies, NWS recommended
to NOAA that selected certification packages--27 out of 83--be held
back because, according to a Weather Service official, not filling
the vacancies would result in degraded weather services at some
locations.  NOAA subsequently held back all certification packages.\4

On June 25, 1997--just 5 days after NWS' Deputy Assistant
Administrator for Modernization recommended holding back the 27
packages certifying nondegradation of service--NOAA's Assistant
Administrator for Weather Services was reassigned.  This action was
taken by NOAA's Under Secretary, citing conflicting information from
NWS on how it would provide essential weather services while
recognizing the need for government agencies to reduce costs. 

In addition, a special adviser was tasked with reporting to the
Secretary of Commerce on the fiscal requirements to operate a
modernized NWS during fiscal years 1998 and 1999.  One result of the
report is the department's plan to hire a Chief Financial Officer for
NWS. 


--------------------
\4 We recently learned that NWS plans to forward about 80
certification packages this year. 


   WEATHER SERVICE MODERNIZATION
   AND THE ADVANCED WEATHER
   INTERACTIVE PROCESSING SYSTEM
---------------------------------------------------------- Chapter 0:2

The mission of the National Weather Service is of critical importance
to all Americans--a fact of which we are reminded whenever
hurricanes, floods, tornadoes, or other severe weather threatens. 
The service's modernization program--vital to improved operations--is
one of the largest in the federal government.\5 Begun in the early
1980s, its goals are to (1) achieve more uniform weather services
nationwide, (2) improve forecasting, (3) provide more reliable
detection and prediction of severe weather and flooding, (4) permit
more cost-effective operations, and (5) achieve higher productivity. 

The Advanced Weather Interactive Processing System--AWIPS--is the
linchpin of the NWS modernization that will integrate--for the first
time--satellite, radar, and other data to support forecaster
decision-making and communications.  Along with AWIPS, the
modernization includes three other major programs:  the Automated
Surface Observing System, the Next Generation Geostationary
Operational Environmental Satellite (GOES-Next), and the Next
Generation Weather Radar (NEXRAD). 

Ongoing problems--both developmental and operational--have, however,
surrounded the modernization.  For example, new radars have not
always been up and running when severe weather threatened, and
ground-based sensors have fallen short of performance and user
expectations.  Performance problems such as these, along with
developmental problems relating to cost and schedule, have led us to
voice continuing concern--and make specific recommendations--over the
past several years.\6 Many of these concerns remain. 

As a result of its continuing problems, Weather Service modernization
has been included--both in 1995 and 1997--on our list of high-risk
government programs.\7 One key element of risk is the lack of an
overall systems architecture--a guiding blueprint to be followed in
the development of a systems modernization to ensure interoperability
and cost-effective maintenance.  As of now, an overall modernization
architecture is not expected before September 30 of this year. 

Through AWIPS--designed to be the "central nervous system" of a
modernized Weather Service--NWS expects to tap a reservoir of data
from its new observing systems, data that its current, aging system
cannot handle.  AWIPS forecaster workstations are being developed
incrementally, in a series of six modules called "builds."\8
Operational testing and evaluation was recently concluded on build 3
and the results showed that users have generally been pleased with
system modules already in use.  Yet full utilization of data from the
observing systems has been prevented by the continuing problems and
delays.  Therefore, exactly when AWIPS will be fully deployed and
functioning properly, at what cost, and with what level of
capability, remain unknown.  Until that time, the Weather Service
will not fully reap the rewards of an investment that has spanned 15
years and is approaching $4.5 billion. 

Although AWIPS is expected to greatly improve NWS' ability to provide
weather services, serious risks continue to be associated with the
system's costs, schedule, development, and maintenance.  The area of
cost, for example, has been in flux.  The cost to develop AWIPS was
estimated at $350 million in 1985.\9 A decade later the figure had
risen to $525 million; however, in testimony and a report issued in
1996, we pointed out the inaccuracy of this estimate due to the
omission of several cost factors, including known contract
increases.\10 The Department of Commerce later committed to a $550
million funding cap.  We testified last spring that it would be
extremely difficult for NOAA to develop and deploy AWIPS within the
$550 million cap if it encounters any major problems.\11 And given
the size and complexity of the development--and recognizing that even
managed risks can turn into real problems--such problems are likely
to occur.\12

The department has since reported its uncertainty over whether AWIPS
could be completed within the $550 million cap.  NWS officials
attribute this to increased software development expenses.  In
accordance with a recommendation we made in 1996,\13

Commerce contracted for an independent cost estimate.  According to
this February 2, 1998, assessment, the likely costs to complete
builds 4 through 6 will jump by $68 million, to a total of $618
million.  The contractor also noted a likely 9-month schedule delay. 

In addition, while AWIPS was planned for full deployment through
build 6 in 1999--at 152 locations nationwide--that schedule is now in
doubt.  The latest schedule calls for build 4--actually build 4.2--to
be completed in March of 1999.  Completion dates for builds 5 and 6
are now uncertain because, we are told, NWS wants to ensure that
requirements for those modules are not extraneous to mission needs,
in order to minimize future cost increases.  This reflects a
recommendation we made in 1996 for all AWIPS builds.\14 Until this
process has been completed and specific mission requirements
identified, the Weather Service will not know what capabilities
builds 5 and 6 will possess, and when they will be available. 

The most critical risk factors underlying questions about AWIPS'
future relate to software development.  Software quality is governed
largely by the quality of the processes used to develop it; however,
NWS' efforts to develop AWIPS software have lacked defined
software-development processes.  Such processes are all the more
essential because of NWS' increased use of software code developed
internally at NOAA's Forecast Systems Laboratory (FSL) in Boulder,
Colorado--a research and development facility that primarily develops
prototype systems.\15 This software code has not been developed
according to the rigorous processes commonly used to develop
production-quality code.\16 Failure to adhere to these processes may
result in unstable software that will continue to cause cost
increases and schedule delays.  The cost assessment delivered last
month also found risk inherent in the development of builds 4 through
6 because of the transitioning of FSL-developed software to AWIPS,
and the uncertainty surrounding requirements for these builds. 

Another risk area concerns the network control facility, which
provides the ability to monitor and maintain AWIPS sites across the
country from a single location.  Through build 3 AWIPS was still
experiencing difficulty with the central location's ability to detect
and respond to problems.  We should note, however, that such problems
concerned only a limited number of sites, and that as more sites come
on line, problems can be expected to increase.  Weather Service
officials acknowledge that the poor performance of the network
control facility to date is a prime concern, and that neither NWS nor
its contractor has experience in developing the capability for
central maintenance.  As a result, they told us that they obtained a
contractor to assist them in this area. 

The last, huge area of risk, inescapably, is whether the AWIPS
builds--and, indeed, all modernization components--will be what is
called Year 2000 compliant.\17 On the basis of information received
from Weather Service officials just last week, our concerns have not
been allayed.  AWIPS to date is not Year 2000 compliant.  Build
4.2--set for completion a year from now--is supposed to make all
AWIPS applications Year 2000 compliant.  But several questions
remain, perhaps the most obvious:  What if it does not?  As a
fail-safe tactic, NWS told us that it has renovated its aging current
system, Automation of Field Operations and Services, in case AWIPS is
not operational by the year 2000. 

Yet even if Year 2000 compliance ceases to be an issue with build
4.2, NWS' companion modernization systems--GOES-Next, and
NEXRAD--will need to be compliant as well because of the amount of
data interchanged among them.  Similarly, it will be essential to
determine whether all systems that provide data to individual weather
stations are themselves Year 2000 compliant; otherwise, "corrupted"
data emanating from an unrenovated system could infect the entire
operation.  According to NWS officials, they are continuing to
evaluate their data exchanges.  They also acknowledged the risks to
weather services if their systems are not Year 2000 compliant.  To
reduce the risk and the potential impact of Year 2000-induced
information system failures on the Weather Service's core business
processes, it is critical that NWS have contingency plans in place to
help ensure continuity of operations through the turn of the century. 


--------------------
\5 See National Oceanic and Atmospheric Administration:  Weather
Service Modernization and NOAA Corps Issues (GAO/T-AIMD/GGD-97-63,
Mar.  13, 1997). 

\6 See Weather Service Modernization:  Risks Remain That Full Systems
Potential Will Not Be Achieved (GAO/T-AIMD-97-85, April 24, 1997),
Weather Forecasting:  Recommendations to Address New Weather
Processing System Development Risks (GAO/AIMD-96-74, May 13, 1996),
Weather Forecasting:  NWS Has Not Demonstrated That New Processing
System Will Improve Mission Effectiveness (GAO/AIMD-96-29, Feb.  29,
1996), Weather Forecasting:  Improvements Needed in Laboratory
Software Development Processes (GAO/AIMD-95-24, Dec.  14, 1994), and
Weather Forecasting:  Systems Architecture Needed for National
Weather Service Modernization (GAO/AIMD-94-28, Mar.  11, 1994). 

\7 High-Risk Series:  An Overview (GAO/HR-95-1, February 1995) and
High-Risk Series:  Information Management and Technology
(GAO/HR-97-9, February 1997). 

\8 Incremental systems development entails building and testing
software products in a series of increments with increasing
functional capability; that is, the software is partitioned into
increments whose development is phased in over the total development
cycle. 

\9 Weather Forecasting:  Cost Growth and Delays in Billion-Dollar
Weather Service Modernization (GAO/IMTEC-92-12FS, Dec.  17, 1991). 

\10 Weather Forecasting:  New Processing System Faces Uncertainties
and Risks (GAO/T-AIMD-96-47, Feb.  29, 1996) and GAO/AIMD-96-74, May
13, 1996. 

\11 GAO/T-AIMD-97-85, April 24, 1997. 

\12 National Oceanic and Atmospheric Administration:  Follow-up on
Weather Service Modernization and NOAA Corps Issues
(GAO/AIMD/GGD-97-75R, April 10, 1997). 

\13 GAO/T-AIMD-96-47, Feb.  29, 1996 and GAO/AIMD-96-74, May 13,
1996. 

\14 See GAO/AIMD-96-29, Feb.  29, 1996. 

\15 FSL's prototype system, called Weather Forecast Office
(WFO)-Advanced, was being developed in parallel with AWIPS as a
risk-reduction tactic.  In both 1995 and 1996, AWIPS program
officials decided to use the WFO-Advanced software to take advantage
of FSL's hydrology and meteorology application software, hoping that
this would enable NWS to deploy AWIPS as quickly as possible. 

\16 See GAO/T-AIMD/GGD-97-63, Mar.  13, 1997, and Weather
Forecasting:  Improvements Needed in Laboratory Software Development
Processes (GAO/AIMD-95-24, Dec.  14, 1994) for discussion of software
quality assurance and software configuration management principles
typical of a production environment. 

\17 Computer systems have long used two digits to represent the year,
such as simply "98" for 1998, to conserve electronic data storage and
reduce operating costs.  In this format, however, 2000 is
indistinguishable from 1900 because both are represented as "00." As
a result, if not modified, systems or applications that use dates or
perform date- or time-sensitive calculations may generate incorrect
results beyond 1999. 


   ERIE/NORTHWESTERN PENNSYLVANIA
   RADAR COVERAGE
---------------------------------------------------------- Chapter 0:3

Finally, at the request of the Chairman of the full Committee and
Congressman English, we reported last September on the effects of NWS
modernization activities on radar coverage in the area of Erie and
northwestern Pennsylvania.\18 As part of its streamlining efforts,
the Weather Service has identified various weather offices nationwide
for closure, with responsibilities to be taken over by neighboring
offices.  Many concerns were voiced, however, that closing the Erie
office would result in a gap in radar coverage of lake-effect snow--a
gap that neighboring NWS offices (in Cleveland, Buffalo, Pittsburgh,
and State College, Pennsylvania) would be unable to fill. 

As we reported at the time, several studies presented evidence that a
degradation in weather service had not occurred in northwestern
Pennsylvania; however, the ability to detect and predict lake-effect
snow remained a concern.  For example, the preliminary conclusions of
an NWS lake-effect snow study indicated that weather services
provided to Erie were not as good as services provided to other lake
communities whose overall service had improved as a result of the NWS
modernization.  Because of this, the Director of NWS' Office of
Meteorology recommended that a new radar be installed for the Erie
area. 

In its response at the time, Commerce officials said they were still
in the process of completing the study of lake-effect snow and
reviewing the need for additional radar at Erie.  As of last week,
however, no decision on this matter had been made. 


--------------------
\18 GAO/AIMD-97-156, Sept.  26, 1997. 


-------------------------------------------------------- Chapter 0:3.1

Mr.  Chairman, this concludes my statement.  We would be pleased to
respond to any questions that you or other members of the
Subcommittee may have at this time. 


*** End of document. ***