Defense Spectrum Management: More Analysis Needed to Support	 
Spectrum Use Decisions for the 1755-1850 MHz Band (20-AUG-01,	 
GAO-01-795).							 
								 
Plans for identifying spectrum to support third generation mobile
wireless systems by July 30, 2001, and to auction licenses by	 
September 30, 2002, are premature. GAO agrees with the Federal	 
Communications Commission (FCC) and the Department of Commerce	 
that delaying the identification of spectrum and auction of	 
licenses for third generation wireless systems could serve the	 
public interest. Adequate information is not currently available 
to fully identify and address the uncertainties and risks of	 
reallocation. The Department of Defense (DOD) and the federal	 
government could make decisions affecting national security	 
without knowing the full extent of risks they face or steps	 
available to reduce those risks. Extending the current schedule  
for identification and auction of licenses for this portion of	 
the spectrum would allow DOD to complete technical and		 
operational assessments and to consider their future spectrum	 
requirements. In addition, a delay would allow time to further	 
consider the adequacy of existing national spectrum strategies	 
affecting international agreements and DOD overseas military	 
operations, to modify these strategies as necessary, and to	 
incorporate these strategies into a DOD long range spectrum plan.
-------------------------Indexing Terms------------------------- 
REPORTNUM:   GAO-01-795 					        
    ACCNO:   A01430						        
  TITLE:     Defense Spectrum Management: More Analysis Needed to     
             Support Spectrum Use Decisions for the 1755-1850 MHz Band        
     DATE:   08/20/2001 
  SUBJECT:   Cellular telephone 				 
	     International agreements				 
	     Radio frequency allocation 			 
	     Telecommunication					 
	     Army Mobile Subscriber Equipment Program		 

******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO Testimony.                                               **
**                                                              **
** 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.                                         **
**                                                              **
******************************************************************
GAO-01-795
     
Report to the Ranking Minority Member, Subcommittee on Readiness and
Management Support, Committee on Armed Services, U. S. Senate

United States General Accounting Office

GAO

August 2001 DEFENSE SPECTRUM MANAGEMENT

More Analysis Needed to Support Spectrum Use Decisions for the 1755- 1850
MHz Band

GAO- 01- 795

Page i GAO- 01- 795 Defense Spectrum Management Letter 1

Results in Brief 3 Background 5 Additional Analysis Is Required for Spectrum
Decisions 10 Conclusions 22 Recommendations for Executive Action 23 Agency
Comments and Our Evaluation 24 Scope and Methodology 24

Appendix I Spectrum Interference From Third Generation Systems to DOD
Satellites 26

Appendix II Replication of DOD?s and Industry?s IMT- 2000 Base Station
Interference Models 36

Appendix III Comments From the Department of Defense 49

Appendix IV Comments From the Department of Commerce 55

Appendix V Comments From the Federal Communications Commission 56

Tables

Table 1: Results of DOD and Industry Interference Calculations 39 Table 2:
Comparison of DOD and Our Interference Results based

upon Our Replication of DOD?s Model 47 Table 3: Comparison of Industry and
Our Interference Results

based upon Our Replication of Industry?s Model 47 Contents

Page ii GAO- 01- 795 Defense Spectrum Management Abbreviations

BBA Balanced Budget Act CTIA Cellular Telecommunications and Internet
Association DMSP Defense Meteorological Satellite Program DOD Department of
Defense DSB Defense Science Board DSCS Defense Satellite Communications
System DSP Defense Support Program FCC Federal Communications Commission
FFRDC Federally Funded Research and Development Center GPS Global
Positioning System IMT- 2000 International Mobile Telecommunications- 2000
ITFS Instructional Television Fixed Service ITU International
Telecommunication Union JTCTS Joint Tactical Air Combat Training System MDS
Multipoint Distribution Service MMDS Multichannel Multipoint Distribution
Service MSE Mobile Subscriber Equipment NATO North Atlantic Treaty
Organization NDAA National Defense Authorization Act NTIA National
Telecommunication and Information Administration OBRA Omnibus Budget
Reconciliation Act PSD Power Spectral Density WARC World Administrative
Radio Conference WRC World Radiocommunication Conference

Page 1 GAO- 01- 795 Defense Spectrum Management

August 20, 2001 The Honorable James Inhofe Ranking Minority Member
Subcommittee on Readiness and

Management Support Committee on Armed Services United States Senate

Dear Senator Inhofe: The use of commercial mobile wireless communications,
such as personal communications services and cell phones, has been
escalating worldwide over the radio frequency spectrum. 1 This spectrum is a
scarce and finite resource, which serves ever increasing and competing
public and private uses. The federal government has supported commercial
services by making spectrum available as these services developed over the
years. Some representatives of the commercial mobile radio service industry
claim additional spectrum is needed to support advanced communications
systems, called third generation systems. Other members of the industry
question the need for, or feasibility of providing, large amounts of
additional spectrum to meet industry requirements. These members point to
the need for increased efficiency in the use of spectrum through new
technologies. Third generation systems are being developed to provide mobile
voice, high- speed data, and Internet capabilities and are expected to
contribute significantly to the economic well- being of the United States.

Access to the radio frequency spectrum is also critically important to
federal, state, and local governments for national security, public safety,
and other governmental functions. Specifically, while the national interest
in a broad sense may be served by a robust commercial mobile wireless
system, national security also requires that the federal government be able
to meet its unique communications requirements to support domestic training
and worldwide military operations. Thus, determining the proper

1 Electromagnetic radiation is the propagation of energy that travels
through space in the form of waves. The radio frequency spectrum is the
portion of electromagnetic spectrum that carries radio waves. Frequency is
the number of waves traveling by a given point per unit of time, in cycles
per second, or hertz. Radio frequency is usually measured in thousands of
hertz or kilohertz (kHz), millions of hertz or megahertz (MHz), and billions
of hertz or gigahertz (GHz).

United States General Accounting Office Washington, DC 20548

Page 2 GAO- 01- 795 Defense Spectrum Management

use of a limited amount of spectrum, today and in the future, is a
challenging and complex task due to competing industry and governmental
demands.

In response to a presidential memorandum released in October of 2000, the
Department of Commerce issued a plan to select spectrum for potential use by
third generation wireless systems in the United States. This plan designated
the National Telecommunication and Information Administration (NTIA) to
study the 1755 to 1850 MHz band. Within the United States, this band is
allocated exclusively to the federal government, particularly for defense
purposes, such as space systems, mobile tactical communications, and combat
training.

To assist the NTIA, the Department of Defense (DOD) issued a report in
February 2001, which focuses primarily on spectrum sharing and relocation
issues in the United States. The report also describes potential operational
impacts and estimates the costs DOD would incur from future reallocation of
this spectrum to private sector users. DOD concluded that vacating the band
could not be accomplished for space systems until at least 2017, and for
most DOD non- space systems until at least 2010. The report notes that DOD
could face significant operational restrictions in any frequency- sharing
situation. In addition, the preliminary results of DOD?s analysis indicate
that comparable spectrum, which is operationally suitable, may not be
available for the relocation of the DOD systems. NTIA?s final report in
March 2001 incorporated DOD?s study results and concluded that unrestricted
sharing of the 1755 to 1850 MHz band is not feasible and any other sharing
option would require considerable coordination. Further, the report notes
that issues involving the availability of comparable spectrum,
reimbursement, and the time required for federal entities to either modify
or replace equipment would need to be addressed before a decision could be
made whether federal users could share or vacate a band of spectrum.

Because of the importance of this issue to private and public interests in
the United States, you asked us to assess DOD?s study of the 1755 to 1850
MHz band and determine whether the study provides a sound basis for
decisions about the reallocation of this spectrum. As part of this
assessment, we reviewed and analyzed an industry consortium report that

Page 3 GAO- 01- 795 Defense Spectrum Management

reached different conclusions on space operations than the DOD study. 2 We
also reviewed a separate ongoing Air Force analysis that has similar
conclusions on space operations to the DOD study. 3 We did not review the
Department of Commerce or the Federal Communications Commission (FCC) report
or each agency?s process to identify and select additional spectrum for
third generation wireless systems.

Spectrum decisions based on either the DOD or the industry study of the 1755
to 1850 MHz band would be premature at this time. Neither study contains
adequate information to make reallocation decisions. 4 In particular, we
found that neither the DOD model nor the industry model is mature enough to
calculate spectrum interference to satellites, and, therefore, cannot
support a near- term decision. DOD recognized in its report that additional
analyses are needed to assess the impacts of any reallocation or sharing
decision. A third model to calculate this interference, developed by the Air
Force, shows promise but is still under development. 5 All three studies
used different assumptions because there is no single agreed upon
methodology or model today to estimate potential spectrum interference from
third generation wireless systems.

2 Report of the Industry Association Group on Identification of Spectrum for
3G Services, an attachment to comments to the Federal Communications
Commission?s notice of proposed rulemaking on 3G submitted by the Industry
Association Group. The Industry Association Group is composed of the
Cellular Telecommunications and Internet Association, the Telecommunications
Industry Association, and the Personal Communications Industry Association.

3 The Air Force study is being prepared by the Aerospace Corporation, a
federally funded research and development center (FFRDC). These centers are
sponsored by government agencies, but are privately administered by
universities and other nonprofit corporations to ensure objectivity and
independence. Aerospace Corporation?s primary customer is the Space and
Missile Systems Center of the Air Force Materiel Command. The primary
responsibility of Aerospace Corporation is to ensure the mission success of
national security space programs.

4 Potential spectrum interference with DOD satellite operations is discussed
later in the report, and appendixes I and II are devoted to this issue. This
focus is the result of a lack of time for us to conduct similar analyses on
other DOD systems and does not mean that problems are not present in other
major tactical radio, air combat training, and precision guided munitions
systems.

5 Aerospace Corporation reviewed the DOD and industry studies as a starting
point and is making changes in methodology as it deems appropriate. The
Corporation will incorporate these changes into future work. Results in
Brief

Page 4 GAO- 01- 795 Defense Spectrum Management

If the current schedule to auction spectrum is maintained, the federal
government will make decisions affecting national security without knowing
the full extent of the risks it faces or steps available to reduce those
risks. While the DOD study provides preliminary information about DOD
systems and operational requirements, it does not contain information on
critical considerations, which are necessary for decisions on DOD?s ability
to vacate the entire band or share all or part of the band with industry.
According to DOD officials, the following considerations (1) were outside
the scope of its study or (2) were not included in its study because of a
compressed time schedule and the lack of information on alternative spectrum
and development of third generation systems.

 First, additional technical and operational analysis is required to
determine the impact of third generation systems on military operations.
Without the proper technical and operational analyses, DOD risks a reduction
in military preparedness or a degradation of systems in the 1755 to 1850 MHz
band that support mission capabilities. Specifically, DOD faces an unknown
risk of operational degradation to its satellite operations that could
include actual loss of control of its satellites and an undetermined risk to
the warfighter.

 Second, because additional analysis is required to estimate costs, DOD
risks not receiving a fair reimbursement value for the costs the agency
incurs to vacate or share the 1755 to 1850 MHz band. The DOD cost estimate
is potentially understated by billions of dollars because the Department did
not include satellite replacement costs.

 Third, because the DOD report did not fully consider the Department?s
future communication requirements, the Department risks losing spectrum that
may be needed for national security. DOD?s future warfare plans assume
unimpeded communications are readily available to deployed forces.

 Fourth, DOD?s study process lacked adequate programmatic, budgetary,
technical, and scheduling guidance for command and operational units.
Without this guidance, command and operational units were unsure which
program or technical alternatives to include in the study.

 Finally, DOD did not include in its study the potential impacts of U. S.
spectrum reallocation decisions on international agreements and overseas
military operations. Reallocation of U. S. spectrum could have a negative
impact, such as loss of existing regulatory protection, on these agreements
and operations.

Page 5 GAO- 01- 795 Defense Spectrum Management

This report contains recommendations to the Secretary of Defense and the
Secretary of Commerce concerning actions needed to complete the analysis of
the impacts from a potential spectrum reallocation of the 1755 to 1850 MHz
band. In commenting on a draft of this report, DOD, FCC, and the Department
of Commerce agreed with our findings and recommendations. The National
Security Council provided only technical comments.

Wireless technologies have become a valuable asset to improve communications
efficiency and reduce costs for industries and governments around the world.
The first generation of wireless technology, analog cell phones, is still
used in many areas of the country. The second generation of wireless
technology added digital ?personal

communications services,? such as voice mail, text messaging, and access to
the World Wide Web. The second generation is growing at an extraordinary
rate and is an essential part of the way the world does business. Today,
industry is also developing a third generation of personal communications,
which are expected to give consumers mobile, high data rate, high- quality
multi- media services.

The most technically suitable spectrum for mobile communications, public and
private, is below three gigahertz. This band of spectrum is the best match
for special spectrum propagation characteristics (such as distance,
capacity, and reliability) required by DOD, other federal agencies, and
commercial wireless firms. As a result, this spectrum is the subject of much
competition among the different users. This competition presents major
spectrum management issues for decisions by governmental organizations
within the United States and by international organizations composed of
sovereign nations.

FCC and NTIA manage the radio spectrum in the United States. NTIA is the
executive branch agency principally responsible for developing and
articulating domestic and international telecommunications policy for the
executive branch. NTIA is also responsible for managing the federal
government?s use of the radio spectrum. FCC, an independent agency of the
federal government, has authority over commercial spectrum use, as well as
the use of spectrum by state and local governments. NTIA and FCC Background

Page 6 GAO- 01- 795 Defense Spectrum Management

manage the spectrum through a system of frequency allocations, allotments,
and assignments. 6

Previously, Congress directed the reallocation of spectrum from federal to
private sector use under title VI of the Omnibus Budget Reconciliation Act
of 1993 (OBRA 93) 7 and later expanded the reallocation of spectrum under
title III of the Balanced Budget Act of 1997 (BBA 97). 8 Under these laws,
NTIA identified a total of 255 MHz for reallocation. 9

Because radio waves transcend national borders and the number of global
services has been increasing, international coordination of spectrum is a
critical component of the spectrum allocation process. The
radiocommunication conferences of the International Telecommunication Union
(ITU) 10 are the principal mechanisms for international spectrum allocation
via treaties. At the 2000 World Radiocommunicaton Conference (WRC- 2000),
the ITU discussed spectrum and regulatory issues for advanced mobile
applications, including third generation services. According to NTIA
officials, the WRC- 2000 and the ITU 1992 World Administrative Radio
Conference (WARC- 92) identified a total of 749

6 The entire radio spectrum is divided into blocks, or bands, that are
allocated for broad categories of radio services, such as fixed, mobile,
broadcasting, or satellite services. FCC establishes rules that further
define the particular types of use that are permitted within each
allocation. For example, a frequency band that is allocated to the mobile
service may be designated in the FCC rules for particular users such as
business users, public safety users, or cellular users. Allotments may be
made within certain services, such as TV broadcasting, where particular
channels are provided in each geographic area. Assignment refers to the
final subdivision of the spectrum in which a party gets an assignment or
license to operate a radio transmitter on a specific channel or group of
channels at a particular location under specific conditions.

7 P. L. 103- 66, Aug. 10, 1993. 8 P. L. 105- 33, Aug. 5, 1997. 9 Eight MHz
of spectrum was subsequently reclaimed per congressional direction. See
section 1062 of the National Defense Authorization Act for Fiscal Year 2000
(NDAA- 2000) (P. L. 106- 65, Oct. 5, 1999). OBRA 93 required FCC to
gradually allocate and assign frequencies over the course of 10 years. The
reallocation of the majority of the 235 MHz identified under that act is
still underway. BBA 97 imposed a stricter deadline for NTIA to identify for
reallocation and FCC to reallocate, auction, and assign licenses by
September 2002 for the additional 20 MHz of federal spectrum.

10 The International Telecommunication Union is a United Nations specialized
agency. The federal government considers the ITU the principal competent and
appropriate international organization for the purpose of formulating
international treaties and understandings regarding certain
telecommunications matters.

Page 7 GAO- 01- 795 Defense Spectrum Management

MHz 11 of spectrum for use by sovereign countries wishing to implement IMT-
2000 (also known as third generation mobile wireless), including the 1755 to
1850 MHz band. 12 The United States agreed that it would study these bands
domestically. Because of the significance of the services in this band of
spectrum, the United States did not commit to providing additional spectrum
for third generation systems.

Subsequent to WRC- 2000, the President issued an executive memorandum on
October 13, 2000, that outlined a policy to encourage cooperation among FCC,
NTIA, other federal agencies, and the private sector to determine if
additional spectrum could be made available for third generation wireless
systems. In addition, the memorandum specified that incumbent users of
spectrum be treated equitably, taking national security and public safety
into account. The memorandum directed the Secretary of Commerce to work
cooperatively with FCC to develop a plan to select spectrum for third
generation wireless systems and to issue an interim report on the current
spectrum uses and the potential for reallocation or sharing of the bands
identified at WRC- 2000. The plan established spectrum sharing and
relocation options to be studied for implementation in the years 2003, 2006,
and 2010. FCC, in conjunction with NTIA, was expected to identify spectrum
by July 2001, and auction licenses to competing applicants by September 30,
2002.

Under the plan, NTIA studied the 1755 to 1850 MHz band and the FCC studied
the 2500 to 2690 MHz band. 13 Within the United States, the 1755 to 1850 MHz
band is allocated on an exclusive basis to the federal government for fixed
and mobile services and satellite control. DOD is the

11 According to NTIA, WARC- 92 identified 1885 to 2025 and 2110 to 2200 MHz
and WRC2000 identified 806 to 960 MHz, 1710 to 1885 MHz, and 2500 to 2690
MHz. 12 According to NTIA officials, WRC- 2000 indicated that identification
of these bands does not preclude the use of these bands by any services to
which they are allocated and does not establish priority for third
generation systems in the radio regulations. A DOD official said
?identification? of spectrum for third generation mobile wireless systems
does not create a treaty obligation on each party to a treaty to use
particular spectrum for third generation systems.

13 The 1755 to 1850 MHz band is a federal government band used primarily for
fixed point to point microwave, air and ground mobile uses, and space
operations. In the United States, the 2500 to 2690 MHz band is currently
used by the Instructional Television Fixed Service (ITFS), Multipoint
Distribution Service (MDS), and Multichannel Multipoint Distribution Service
(MMDS).

Page 8 GAO- 01- 795 Defense Spectrum Management

predominant user, although 13 other federal agencies operate extensive fixed
and mobile systems in this band throughout the United States.

To support NTIA?s efforts to study the 1755 to 1850 MHz band, DOD issued a
final report in February 2001. The report was prepared for DOD by the Office
of Spectrum Analysis and Management, which is part of the Defense
Information Systems Agency. In addition, responsibility for assisting in the
development of cost estimates associated with implementing study results was
assigned to the Department?s Cost Analysis Improvement Group. 14 The study
addressed whether the Department could share or vacate this band.

The Office of Spectrum Analysis and Management grouped military systems into
five major categories. These categories included satellite operations,
tactical radio relay, air combat training, weapons data links, and
miscellaneous systems. Next, technical analyses were conducted to calculate
the effects of potential spectrum interference between proposed commercial
(third generation) and major military systems. 15 For example, DOD
calculated the potential interference between selected DOD satellite ground
stations and anticipated third generation wireless systems (mobile and
fixed- base stations). DOD also calculated the potential interference from
third generation wireless systems on DOD satellite operations. In another
example of technical analysis, DOD selected two air combat training ranges,
which were considered representative of all training areas in the United
States, for analysis of potential spectrum interference. Finally, the
military commands and the operational communities responsible for each major
category of systems used the results of the technical analysis to estimate
the extent of any operational impacts on their missions. Then, acquisition
program officers provided cost estimates for the options of sharing or
vacating the band.

DOD?s spectrum report concluded that loss of access to spectrum, above and
beyond the spectrum already transferred as a result of OBRA 93 and BBA 97,
would jeopardize DOD?s ability to execute its mission. Specifically, DOD
concluded that it is unable to totally vacate the 1755 to 1850 MHz band
until at least 2017 for space systems and at least 2010 for

14 The Cost Analysis Improvement Group provides overall DOD guidance for
accuracy of DOD cost analyses used in program acquisition decisions. 15 DOD
officials said their analysis incorporated expected performance
characteristics of third generation mobile wireless systems derived from
published sources and coordinated with NTIA and FCC.

Page 9 GAO- 01- 795 Defense Spectrum Management

non- space systems. DOD also found that full band sharing is not feasible.
However, DOD said that the compressed schedule, initiated by the
presidential memorandum and the Department of Commerce?s plan, did not
provide time for a thorough analysis and review of these complex issues.

NTIA issued a spectrum report on March 30, 2001. 16 Based in part on the DOD
report, NTIA found that the unrestricted sharing of the 1755 to 1850 MHz
band is not feasible and that any other sharing option would require
considerable coordination between industry and DOD before third generation
systems can be operated along with federal systems. Specifically, NTIA
stated that there are several issues that must be resolved before any
spectrum can be made available for reallocation in the 1755 to 1850 MHz
band. These include reimbursement issues 17 and the assurance of
availability of comparable spectrum if DOD must surrender spectrum in this
band. 18

On June 26, 2001, the Chairman of FCC wrote to the Secretary of Commerce
stating that additional time is necessary to allow the Commission and the
executive branch to complete evaluations of the various options available
for advanced wireless services. FCC sought additional time to identify and
schedule the auction of spectrum for third generation wireless services. FCC
stated in its letter that it wanted to work with the executive branch and
appropriate congressional committees to come up with a revised allocation
plan and auction timetable for third

16 NTIA?s final report examined the potential for accommodating third
generation mobile wireless systems in the broader 1710 to 1850 MHz band. 17
The Strom Thurmond National Defense Authorization Act for the Fiscal Year
1999 (NDAA- 99) (P. L. 105- 251, Oct. 17, 1998) authorized federal entities
to accept compensation payments when they relocate or modify their frequency
use to accommodate non- federal users of the spectrum. These reimbursement
provisions would be applicable to any spectrum that is or would be
reallocated in the 1755 to 1850 MHz band. Both NTIA and FCC have issued
notices of proposed rulemaking that address reimbursement provisions.

18 The National Defense Authorization Act for Fiscal Year 2000 (NDAA- 2000)
(P. L. 106- 65, Oct. 5, 1999) specified a number of conditions that have to
be met if spectrum in which DOD is the primary user is surrendered. This act
requires that NTIA, in consultation with FCC, identify and make available to
DOD for its primary use, if necessary, an alternate band( s) of frequency as
a replacement for the band surrendered. Further, if such bands of frequency
are to be surrendered, the Secretaries of Defense and Commerce and the
Chairman of the Joint Chiefs of Staff must jointly certify to relevant
congressional committees that such alternative band( s) provide comparable
technical characteristics to restore essential military capability.

Page 10 GAO- 01- 795 Defense Spectrum Management

generation systems that will allow completion of necessary work. In a July
19, 2001, letter responding to the FCC request, the Secretary of Commerce
directed NTIA to work with FCC to develop a new plan for the selection of
spectrum for third generation mobile wireless systems. The Secretary asked
that this work be coordinated with appropriate executive branch entities,
such as the National Security Council, the National Economic Council, the
Office of Management and Budget, and DOD. He also encouraged participants in
this process to consider ways to achieve flexibility with respect to the
statutory auction dates if flexibility is needed to implement the new plan.
For example, the 1710 to 1755 MHz band that is being considered for third
generation purposes is under a statutory deadline for the auction of
licenses by September 2002. 19

DOD?s February 2001 study does not provide a basis for decisions about
reallocation of spectrum in the 1755 to 1850 MHz band. The study was
constrained by lack of adequate guidance and by inadequate time and
information. Thus, major considerations either were not addressed or were
not adequately addressed in the final report. These considerations include
complete technical and operational analyses of anticipated spectrum
interference; cost estimates supporting DOD reimbursement claims, spectrum
requirements supporting future military operations; programmatic, budgeting,
and schedule decisions needed to guide analyses of alternatives; and
potential impacts of U. S. reallocation decisions upon international
agreements and operations. As a result, DOD?s analysis was limited in its
ability to adequately describe and document potential technical,
operational, and cost impacts should the Department be required to vacate
the 1755 to 1850 MHz band or to share it with commercial users.

In addition, neither the industry analysis nor the Air Force analysis of
potential spectrum interference to DOD satellite systems provides an
adequate basis for decisions about reallocation of the 1755 to 1850 MHz band
from governmental to nongovernmental uses. The findings of both analyses,
along with DOD?s analysis of satellite interference, are not reliable or
complete.

19 Section 3007 of P. L. 105- 33 provides that the Commission shall conduct
the competitive bidding in such a manner as to ensure that all proceeds are
deposited not later than September 30, 2002. Additional Analysis Is

Required for Spectrum Decisions

Page 11 GAO- 01- 795 Defense Spectrum Management

Before making reallocation decisions with a significant impact on national
security and the economic welfare of the nation, the federal government
should approach the alternatives with knowledge gained from a sound and
complete analysis. Given an adequate amount of time, information, and
guidance, a study of DOD?s spectrum needs and requirements could reduce
operational and cost risks presented by critical spectrum reallocation
decisions. The alternatives considered for making radio frequencies
available to industry include DOD vacating the entire spectrum band or
sharing all or part of the spectrum band with industry.

DOD based its analysis of potential operational and cost impacts on the
premise that it could not accept any degradation of current mission
capability from a complete or partial reallocation of its spectrum to other
users. DOD also stated in its report that unrestricted sharing of the entire
band with third generation mobile systems would place unacceptable
operational restrictions on both DOD and commercial users of the band. These
factors caused DOD to establish critical conditions that it believes must be
met before it vacates or shares the 1755 to 1850 MHz band. These conditions
include (1) alternative spectrum must be provided comparable to what the
Department loses; (2) cost reimbursement must be timely as required under
current law; (3) the Department must receive the same regulatory protection
in any new spectrum as it now enjoys in the 1755 to 1850 MHz band; (4)
defense systems must receive timely certification to operate in any new
spectrum band; and (5) new commercial users in the 1755 to 1850 MHz band
must be prohibited from interfering with DOD?s legacy systems while they
migrate into their new spectrum. The Department concluded in its report that
these conditions could not be met in the short term and that it must have
continued access to the 1755 to 1850 MHz band until at least the year 2017
for satellite systems and until approximately 2010 for other systems.

In its June 26, 2001, letter to NTIA, the Chairman of the FCC stated that
the entire federal government faces a challenging set of issues in
addressing how best to make available sufficient U. S. spectrum for advanced
wireless services and that the public interest would be best served by
additional time for informed consideration of these issues. DOD?s February
2001 report also recognizes that additional analyses are needed to fully
assess operational impacts and develop estimates of costs resulting from any
spectrum reallocation or sharing. Importance of Analyzing

DOD?s Spectrum Needs and Requirements

Page 12 GAO- 01- 795 Defense Spectrum Management

Based on our experience and previous work on spectrum management issues, 20
we believe an analysis of spectrum reallocation affecting DOD would be more
comprehensive if it included the following considerations:

 Additional technical and operational analyses to more completely reveal
the impact of third generation wireless systems on military systems and any
potential operational degradation of DOD systems.

 Additional analysis to estimate the cost of vacating or sharing the
frequency band and the level of reimbursement.

 Identification of the expanding future communication requirements to allow
DOD to include those requirements necessary for the envisioned warfare
strategies, which rely heavily on wireless communications.

 Appropriate programmatic, budgeting, schedule, and technical guidance to
the services and units conducting the analyses of the individual systems to
define the scope and breadth of the analysis and prepare an accurate
assessment of operational and cost impacts.

 Consideration of the impact of reallocation decisions on international
agreements and operations. A national spectrum strategy could give DOD, FCC,
and NTIA a guiding framework for decisions affecting training operations
with allies, overseas deployments, and international treaty obligations.
Spectrum reallocation impacts all of these areas.

Due to time and information constraints on DOD?s initial technical and
operational analyses, further study is required in these areas. For example,
the DOD report predicts interference to satellite operations from third
generation wireless systems by the year 2006 and states that this
interference could impede command and control of DOD satellites, especially
low- earth orbit satellites. However, officials from the telecommunications
industry have a different view. An industry analysis states that
interference from third generation mobile wireless systems to DOD satellite
receivers will be at acceptable levels and that sharing between these
systems and DOD satellites is possible without any efforts to mitigate
interference levels from the commercial systems. Our review of

20 Defense Communications: Federal Frequency Spectrum Sale Could Impair
Military Operations (GAO/ NSIAD- 97- 131, June 17, 1997). Additional
Technical and

Operational Analyses Are Required for Spectrum Use Decisions

Page 13 GAO- 01- 795 Defense Spectrum Management

the DOD and industry reports, as described below, however, found that the
two parties used different assumptions to calculate the extent of potential
interference and the impact this interference will have on satellites. 21

A third view of the potential interference from the commercial systems on
DOD satellites is being developed by the Aerospace Corporation, a federally
funded research and development center, for the Air Force. Aerospace
officials said they are using many of the same assumptions industry used in
its analysis but that they are also using techniques and assumptions not
included in either the DOD or industry analysis. These officials told us
that their results to date confirm the DOD position that third generation
mobile wireless in the 1755 to 1850 MHz band could interfere with DOD
operation of its satellites. However, Aerospace officials said they disagree
with DOD on which satellites will be affected. According to these officials,
the satellites affected by spectrum interference from third generation
mobile wireless systems will be medium- earth orbit (20,000 kilometers) and
high- earth orbit (36,000 kilometers) satellites, not the satellites in low-
earth orbit as forecasted by DOD. When we spoke to Aerospace officials, they
were starting to assess the effects of spectrum interference on the
operations of specific types of satellites. These operational analyses are
important for understanding the full effect of potential spectrum
interference on satellite performance. For example, our review of Aerospace
data suggests that estimated interference levels from third generation
mobile wireless systems are high enough to adversely affect successful
contact with the Global Positioning System (GPS). 22

Each analysis led to a different conclusion because, while certain general
engineering principles apply to estimating spectrum interference, no single
methodology or model exists today to estimate potential spectrum
interference to DOD satellite operations from third generation mobile
wireless systems. As a result, each party used different methodologies and
assumptions. In addition, our preliminary analysis of the DOD and

21 See also appendix I. 22 GPS provides worldwide navigation and timing data
to both military and civilian users. The civilian market is estimated at
approximately $10 billion dollars annually.

Page 14 GAO- 01- 795 Defense Spectrum Management

industry analyses indicates that questionable assumptions, 23 inadequate
information, and a compressed schedule negatively impacted their analyses.
One questionable assumption that was used in DOD and industry analyses was
the assumption that cities would generate most of the spectrum interference
to satellites and that contribution from suburban areas would be marginal.
DOD officials said that including interference from suburban and rural areas
or along interstate highways increases the projected amount of interference
from third generation systems on the satellites. They said this interference
could be severe enough to disrupt the command links to many satellites.
However, DOD officials did not have time to revise their published analysis
to incorporate this new information. Aerospace officials said they included
an estimate for suburban and rural interference that neither industry nor
DOD recognized in their reports.

With respect to inadequate information, all three analyses lacked essential
information from industry about its plans for building and deploying third
generation systems. The lack of good information about future industry plans
for geographic coverage and density of third generation mobile wireless
systems creates a very high level of uncertainty about the levels of energy
from these systems that can cause interference with satellite operations and
lessens the reliability of all three estimates. NTIA officials said this
information was requested from industry representatives during a series of
government- led industry outreach meetings between November 2000 and
February 2001. NTIA officials said, however, that industry representatives
refused to provide such information because it is proprietary and could not
be shared with competitors. In addition, as we describe in appendix I, DOD
may have significantly underestimated potential interference to control of
its satellites because it incorrectly estimated the size of cities in its
population database.

Another example of an incomplete technical and operational analysis is the
DOD assessment of third generation systems on airborne, precision guided
weapons training programs. These training programs use radio spectrum for
data links between the aircraft and the air launched weapons within both the
1710 to 1755 and 1755 to 1850 MHz bands. The 1710 to 1755 MHz band was
designated for reallocation from federal governmental to non- governmental
use pursuant to congressional direction, but federal

23 We analyzed the DOD and industry models to the point that we could
replicate their estimates of the level of spectrum interference generated by
the worldwide build- out of third generation systems. We reviewed the Air
Force's model but did not attempt to replicate its findings.

Page 15 GAO- 01- 795 Defense Spectrum Management

operations can continue in the band within 16 protected zones. 24 According
to NTIA officials, these sites were initially established by NTIA to protect
DOD ground and air training functions. 25 However, in its March 2001 report,
NTIA proposed, among other options, eliminating the 16 protection zones and
relocating all systems in the 1710 to 1755 MHz band to the 1755 to 1850 MHz
band or a higher band. According to NTIA, this proposal was necessary
because no other solution was available to share the 1710 to 1755 MHz band
with industry. The NTIA report noted that this proposal was not in the FCC
and Commerce plan to identify spectrum for potential reallocation and had
not been evaluated by DOD. Therefore, NTIA did not know how this proposal
would affect DOD operations. DOD had not completed a review of this proposal
at the time of our review.

The Air Force identified concerns to us about loss of spectrum in either
band. Air Force officials said the frequencies in the 1710 to 1755 MHz and
the 1755 to 1850 MHz bands operate as a pair, 26 and loss of either
frequency would adversely affect the training operations in the other
frequency. For example, Air Force officials told us the existing 16
protection zones in the 1710 to 1755 MHz band are already too small in land
area to simulate realistic combat conditions, but eliminating them entirely
would stop all operationally realistic training at these sites. In addition,
an Air Force official stated that many training missions now flown on low
level training routes over most of the continental United States would be
severely degraded if further spectrum is lost in either band.

However, Air Force officials said that no new spectrum has been selected for
training on precision guided weapons and that the total cost and operational
impact of changing frequency bands have not been fully assessed. These
officials stated that, at this time, any studies or analyses have been based
on numerous assumptions and, thus, study results are preliminary. The cost
of changing frequency bands could be substantial. Recognizing that these are
preliminary estimates, Air Force officials believe that loss of spectrum in
either band and moving to a higher band

24 The Federal Power Administrations and public safety fixed links will also
be protected. 25 FCC disagreed with NTIA. According to FCC technical
comments, Air Force operations were not identified to remain in this
spectrum. FCC said retaining such operations would have a detrimental impact
on any significant use of the spectrum for nongovernmental operations.

26 Frequencies in both bands are used for the same missions.

Page 16 GAO- 01- 795 Defense Spectrum Management

could cost up to $580 million in new equipment development and take up to 10
years to complete.

Air Force and Navy officials managing precision guided weapons programs also
cited lack of time and information as the main reasons for their inability
to perform a detailed analysis. For example, these officials stated that
they did not have adequate time to receive input from training bases and
obtain technical information from the commercial providers of the weapons
systems to determine the feasibility of band sharing or segmentation. DOD
also said that it could not determine the amount of new engineering work
required for the communications components of the weapons without knowledge
of a new operating band.

The problems of a compressed time schedule and a lack of information make it
difficult for federal agencies to reach a reasonable decision about
reallocation of the 1755 to 1850 MHz band to nongovernmental uses. For
example, a decision to exclude third generation mobile systems from this
band because of potential interference to satellite or other operations
could mean economic loss to industry. On the other hand, allowing these
systems into the 1755 to 1850 MHz band when in fact they could interfere
with satellite or training operations could mean a reduction in military
preparedness, degradation of satellite performance, or even loss of
satellites in orbit.

We found that DOD- within time constraints, extensive programmatic
uncertainty, and available guidance- produced reasonable cost estimates for
the assumptions used in the studies. The cost estimates in DOD?s report
range from at least $2.8 billion to relocate major defense communications
systems from segments of the band to in excess of $4.3 billion to fully
vacate the band. However, the Department?s cost estimates are incomplete
because of program, budget, and technical uncertainties and could be
underestimated by billions of dollars.

The DOD report acknowledges that its cost estimates are preliminary and
states that they are not conclusive. According to the report, all of the
cost estimates are sensitive to many complex technical and budgetary
unknowns. For example, the report notes that implementation of interference
mitigation measures can greatly enhance opportunities for spectrum sharing,
but employing any of these techniques would require a new cost assessment
that could dramatically alter the cost estimates in the report. In addition,
Air Force officials told us they did not determine the cost of replacing
entire satellite systems to make room for third generation Additional DOD
Work

Required to Estimate Costs

Page 17 GAO- 01- 795 Defense Spectrum Management

mobile wireless systems before the year 2017. The DOD report states that the
1755 to 1850 MHz band is used to control over 120 satellites in orbit and
that loss of this band before the year 2017 means it could no longer control
satellites in orbit and would have to replace them. According to Air Force
officials, they assumed continued access to the band for the life of
existing satellites. 27 They said satellite systems, including spacecraft
and related ground infrastructure, costing billions of dollars, would become
useless if DOD were forced to vacate the 1755 to 1850 MHz band before the
year 2017. While replacing these satellite systems would cost billions of
dollars, Air Force officials also questioned whether industrial base or
launch facilities exist to build and launch significant numbers of new
satellites before the year 2017. The total system costs of these satellite
systems suggest that the replacement costs would be significant. 28 For
example, DOD estimates total GPS program costs at $18.4 billion over a 43-
year period- fiscal years 1974 through 2016. 29 In a second example, the
Defense Meteorological Satellite Program (DMSP) has a total program cost
estimate through the year 2012 of $2. 4 billion. 30 Therefore, the cost
estimates could increase significantly if DOD is forced to vacate the 1755
to 1850 MHz band before 2017 and had to replace existing satellites before
the end of their normal life cycle. 31

Finally, DOD cost estimates on vacating the 1755 to 1850 MHz band cannot be
completed until the alternative spectrum for DOD is identified. To date,
NTIA has not been able to identify alternative and comparable spectrum
available for federal use to replace the 1755 to 1850 MHz band. DOD?s report
states that relocation costs could vary depending upon the bands selected as
replacements for lost spectrum in the 1755 to 1850 MHz band.

27 Air Force officials assumed that all satellites launched after the year
2010 would be capable of using the Unified S- band (2025 to 2110 MHz band).
However, as of June 2001 Air Force officials told us no decision had been
made to move to this band.

28 The Acting Assistant Secretary of Defense for Command, Control,
Communications, and Intelligence testified on July 31, 2001, before Congress
that the more than 120 satellites represent a cumulative investment of about
$100 billion.

29 The GPS system is totally DOD funded, about $9. 0 billion has been
invested to date and about $9. 4 billion is planned to be invested. 30 The
DMSP system is totally DOD funded and provides weather information primarily
for the military. 31 The National Security Council said that if relocation
costs were incurred because of international developments, the federal
government may not be entitled to full reimbursement.

Page 18 GAO- 01- 795 Defense Spectrum Management

A change in spectrum bands for weapons data links, for example, could
require either an extensive engineering redesign of antennas and other radio
equipment on both weapons and the aircraft delivering the weapons or
essentially a new major systems development program - depending on the new
spectrum band selected. In another example, changes to frequencies used by
existing satellites awaiting launch could delay the launch by years in order
to develop and manufacture key components for the new frequencies.

Without complete cost estimates for the reallocation of spectrum, the
Department cannot ensure that it is receiving a fair reimbursement value for
the costs the Department incurs to vacate or share the 1755 to 1850 MHz
band. Ultimately, if the Department is not fully reimbursed for the costs of
reallocation, the government would be responsible for the funds needed to
ensure that national defense is not degraded.

DOD?s report does not describe future spectrum requirements necessary to
meet the Department?s growing communications needs. The Defense Science
Board?s (DSB) November 2000 study on spectrum issues concluded that the
Department?s need for spectrum is escalating rapidly as

?information superior? forces become a reality and deploy. 32 DSB said
wireless communication is particularly critical for the type of
geographically dispersed warfare contemplated in future concepts of the
individual services, such as the Marine Expeditionary Forces. 33 The Board?s
study stated that the Department requires a proactive, needsbased strategy
supported by detailed knowledge of DOD?s spectrum requirements. DSB
recommended the Department expand an ongoing internal requirements study
into an inventory of current and future defense spectrum needs linked to
military capabilities.

The Department?s spectrum report acknowledged that it is highly likely that
new defense requirements for this band and other military spectrum bands
will arise. However, the DOD spectrum report does not discuss future
spectrum requirements in any depth and does not attempt to

32 Joint Vision 2010 establishes the DOD warfighting vision and defines
?information

superiority? as the key enabler of the operational concepts of Joint Vision
2010. 33 A Marine Expeditionary Force is a warfighting arm of the Marine
Corps that is composed of about 45,000 personnel from a Marine division,
various support activities, and a Marine aircraft wing. Future Military
Spectrum

Requirements Not Considered

Page 19 GAO- 01- 795 Defense Spectrum Management

quantify the requirements for the new systems. A Joint Chiefs of Staff
official said that an analysis of future requirements was outside the scope
of the report. After the report was issued, DOD provided us general forecast
information about fixed and mobile spectrum requirements. This information
projected an increase in mobile spectrum requirements, below 3 gigahertz, of
92 percent by the year 2005; fixed requirements increasing by 60 percent by
the year 2007; as well as more than 600 MHz of spectrum for training by the
year 2005. According to the information provided, any sharing arrangement
with third generation mobile wireless users in the 1755 to 1850 MHz band may
not be workable in the longterm, unless DOD freezes its spectrum
requirements in this band.

We recognize that DOD?s requirements are likely to change as new systems,
technologies, and strategies are developed for the nation?s future
warfighting force. However without a better understanding of future
requirements, DOD increases its risk of losing access to bands of spectrum
necessary for future mission needs.

We also found that key programmatic, budgeting, and schedule decisions had
not been provided to appropriate command and operational units to help
prepare DOD operational and cost estimates. These decisions are necessary to
guide the Department?s analyses of alternative courses of action to either
share the spectrum or vacate the spectrum band.

For example, Army Mobile Subscriber Equipment (MSE) program officials were
uncertain about how band sharing would be accomplished for operational and
cost analyses. Because of this uncertainty and the time constraints to
complete the study, they chose a general, high- level approach that did not
consider important factors with operational and cost implications in their
assessment of the MSE program. 34 Under this high- level approach, the Army
did not include the operational and cost impacts of relocating reserve units
to accommodate training requirements or costs at individual bases to
implement band sharing. In addition, the Army based its analysis on the
assumption that the accelerated development and production of the High
Capacity Line of Sight radio would replace the MSE radio. However, Army
assumptions about accelerating production of the replacement radio have not
been approved in the DOD budget and would require additional funding and

34 The MSE system consists of line of sight trunk radios linking switching
centers. DOD Analysis Requires

Additional Programmatic, Budgeting, and Schedule Guidance

Page 20 GAO- 01- 795 Defense Spectrum Management

reprogramming of funds to earlier years. In addition, replacing the MSE
radio with a new radio may not solve the Army?s tactical radio communication
problems should DOD have to share the 1755 to 1850 MHz band with third
generation mobile wireless users.

In another example of inadequate program and budget guidance, the Joint
Program Office for the Joint Tactical Air Combat Training System (JTCTS)
based its operational analysis on a plan to accelerate development and
fielding of that system, 35 but Navy and Air Force sponsors of JTCTS had not
formally reviewed or approved this accelerated plan. In July 2001, DOD
officials told us the contract for this program had been cancelled because
of schedule and performance problems. According to these officials, no
schedule has been approved for a new program, and any equipment from a new
program could not be fielded until at least the year 2014.

Without programmatic, budgetary, and scheduling guidance, command and
operational units risk uncertainty when assessing alternatives and making
assumptions in their analyses. Thus, the resulting operational and cost
estimates will also be uncertain.

The DOD report does not recognize or discuss planned development of
commercial wireless systems in other countries. Instead, it focuses on
potential operational degradation that may be caused by sharing the 1755 to
1850 MHz band with commercial wireless systems in the continental United
States. The services may need to operate on the same frequencies overseas as
they train on in the United States, and spectrum allocation decisions in
foreign nations could prevent the United States from using these frequencies
in other countries. According to a Joint Chiefs of Staff official, the Joint
Staff has produced several documents advocating greater flexibility through
the use of multiple frequencies in new defense systems for overseas
operations. However, the official said an analysis of overseas operations
was outside the scope of the DOD report.

Omission of any discussion of foreign spectrum developments creates an
unrecognized risk for DOD overseas operations, particularly DOD?s

35 JTCTS is the next generation of aircrew training systems displaying
weapon and aircraft information in real- time. It is intended to support
?rangeless? training that current systems cannot support. Reallocation
Effects on

International Agreements and Overseas Operations Not Described

Page 21 GAO- 01- 795 Defense Spectrum Management

control of its satellites. The International Telecommunication Union has
identified the 1710 to 1885 MHz band as one of several bands for possible
use by third generation mobile wireless systems worldwide. Overseas
development in this band over the long term by current second generation
commercial wireless systems or future development of this band by third
generation wireless systems could result in spectrum interference with U. S.
satellites. 36 Thus, DOD may have problems in the future controlling its
medium- to higher- earth orbit satellites from ground stations in the United
States because spectrum interference from overseas development of commercial
wireless systems in this band could be visible to these satellites. 37 This
interference could occur even if the United States does not allow commercial
use of the 1755 to 1850 MHz band in this country.

Unilateral reallocation of the 1755 to 1850 MHz band by the United States
could also have a potential negative impact on U. S. international spectrum
agreements and overseas military operations. For example, DOD provides
communications support to Great Britain and the North Atlantic Treaty
Organization (NATO) within the 1755 to 1850 MHz band in the United States.
Further, military air combat training systems in this same band are used by
allied nations during training in the United States. The National
Telecommunication and Information Administration Organization Act, as
amended, 38 requires, among other things, that NTIA, before acting on a
petition for relocation, determine that any proposed use of spectrum
frequency to which a federal entity will be relocated is consistent with
obligations undertaken by the United States in international agreements,
national security and public safety interests. Also, the proposed use must
be suitable for the technical characteristics of the band.

The November 2000 DSB study noted that the United States does not have a
national spectrum strategy that addresses international issues. The Board
said that international spectrum usage by DOD is governed by treaty, status
of forces agreements, 39 and other arrangements with allied

36 NTIA officials said the ITU identified several other bands for potential
third generation wireless use but that such identification does not preclude
use of any of these bands by any services to which they are now allocated
and that no priority is accorded to third generation wireless systems in the
radio regulations.

37 NTIA officials expect deployment overseas will not occur first in the
1710 to 1885 MHz band but in other bands recently auctioned throughout the
world. 38 47 U. S. C. sec. 923( g)( 2)( D).

39 Status of forces agreements and other arrangements govern interactions of
U. S. forces deployed overseas with foreign governments.

Page 22 GAO- 01- 795 Defense Spectrum Management

nations. The Board also recognized that other nations are asserting their
sovereign rights to manage their own spectrum, complicating deployments of
U. S. forces abroad. The Board recommended that the National Security
Council develop a national strategy given the increasing domestic and
international private sector demand for spectrum and the importance of
spectrum to national security. NTIA officials told us that federal agencies
requiring spectrum in other countries must work directly with foreign
countries because NTIA does not have responsibility or authority to develop
plans for federal use of foreign spectrum. For example, they said DOD must
press its strategic plans in NATO via country- to- country alliances and
develop international support for its requirements at international radio
conferences.

In the United States, there is a national allocation table wherein some 45
radiocommunication services are allocated spectrum. According to NTIA, the
allocation table and existing spectrum management processes constitute a
basic U. S. strategic spectrum plan, which covers all cases of spectrum use.
However, the national allocation table only reflects the current landscape
of spectrum use and does not provide a framework to guide spectrum decisions
into the future.

The Chairman of the FCC has also expressed support for overall improvement
of spectrum planning through a more coherent, nationally harmonized spectrum
policy. He said FCC is trying to improve overarching coordination of the
many existing policies within the Commission, but critical spectrum is also
controlled by other parts of the federal government, each with its own area
of responsibility. He said the administration and Congress both play
critical roles in allocating scarce spectrum resources to the highest and
best uses.

Original plans for identifying spectrum to support third generation mobile
wireless systems by July 30, 2001, and to auction licenses by September 30,
2002, were premature. We agree with FCC and the Department of Commerce that
delaying the identification of spectrum and the auction of licenses for
third generation wireless systems could serve the public interest. Adequate
information is not currently available to fully identify and address the
uncertainties and risks of reallocation. Thus, DOD and the federal
government could make decisions affecting national security without knowing
the full extent of risks they face or the steps available to reduce those
risks. Conclusions

Page 23 GAO- 01- 795 Defense Spectrum Management

Extending the current schedule for identification and auction of licenses
for this portion of the spectrum would allow DOD to complete technical and
operational assessments and to consider future spectrum requirements of DOD
systems. In addition, a delay would allow the federal government and DOD to
further consider the adequacy of existing national spectrum strategies
affecting international agreements and DOD overseas military operations, to
modify these strategies as necessary, and to incorporate these strategies
into a DOD long- range spectrum plan. Also, identifying potential
alternative bands of spectrum would provide DOD with needed information to
complete its technical, operational, and cost assessments. Including
relevant national and DOD strategies into spectrum reallocation decisions
would allow a more informed decision that balances national security
interests and private interests. These strategies could also guide detailed
studies that develop adequate measures to reallocate that spectrum and
develop detailed cost estimates before any auction is scheduled.

To more accurately assess the potential impacts to DOD if the 1755 to 1850
MHz band is selected for third generation systems in the United States, we
recommend that the Secretary of Defense

 complete a system- by- system analysis to determine existing and future
spectrum needs and requirements of systems in the 1755 to 1850 MHz band;

 prepare a long- range spectrum plan and make programmatic decisions
necessary to carry out that plan; and

 complete the technical, operational, and cost assessments of satellite
systems in the 1755 to 1850 MHz band and review and complete assessments of
other systems as necessary.

To provide DOD with adequate time and guidance to complete its plans and
analysis, we recommend that the Secretary of Commerce

 incorporate a sufficient amount of time into the new NTIA plan to select
spectrum for third generation mobile wireless systems to address the issues
discussed in this report, specifically with respect to satellite operations;

 direct NTIA, in conjunction with FCC, to identify comparable alternative
spectrum for use by the DOD systems before a decision is made to
Recommendations for

Executive Action

Page 24 GAO- 01- 795 Defense Spectrum Management

reallocate the 1755 to 1850 MHz band, should such an action be contemplated;
and

 coordinate with appropriate executive branch agencies to review existing
national spectrum management plans and policies, and, if necessary, to
establish a clearly defined national spectrum strategy reflecting DOD
requirements for international agreements and spectrum requirements to
operate overseas.

In commenting on a draft of this report, DOD, FCC, and the Department of
Commerce agreed with our findings and recommendations. All three agencies
and the National Security Council also provided technical comments to ensure
completeness and accuracy, and to provide clarity and balance. These
comments were reviewed and incorporated into our report as necessary.

We also clarified our recommendation that the Secretary of Commerce identify
alternative spectrum for DOD use before any reallocation decision. We
included DOD's suggestion that NTIA work in conjunction with FCC to identify
possible nongovernmental spectrum for DOD use.

DOD, the Department of Commerce, and FCC comments are reprinted in
appendixes III, IV, and V, respectively.

To determine whether the DOD process to define defense spectrum needs in the
1755 to 1850 MHz band experienced material constraints and whether the final
analysis and findings of the DOD report are complete or need further work,
we reviewed the Department?s data collections plans and visited selected
field locations, commands, and program offices to review implementation of
these plans. We also discussed data collection issues and related cost
issues with officials from the Office of the Assistant Secretary of Defense
for Command, Control, Communications, and Intelligence; the Joint Chiefs of
Staff Communications Division; the Office of Spectrum Analysis and
Management; and individual service spectrum management offices. In addition,
we reviewed findings and supporting material in the Department?s February
2001 spectrum report. We reviewed selected issues within the Department?s
technical analysis and a separate technical analysis prepared by a
communications industry working group. We also met with DOD and industry
officials responsible for preparing these technical analyses to obtain
further information and to discuss our findings on these issues. We reviewed
material presented by Aerospace officials and conducted a telephone
conference with those officials. Agency Comments

and Our Evaluation Scope and Methodology

Page 25 GAO- 01- 795 Defense Spectrum Management

(See appendix II for details of our analysis.) In addition, we met with NTIA
and FCC officials to discuss proposed rulemaking actions on spectrum related
issues.

We conducted our work from November 2000 through June 2001 in accordance
with generally accepted auditing standards.

We are sending copies of this report to appropriate congressional
committees. We are also sending this report to the Secretary of Defense, the
Secretary of the Air Force, the Secretary of the Army, the Secretary of the
Navy, the Secretary of Commerce, the Chairman of the Federal Communications
Commission, and the National Security Council. We will make copies available
to others upon request. The report will also be available on our homepage at
http:// www. gao. gov.

If you or your staff have any questions concerning this report, please call
me on (202) 512- 4841 or Charles Rey on (202) 512- 4174. Other major
contributors to this report were Rahul Gupta, Arthur Fine, Robert Hadley,
Judy Lasley, Gary Middleton, Keith Rhodes, Joseph Rizzo, Jay Tallon, and Dr.
Hai Tran.

Sincerely yours, Allen Li Director Acquisition and Sourcing Management

Appendix I: Spectrum Interference From Third Generation Systems to DOD
Satellites

Page 26 GAO- 01- 795 Defense Spectrum Management

While certain general engineering principles apply to any estimation of
spectrum interference, no single agreed- upon methodology or model exists
today to estimate potential interference to Department of Defense (DOD)
satellite operations from third generation mobile wireless systems. DOD
satellite operations control more than 120 satellites and their payloads,
including launch and early orbit operations, transmission of mission data,
on- orbit operations, and emergency and end- of- life operations. These
satellites provide for missile warning, navigation, military communications,
weather, and other defense missions. Specific satellites controlled or
served by DOD include (1) the Global Positioning System, which provides
navigational data and precise time transfer capability to military and
civilian users worldwide and (2) the Defense Satellite Communications
System, which provides essential command and control communications. DOD
controls these satellites through a set of common and dedicated Air Force
Space Command and Navy satellite control sites both inside and outside the
United States. This control is exercised through the Space Ground Link
Subsystem of the Air Force Satellite Control Network or through other
dedicated (mission specific) satellite control networks. Agencies and other
government users served include the National Command Authority, DOD,
Combatant Commanders, the Federal Aviation Administration, and National
Aeronautics and Space Administration. Because some DOD satellites- most
notably the Global Position System- broadcast information on an unrestricted
basis, DOD satellite control also supports civil and commercial interests.

Developing a standard methodology to estimate potential spectrum
interference to DOD satellite operations from third generation mobile
wireless systems is particularly challenging because key assumptions must be
made not only about the technical characteristics of mobile wireless systems
to be fielded many years in the future, but also about the extent to which
these systems will be deployed worldwide in urban, suburban, and rural
environments. Unlike other DOD systems, satellites are unique in that they
may be exposed to the aggregate level of interference emanating from
significant portions of the globe. Moreover, because some DOD satellites may
need to operate within the 1755 to 1850 MHz band until at least 2017 (and
perhaps as late as 2030), interference estimates must consider the extent of
third generation mobile wireless systems (IMT- 2000) build- out over the
next several decades. We reviewed, in detail, two models- a DOD model and an
industry model- that took different approaches to estimating the level of
interference DOD satellites could experience from this worldwide build- out.
The level of interference predicted by these two models differed
significantly. We also performed a limited review of a third Air Force model
being developed by the Appendix I: Spectrum Interference From

Third Generation Systems to DOD Satellites

Appendix I: Spectrum Interference From Third Generation Systems to DOD
Satellites

Page 27 GAO- 01- 795 Defense Spectrum Management

Aerospace Corporation, a federally funded research and development center.
The Aerospace Corporation is developing its model in support of the Air
Force?s Space and Missile Systems Center and Air Force Space Command to
better understand why the DOD and industry models produced such different
results and to attempt to develop a trusted and credible interference model
to generate results that all stakeholders can agree upon and that can be
adopted as the standard in support of key decisions. 1

We found that neither the DOD nor industry model is fully mature and that
they differ strikingly from one another in many of their key assumptions. In
our opinion, neither model is mature enough to support a near- term decision
by the United States regarding whether DOD satellites can share spectrum
with third generation mobile wireless systems, and both incorporate
questionable assumptions that drastically influence their predicted level of
spectrum interference. Moreover, because IMT- 2000 will be a worldwide
system and because the International Telecommunication Union has identified
the 1710 to1855 MHz band (among other bands) for use by countries wishing to
implement IMT- 2000, some degree of spectrum sharing could become
unavoidable regardless of any reallocation decisions made within the United
States. If the aggregate level of the non- U. S. interference in the 1755 to
1850 MHz band proves to be significant, DOD satellites could be adversely
affected by overseas deployment of IMT- 2000.

As a result of this uncertainty, DOD faces an unquantified risk of
operational degradation to its satellite operations if a decision is made to
allow industry into the 1755 to 1850 MHz band in the United States before
more accurate information is available upon which to base a decision.
However, DOD may face such a risk irrespective of that decision due to
overseas build- out of IMT- 2000.

1 We analyzed the DOD and industry models to the point that we could
reproduce their estimates of the level of spectrum interference generated by
the worldwide build- out of third generation systems. We reviewed the Air
Force model but did not attempt to replicate its findings. Summary

Appendix I: Spectrum Interference From Third Generation Systems to DOD
Satellites

Page 28 GAO- 01- 795 Defense Spectrum Management

DOD and industry studies do not agree about how to predict the level of
potential interference to satellite receivers from the worldwide build- out
of third generation wireless systems or about the effect of this
interference on satellite operations. DOD predicts interference levels
significant enough to affect satellite operations by the year 2006,
particularly for lowearth- orbit satellites. Industry, on the other hand,
predicts interference levels that are too low to adversely affect DOD
satellite operations until at least the year 2015 and perhaps never. The Air
Force model incorporates features more closely aligned with industry?s
methods. However, the Air Force model?s worst- case results for operating
under nominal conditions agree with the DOD report?s overall conclusion that
third generation wireless systems present potential interference problems to
DOD satellites. The Air Force model?s analysis suggests, however, that
interference is much less likely to occur with low earth- orbit satellites
(as DOD had concluded) than with medium and high earth- orbit satellites.
For example, the Air Force model?s analysis suggests that, under certain
modeling conditions, the Air Force may experience difficulty communicating
effectively with Global Positioning System and Defense Support Program
satellites because of interference from the build- out of third generation
systems.

A number of key data elements are necessary to finalize estimates of
potential third generation wireless systems interference to DOD satellite
operations. First, the locations and sizes of the geographic areas most
likely to be served by the mobile wireless industry must be properly
estimated, and translated into an estimate of the number of third generation
base stations needed to serve each geographic area. Second, the amount of
power needed by each base station to communicate with mobile units within
its area of coverage (typically referred to as a mobile wireless base
station?s ?cell?) must be calculated. Third, the fraction of that power that
will reach the satellite must be estimated. In addition, assuming some level
of consensus can be reached regarding the best assumptions and methodology
for predicting the level of interference at satellite orbital altitudes from
the worldwide build- out of third generation wireless systems, the
operational impact of this interference on DOD?s ability to communicate with
particular types of satellites from specific satellite control sites (both
within and outside the United States) under specific operating conditions
would need to be fully analyzed and presented to key decisionmakers before
reasonable decisions can be made. Different Analytical

Approaches

Appendix I: Spectrum Interference From Third Generation Systems to DOD
Satellites

Page 29 GAO- 01- 795 Defense Spectrum Management

While they disagree on the specifics of the interference estimates, both DOD
and industry agree that low- power mobile stations are much less likely to
cause significant interference with DOD satellite operations. However, they
disagree about the potential interference to satellite operations from much
higher- powered base stations. Each base station services a cell of a given
size- up to about 10 kilometers in radius (or 314 square kilometers),
although smaller base stations are more likely to be used within dense urban
environments. Both DOD and industry agree that the first step in determining
the interference generated by IMT- 2000 base stations is to estimate the
area of each geographic region likely to be served by third generation
mobile wireless systems. However, neither DOD nor industry was able to
locate a comprehensive database of the actual areas (geographic dimensions)
of the world?s population centers. The only information readily available to
both was the geographic location and population of the world?s largest urban
centers- generally those with a population of 100,000 or more. Thus, both
DOD and industry used the population of each urban center to estimate how
much land area that population lives on. Those area estimates were then used
to determine the interference contribution from base stations within that
area. DOD?s approach did not explicitly estimate the number of base stations
required to service a given area. Rather, DOD assumed that the power
radiated per unit area from IMT- 2000 base stations would be a constant
derived from reports and recommendations of the International
Telecommunication Union, and then multiplied this constant by the estimated
size of each urban area to determine the power radiated from that area.
Industry?s methodology used their estimated size of each service area to
calculate the number of maximum- sized (10 kilometer radius) base stations
that would be needed both to service that area, and to serve as an

?upper bound? on the estimate of total power radiating from that area if
smaller, more numerous base stations were used.

The first problem we identified is that both the DOD and industry models are
likely to have significantly underestimated the total number and size of
areas (urban, suburban, and rural) that are potential markets for third
generation systems. This problem exists, in part, because no standard
definitions exist of the most likely market areas for third generation
mobile wireless systems. The International Telecommunication Union
identified potential uses of third generation systems for urban, suburban,
and rural markets, but detailed forecasts of IMT- 2000 build- out are not
available. Thus, both models are based on incomplete and inconsistent
databases as a source to calculate the number of urban centers and other
potential areas for service delivery. Third Generation Mobile

Wireless Base Station Coverage Areas

Appendix I: Spectrum Interference From Third Generation Systems to DOD
Satellites

Page 30 GAO- 01- 795 Defense Spectrum Management

For example, DOD used a database of 2,763 urban centers worldwide, of which
approximately 320 were urban centers within the United States. 2 Industry
used a database of 3,312 urban centers worldwide, of which approximately 209
were urban centers within the United States. However, both of the databases
included only capitals and urban centers with populations over 100, 000. The
databases excluded many urban centers below 100,000, as well as suburban and
rural areas. The Air Force model uses the same database of 2,763 urban
centers used by DOD. However, the Air Force?s model also includes an
estimate of rural geographic areas that could be served by third generation
systems. The Air Force is working to improve the overall quality of its
population database. A 1990 U. S. Census Bureau report indicates that,
within the United States, there are 224 urbanized areas 3 with a population
of more than 100,000, and 396 urbanized areas with a population of more than
50,000; many of the additional 172 urbanized areas with a population between
50,000 and 100,000 are not included in any of the three models.

The population estimates for both the DOD and industry models are also
incomplete because they do not include any estimate of population growth.
The Air Force?s model includes an estimation of the population growth based
on population databases from 2 different years. DOD estimates the lifetime
of a number of satellites may extend until 2017 and beyond, and perhaps
until as late as the year 2030. Another indicator that service areas may be
significantly underestimated is the FCC report on broadband (high data rate)
use in the United States. This report documents high speed Internet access
in 59 percent of the postal zip codes in the United States; according to the
report, 91 percent of the U. S. population is in these zip code areas. We
believe it is reasonable to expect that these counties would be candidates
for third generation systems.

The second problem is that the DOD and industry models used different
equations to estimate the geographic area covered by their urban centers.
Both models analyzed 1990- era census data for a number of urban centers
whose areas are known to estimate the size of the rest of urban centers in
their databases. However, each model used different coefficients to

2 Best estimate based on urban center locations; the database of urban
centers used by DOD was not coded by country. Moreover, the origins of this
database were not well documented.

3 The United States? urbanized areas are defined by the U. S. Census Bureau.
Each consists of at least one central city or place and its urban fringe.

Appendix I: Spectrum Interference From Third Generation Systems to DOD
Satellites

Page 31 GAO- 01- 795 Defense Spectrum Management

calculate urban center size from population. Industry urban center sizes
were bigger than DOD urban center sizes, but the Air Force model estimates
urban center sizes are bigger than either DOD or industry estimates. In
addition, DOD used only three urban centers in the United States, and five
urban centers outside the United States, to obtain an average and appears to
have used incorrect numbers for urban center size for two of the three U. S.
urban centers. Thus, DOD calculated an average urban center size of 144.2
square kilometers per 1 million of population. Using 1990 U. S. census data,
we calculated an average size for the 396 urbanized areas of about 1, 000
square kilometers per one million people, or about seven times as much area
for a population of the same size.

DOD recognizes that its database may significantly underestimate potential
interference coming from rural and interstate highway base stations and
smaller urban areas that will likely be served by third generation systems.
According to DOD officials, their preliminary calculations suggest that even
partial coverage of rural U. S. interstate highways alone will result in the
deployment of thousands of additional high- power base stations. The
industry model did not include any estimate for rural interference. The Air
Force?s model included estimates of interference from urban, suburban, and
rural areas. According to Aerospace Corporation representatives, results
from this model showed that significant interference came from rural areas.

Once the number of base stations is determined, the next required data
element is how much power each base station must use to service mobile
stations within its cell. DOD and industry disagree markedly in their
estimate of power radiating from the area covered by an individual 10-
kilometer radius cell served by a single base station. Industry placed an
effective upper bounds of 30 watts of power for this 314 square kilometer
area. 4 Using this set of assumptions, DOD?s ITU- based method for
calculating power from this 314 square kilometer area would be about 2,600
watts of power, or nearly 100 times as much power. As noted above, DOD?s
analysis did not explicitly label this area as the service area of a large
base station. The Air Force?s model calculated the level of power required
to service a geographic area, taking into consideration the propagation
model suitable to a particular area. Their results showed that

4 Based on IMT- 2000 technology over a 200 kHz bandwidth. Power Requirements
for

Each Base Station

Appendix I: Spectrum Interference From Third Generation Systems to DOD
Satellites

Page 32 GAO- 01- 795 Defense Spectrum Management

power radiating from an urban area would be significantly greater than
predicted by industry, but considerably less than predicted by DOD.

According to Aerospace Corporation representatives, the Air Force?s model
uses a higher power level than industry because their investigation of
second generation wireless systems suggests that power loss from each base
station is higher in urban areas than industry assumed. Buildings in urban
areas significantly lower distance covered by base stations. Industry
assumed that the power radiated from IMT- 2000 base stations would be
proportional to the square of the base station cell?s radius irrespective of
the propagation environment. For example, a 314 square kilometer rural area,
served by a single 10- kilometer base station, would radiate essentially the
same amount of power as an urban area of the same size, served by 100 1-
kilometer base stations. According to Aerospace Corporation representatives,
the Air Force?s model relies upon calculations of the specific distances
that a base station could complete a call to a mobile unit, resulting in
significantly higher estimate of power levels required of urban base
stations within each area served by those base stations.

The third required data element is a calculation of the amount of power from
all base stations that reaches the satellites. Power from third generation
systems reaching a DOD satellite is the sum of contributions from all of the
base stations in view of that satellite at any given moment in time, and
that contribution is the power each station generates times the fraction of
the amount of that power that arrives in space at the satellite?s orbital
altitude. After attempting to determine the number of base stations and the
power level required by urban base stations, the next issue is to determine
how much of the power generated by each base station actually reaches DOD
satellites. The International Telecommunication Union provides little
guidance for estimating interference from IMT- 2000 ground base stations to
satellite operations.

The answer to this issue depends on two factors: (1) the way in which a base
station?s antenna concentrates the power it transmits in certain preferred
directions (generally referred to as the antenna?s gain pattern), and (2)
how much power radiated by a base station?s antenna is lost to atmospheric
and other environmental effects before reaching orbital altitude. DOD
assumed that each base station?s antenna radiates power isotropically- that
is, equally in all directions, but that 90 percent of this power would be
lost due to environmental factors. Industry, on the other hand, employed an
antenna gain pattern that assumes that the power Base Station Power

Radiated Into Space

Appendix I: Spectrum Interference From Third Generation Systems to DOD
Satellites

Page 33 GAO- 01- 795 Defense Spectrum Management

radiated from a base station would be greatest at low elevation angles
(directed just below the horizon), and that only a small fraction of the
base station?s power would be directed toward satellites that are not at or
near the horizon. Industry further assumed that, at low elevation angles, 90
percent of the base station?s power transmitted in that direction would be
lost to the environment, but that most or all of the power radiated at high
elevation angles would reach space. However, as noted above, overall the
power radiated at high elevation angles represents a very small portion of
the base station?s radiated power. For intermediate elevation angles,
industry assumed that power loss would fall between these two values - 0 and
90 percent. The Air Force?s model employed the same antenna gain pattern as
industry, but assumed that, other than at very low elevation angles, only a
small fraction of this power would be lost to the environment before
reaching space. As noted earlier, the Air Force?s model still found
potential problems from third generation wireless systems to satellite
operations.

Calculating power levels reaching satellites is also difficult because data
do not exist today to prepare a baseline, which can be compared to power
levels projected for third generation systems. Commercial systems now
operate in the 1755 to 1850 MHz band overseas, and DOD has military systems
operating in this band within the United States and overseas. However, DOD
officials said they have no data on existing second generation system base
stations and deployment patterns. DOD officials told us that no interference
has been experienced to date with satellite operations because overseas
mobile unit densities on a given satellite control channel are much lower
than those expected for third generation systems.

Assuming some consensus can be reached regarding how much power from IMT-
2000 systems is likely to reach orbital altitudes, agreement would still be
required on how this level of interference is likely to affect satellite
operations. This analysis is complicated by the fact that, under any
modeling assumptions, IMT- 2000 interference is likely to vary greatly over
the orbital ?shell? at a given latitude, particularly for low- earth orbits.
Moreover, each type of satellite can be affected differently by interference
from third generation wireless systems because (1) any specific type of
satellite can follow a different orbit, (2) the technical characteristics of
satellites are different, and (3) each type of satellite has different
abilities to respond to DOD commands in the presence of interference because
the mission of each satellite type is different. Operational Impact of

Potential Interference

Appendix I: Spectrum Interference From Third Generation Systems to DOD
Satellites

Page 34 GAO- 01- 795 Defense Spectrum Management

Neither DOD nor industry attempted a technical or an operational analysis on
each satellite to determine how specific satellites can be affected by third
generation interference from specific satellite control sites. 5 Instead,
they both assumed a standard response irrespective of specific satellite
characteristics. However, DOD does acknowledge that parameters that can
affect its ability to communicate vary from satellite to satellite. As a
result, accurate and complete information is not available to determine if
operational workarounds can be planned to avoid loss of control or degraded
satellite operations.

The Air Force?s model suggests potential interference problems from third
generation base stations for DOD satellites. The Air Force model?s results
were, in fact, close to the DOD results, even though its methodology was
closer to that of industry. However, because it recognized the differences
between satellite types, the Air Force is also using its model to analyze
the potential effects from third generation wireless systems on specific
types of satellite operations. The Air Force?s analysis calculates
interference values for several types of satellites through their entire
orbits, and calculated probability of successful contacts for those
satellites at any given point in their orbits. For example, the Air Force
model?s output data show that interference levels for the Global Positioning
Satellite were high enough, under certain modeling assumptions, to prevent
successful contact, with an acceptable margin of safety, for approximately
60 percent of an orbit. The Air Force model?s output data, however, has not
yet presented interference levels for the period of time that the satellite
is in sight of a specific ground station. Thus, the probability of
successful contact is not available for each satellite at each ground
station. Such an analysis is necessary to determine if alternative satellite
contact plans can be devised to enable successful contacts.

The Air Force study recognizes that specific operational impact analyses
should be performed for all ground stations and DOD satellites. The study
recommends that the Air Force model be evolved to include (1) improved
population modeling, (2) modeling of satellite contacts from specific

5 We recognize that industry may lack the information to perform such a
detailed analysis.

Appendix I: Spectrum Interference From Third Generation Systems to DOD
Satellites

Page 35 GAO- 01- 795 Defense Spectrum Management

ground stations (3) analysis of operations under both nominal and adverse
conditions, and (4) inclusion of more types of DOD satellites.

Appendix II: Replication of DOD?s and Industry?s IMT- 2000 Base Station
Interference Models

Page 36 GAO- 01- 795 Defense Spectrum Management

DOD and industry each developed models to predict the level of interference
to DOD satellites that could arise from the worldwide buildout of IMT- 2000
base stations. The level of interference predicted by these two models
differed significantly. In order to determine whether DOD?s and industry?s
interference calculations could be reproduced and to obtain a better
understanding of the factors contributing to the differences between DOD?s
and industry?s results, we replicated DOD?s and industry?s interference
models. We also reviewed a third IMT- 2000 interference study being
developed for the Air Force by the Aerospace Corporation, but did not
attempt to reproduce its findings. This appendix briefly describes DOD?s and
industry?s approaches to modeling IMT- 2000 interference and our approach to
reconstructing those two models, and offers several observations regarding
the differing assumptions used by DOD and industry for various components of
their interference models.

Our replication of DOD and industry models was based on (1) the mathematical
descriptions of the models provided in the referenced studies, (2) copies of
the population databases used in each study (which we obtained from DOD and
industry representatives), and (3) equations adopted from a standard space
mission planning textbook that describe the geometrical relationship between
a point on the surface of the earth and a point in space at a satellite?s
orbital altitude. To clarify certain matters regarding calculations used in
these models, we also met and corresponded with DOD and industry
representatives on several occasions. As shown in tables 2 and 3, we were
able to reproduce both DOD?s and industry?s estimates for peak IMT- 2000
interference levels to within about 0.1 dB for most orbital altitudes.

Each analysis led to a different conclusion because, while certain general
engineering principles apply to estimating spectrum interference, no single
methodology or model exists today to estimate potential spectrum
interference to DOD satellite operations from third generation mobile
wireless systems. DOD?s IMT- 2000 interference model is described in DOD?s
final report on accommodating IMT- 2000 within the 1755 to 1850 Appendix II:
Replication of DOD?s and

Industry?s IMT- 2000 Base Station Interference Models

Summary DOD?s and Industry?s Approaches to IMT- 2000 Interference Modeling

Appendix II: Replication of DOD?s and Industry?s IMT- 2000 Base Station
Interference Models

Page 37 GAO- 01- 795 Defense Spectrum Management

MHz band. 1 Industry?s model is described in the Report of the Working Group
on Satellite Control Systems. 2

Each model incorporated a database describing the physical location
(longitude and latitude) and population of the world?s largest urban
centers, assumed that IMT- 2000 service would be provided within the 1755 to
1850 MHz band at each of these locations, and then approximated the size
(geographic extent) of each urban center using a model- specific parametric
equation to relate an urban center?s population to its geographic area. For
each geographic area, both models then computed the aggregate power spectral
density (PSD) that would radiate from all of the IMT- 2000 base stations
needed to provide service within that geographic area. To reduce the
computational complexity of this calculation, both models assumed that, for
typical orbital altitudes, the aggregate PSD received from all of the IMT-
2000 base stations within a given urban area could be closely approximated
by a point source at the center of the geographic area whose PSD is the sum
of the PSD from each IMT- 2000 base station operating within the area
served. Both models then employed link budget calculations to compute the
interference contribution from each urban center at any given point S on an
?orbital

shell? defined by the altitude of a class of satellites. In logarithmic
form, the interference contribution for each visible urban center is given
by:

E MHz km T T R L f r G I I ï¿½ ï¿½ + ï¿½ = ) log( 20 ) log( 20 44 . 32 ) log( 10
[1] where:

R I = power spectral density received at S from the urban center in decibel
watts per Hertz

1 Department of Defense: Investigation of the Feasibility of Accommodating
the International Mobile Telecommunications (IMT) 2000 Within the 1755- 1850
MHz Band;

(Feb. 9, 2001). 2 Evaluation of Sharing between International Mobile
Telecommunications (IMT) 2000 Technology and Satellite Control Systems
Operating in the Band 1755- 1850 MHz

(Feb. 19, 2001) [Note: This report was filed with the Federal Communications
Commission as Attachment II of the Report of the Industry Association Group
on Identification of Spectrum For 3G Services (Feb. 22, 2001) in response to
the Commission?s notice of proposed rulemaking on 3G.]

Appendix II: Replication of DOD?s and Industry?s IMT- 2000 Base Station
Interference Models

Page 38 GAO- 01- 795 Defense Spectrum Management

T I = power spectral density transmitted from the urban center in watts per
Hertz

T G = gain of the transmitting antenna relative to an isotropic antenna
(dBi)

km r = slant range from the urban center to the point S in kilometers

MHz f = operating frequency in megahertz

E L = environmental loss in dB Note that specific assumptions about the gain
of the satellite receiver?s antenna are suppressed in equation [1] in order
to consider the interference environment generated by IMT- 2000 base
stations irrespective of a particular satellite receiver?s antenna
specifications. However, it is assumed the satellite receiver?s antenna is
gain- limited.

Calculating the interference contribution for each urban center required
analyzing the particular earth- space geometry between the point S and the
location of each urban center. Specifically, only those urban centers
visible from the point S contribute to the aggregate interference level.
Thus, the elevation angle of the point S as seen from each urban center had
to be calculated to determine whether that urban center was visible (above
the local horizon). Moreover, for the industry model, the transmitting
antenna?s gain, and the environmental loss were functions of this local
elevation angle. In addition, the slant range between each urban center and
the point S had to be calculated to complete the interference calculation.
After obtaining solutions for equation [1] for each visible urban center,
both models then summed these contributions to arrive at the aggregate
interference level at the point S.

Finally, both models calculated the aggregate interference at each point Sï¿½
on a lattice that spanned the orbital shell from -180 degrees to 180 degrees
longitude, and from -90 degrees to 90 degrees latitude to find the
approximate location and value of the highest predicted interference level
worldwide. This interference level was then used to examine the ?worst

case? effects of IMT- 2000 interference on satellite receivers. Each model
performed this analysis for four orbital shells representing four typical
satellite orbital altitudes: (1) 250 km, the typical orbital altitude of the
space shuttle; (2) 833 km, the orbital altitude of DOD?s Defense
Meteorological Satellite Program (DMSP) satellites; (3) 20,200 km, the

Appendix II: Replication of DOD?s and Industry?s IMT- 2000 Base Station
Interference Models

Page 39 GAO- 01- 795 Defense Spectrum Management

orbital altitude of DOD?s Global Positioning System (GPS) satellites; and
(4) 35,784 km, 3 the orbital altitude of several types of geosynchronous
satellites including the Defense Support Program (DSP) and the Defense
Satellite Communications System (DSCS). Each step in this calculation is
described in some detail later in this appendix.

DOD?s and industry?s results for peak IMT- 2000 interference differed
significantly. As shown in table 1, DOD?s peak PSD levels are about 5 dB
higher than industry?s levels for each of the orbital shells analyzed. For
both models, values in the table refer to the PSD at the front end of the
satellite?s antenna and are, therefore, independent of the specific
characteristics of this antenna. 4

Table 1: Results of DOD and Industry Interference Calculations Satellite
altitude

(km) DOD peak PSD (dBw/ Hz) Industry peak PSD

(dBW/ Hz) Difference (dB)

250 - 161.5 - 166.2 4. 7 833 - 166.6 - 173.2 6. 6 20,200 - 186.7 - 191.9 5.
2 35,784 a - 191.1 - 195.8 4. 7 a Industry used a slightly smaller value for
this geosynchronous earth orbit altitude; the use of DOD?s

slightly larger value would likely decrease industry?s peak PSD for this
altitude by less than 0.01 dB

In order to determine whether DOD?s and industry?s interference calculations
could be reproduced and to obtain a better understanding of the factors
contributing to the differences between DOD?s and industry?s results, we
replicated DOD?s and industry?s interference models. In general, sufficient
information was available in the published DOD and industry reports to
recreate the two models once we had obtained the models? urban center
population databases. Both the DOD and industry reports provided their
versions of equation [1] used to calculate the

3 Industry used a slightly smaller assumed value for its analysis of the
geosynchronous orbital shell. 4 It should be noted that industry?s report on
IMT- 2000 interference to satellites shows interference results that are 5
dB lower in each case than those given in table 1 because industry?s
interference levels were calculated at the front end of the satellite
receiver after accounting for the (assumed) - 5 dB gain of the satellite
receiver?s antenna. Because DOD?s reported PSD levels are taken at the front
end of the satellite antenna, industry?s numbers have been adjusted here to
make possible an ?apples- to- apples? comparison. Results of DOD and

Industry Interference Calculations

Analysis of DOD?s and Industry?s IMT- 2000 Interference Calculations

Appendix II: Replication of DOD?s and Industry?s IMT- 2000 Base Station
Interference Models

Page 40 GAO- 01- 795 Defense Spectrum Management

interference from a single urban center to a point S on the orbital shell at
a given slant range and local elevation angle. To clarify certain matters
regarding calculations used in these models, we also met and corresponded
with DOD and industry representatives on several occasions. We developed
versions of DOD?s and industry?s models that incorporated all of the salient
features of the original models.

DOD?s and industry?s databases containing the longitude, latitude and
population of the world?s largest urban centers were obtained from DOD and
industry representatives and incorporated into our models. For each model,
we first developed a ?single- point? PSD calculation module that could
calculate the aggregate PSD from all of the urban centers in the model?s
database within view of any single point S (defined by its longitude,
latitude and orbital altitude). We applied slant range and satellite viewing
angle formulas from a standard textbook on space mission planning. 5
Specifically, we applied suitable forms of equations from this source in our
replication of DOD?s and industry?s models. Equations derived from the DOD
and industry reports that calculated an urban center?s geographic area, the
aggregate PSD radiating from an urban center, the base station antenna gain,
and base station PSD environmental losses were also incorporated into our
models. Finally a suitable form of equation [1] was incorporated into our
models to calculate the interference contribution from each visible urban
center at the point S.

To calculate the PSD for a lattice of points that spanned the entire orbital
shell, we incorporated our ?single- point? PSD calculation modules into an
iterative routine that calculated and stored the location information
(longitude and latitude) and associated PSD value for each point on the
orbital shell?s lattice. The data generated were then analyzed to determine
the value and location of the peak PSD for the orbital shell. We carried out
calculations over a 2- degree by 2- degree lattice using the same four
orbital shells analyzed by DOD and industry. Each key element of the
modeling exercise is discussed briefly below. In addition, selected
observations on DOD?s and industry?s assumptions with respect to those
elements of the models are also included.

5 Space Mission Analysis and Design; 2nd. ed. Wiley J. Larson and James R.
Wertz (editors); Kluwer Academic Publishers (1992); Chapter 5; ?Space
Mission Geometry?; pp. 110- 111.

Appendix II: Replication of DOD?s and Industry?s IMT- 2000 Base Station
Interference Models

Page 41 GAO- 01- 795 Defense Spectrum Management

1. IMT- 2000 Base Station Service Area for Each Urban Center - DOD and
industry both used databases describing the location (longitude and
latitude) and population of the world?s largest urban centers, and then
approximated the size (geographic extent) of each urban center using a
model- specific parametric equation to relate an urban center?s population
to its geographic area.

1.1 Population Databases - DOD?s model used a database describing the
physical location (longitude and latitude) and population of 2,763 of the
world?s most populous urban centers, with a total population of 1.33
billion. Data were obtained from a variety of sources, including United
Nations statistical data and U. S. Census Bureau data. Industry used a
similar (though not identical) database of 3,312 of the world?s most
populous urban centers obtained from United Nations data with a total
population of 1.64 billion. Both of these databases were generally limited
to urban centers with populations of 100,000 or more, and both were based on
1990- era census data. In contrast, the United Nations Population Division
has estimated the world?s urban population at about 2.85 billion in 2000 and
projects that this figure will grow to about 3.82 billion by 2015.

1.2 Urban Center Area Estimation - Both DOD and industry approximated the
size (geographic extent) of the urban centers in their respective database
by applying a parametric equation of the form:

P RP ï¿½ = [2] to each urban center where:

P R = radius of the urban center with population P

P = urban center?s population

= constant of proportionality

= scaling factor DOD used data on the population and land area of 8 large
urban centers around the world (including 3 urban centers in the United
States) to calculate an average inverse population density of 144.2 km 2 per
million people, and applied this constant of proportionality to all 2,763
urban centers in its database. Assuming that the area of each

Appendix II: Replication of DOD?s and Industry?s IMT- 2000 Base Station
Interference Models

Page 42 GAO- 01- 795 Defense Spectrum Management

urban center is a circle, this relationship can be expressed in the
parametric form of equation [2] with:

006775 . 000 , 000 , 1 2 . 144 ï¿½ = DOD [2A]

5 . 0 = DOD [2B] Industry used a more extensive sampling of data from the
1990 U. S. census to obtain a ?best fit? to the parametric equation
described in equation [2] for cities within the United States, and used
United Nations data to obtain a ?best fit? for cities outside the United
States. For cities within the United States, industry calculated values of:

035 . 0 = Industry [2C]

44 . 0 = Industry [2D] According to industry?s report, for cities outside
the United States, this scaling factor remains unchanged, but a smaller
constant of proportionality was calculated to account for the fact that
urban centers outside the United States tend to be more densely populated
than urban centers within the United States. However, the principal author
of industry?s study informed us that he had, in fact, used the single, more
conservative constant shown in equation [2C].

Our models replicate these calculations to obtain estimates for the
geographic area of each of the urban centers. It should be noted that
industry?s model assumed that urban centers are 4 to 7 times larger than DOD
assumed. For example, while DOD assumed that an urban center of 1 million
people would have a total land area of 144.2 km 2 , industry assumed that
this urban center would have a total land area of 733.2 km 2 -about 5 times
as large. We also compared these values with the average inverse population
density calculated from our analysis of data from the 1990 U. S. Census
Bureau report on the population and

Appendix II: Replication of DOD?s and Industry?s IMT- 2000 Base Station
Interference Models

Page 43 GAO- 01- 795 Defense Spectrum Management

land area of the 396 urbanized areas in the United States. 6 We found that
the average inverse population density for these 396 urbanized areas was
about 1,000 km 2 /million people. 7

2. PSD Radiated by IMT- 2000 Base Stations from an Urban Center - Once the
geographic area of each urban center had been approximated, DOD and industry
models each calculated the aggregate PSD for all of the IMT- 2000 base
stations that were assumed to be operating within the boundaries of each
area. DOD?s approach applied a constant PSD per unit area derived from
reports and recommendations of the International Telecommunication Union.
Industry?s approach was based upon consideration of the technical
characteristics of IMT- 2000 base stations employing UWC- 136 (TDMA)
technology. Because DOD?s approach was much simpler than industry?s, the two
approaches are described separately here.

2.1 DOD?s Approach - DOD?s approach to calculating the PSD radiated by IMT-
2000 base stations from a given urban center was very simple. DOD assumed
that IMT- 2000 base stations serving all urban land areas would radiate a
constant PSD per unit area of 41 ï¿½ watts/ Hz/ km 2 , or - 43.9 dBW/ Hz/ km 2
. This figure was derived from one report and one recommendation of the
International Telecommunication Union (ITU). 8 DOD did not attempt to
explicitly calculate the number of base stations to which this PSD value
corresponded. Rather, DOD?s model multiplies this value by the size of the
urban center calculated using equation [2]. We incorporated a suitable form
of this calculation into our model.

6 1990 Census of Population and Housing: Supplementary Reports - Urbanized
Areas of the United Sates and Puerto Rico (1990 CPH- S- 1- 2). U. S.
Department of Commerce; Economics and Statistics Administration; Bureau of
the Census (December 1993); table 2,

?Rank of Urbanized Areas by 1990 Population; Housing Units, Area
Measurements and Density: 1990.? [Note: Nearly identical information on the
1990 population and land area of these 396 urbanized areas was also found on
a commercial website -

www. demographia. com]. 7 Calculated as {Total land area for the 396
urbanized areas} ï¿½ 1,000,000/{ Total population of the 396 urbanized areas}.
8 Recommendation ITU- R M. 687- 2 (1990) estimated a PSD of 38 W/ Hz/ km 2
for IMT- 2000 base and mobile stations. DOD updated this figure using
information from ITU- R Report M. 2023. We did not attempt to replicate
DOD?s calculation but accepted the 41 W/ Hz/ km 2 as the input PSD for our
reproduction of DOD?s model.

Appendix II: Replication of DOD?s and Industry?s IMT- 2000 Base Station
Interference Models

Page 44 GAO- 01- 795 Defense Spectrum Management

2.2 Industry?s Approach - Industry?s approach to calculating the PSD
radiated from IMT- 2000 base stations from a given urban center was more
complex. Industry first estimated the number of base stations of a
particular size and deployment penetration factor required to serve the
urban center?s area (as calculated using equation [2].) Industry assumed
that each base station would serve a circular area of radius

h R . Industry then calculated that while the relative mix of differentsized
base stations might change from one urban center to the next, the maximum
number of any particular size of base stations of service radius h R ,
needed to fully serve an urban center could be calculated using equation (6)
from industry?s report. Finally, industry assumed that the PSD radiated from
a base station with a smaller service radius

h R relative to the PSD of a base station with the largest assumed service
radius (10 km) h max R would be directly proportional to the square of ratio
of these two radii, that is:

2 )

) =

h max h

h max h R R P(R P(R [3]

Industry argued that adding 2 to the number of maximum- sized base stations
calculated using equation (6) from the industry report would result in an
aggregate PSD for each urban center that was as least as large as the
aggregate PSD from any other relative mix of 3 differentsized base stations.
Industry?s ?upper bound? method is described in more detail in its published
report. We applied suitable forms of equations derived from industry?s
?upper bond? method to our model.

Unlike DOD, industry did not assume that each urban center in its database
would be fully ?built out.? Rather, industry developed a population density-
dependent deployment penetration factor to account for the fact that less
densely populated urban centers may not achieve the same level of IMT- 2000
penetration as more densely populated urban centers. Industry?s deployment
penetration factor curve appears as figure 2 of its report. However, because
the industry report did not explain, in detail, its assumed values for this
curve, we did not initially have enough data to replicate this curve
precisely for all population densities. Consequently, we approximated this
curve by visually extracting data points from the published curve and then
fitting this data to a curve that would adequately reproduced the

Appendix II: Replication of DOD?s and Industry?s IMT- 2000 Base Station
Interference Models

Page 45 GAO- 01- 795 Defense Spectrum Management

published curve. To approximate industry?s penetration curve, we used a
four- parameter equation of the form:

( ) ) exp( 1 ) ( e b P cD aD D + = [4]

where:

) (D P = IMT- 2000 deployment penetration factor

D = urban center?s population density (people/ km 2 ).

a, b, c, e are a set of ?best fit? parameters During subsequent
correspondence with the principal author of the industry study, we were
provided with the table of values for population density versus deployment
penetration factor used in the industry model. However, a comparison between
the estimated total number of base stations worldwide predicted using our
?best fit? penetration curve and the total number predicted using industry?s
approach showed less than a 0.1 percent difference. Consequently, we did not
adjust this factor in our replication of industry?s model.

We noted that industry?s introduction of a deployment penetration factor
into its model ultimately made very little difference to the predicted
aggregate PSD level because most urban centers in industry?s database had
relatively large (calculated) population densities, resulting in penetration
factors close to one, and because industry?s ?upper bound? method tended to
suppress the effects of the generally lower penetration factors calculated
for the smaller urban centers in industry?s database. For example, we
calculated that the 3,312 urban centers in industry?s database would be
served by a total of 11,974 base stations based upon our ?best fit?
penetration curve, but that using a 100 percent penetration factor for all
urban centers increased this total by only 203 base stations (to 12,177) or
by about 1.7 percent.

3. PSD Received at Point S - Once the aggregate PSD radiated from each urban
center is calculated, equation [1] is used to calculate the contribution
from each visible urban center at the point S. This requires consideration
of the IMT- 2000 base stations vertical antenna gain pattern and elevation-
dependent environmental losses.

Appendix II: Replication of DOD?s and Industry?s IMT- 2000 Base Station
Interference Models

Page 46 GAO- 01- 795 Defense Spectrum Management

3.1 IMT- 2000 Base Station Antenna Gain Pattern - DOD?s model assumed that
IMT- 2000 base stations would radiate power isotropically. Thus, DOD assumed
that in calculating an urban center?s interference contribution using
equation [1], 0 = T G in all cases. Our reproduction of DOD?s model
incorporates this assumption. Industry?s model used an antenna gain pattern
based upon an ITU document 9 to calculate an elevation angle- dependent
antenna gain ) ( T G , which is described by equations (11a) and (11b) in
industry?s report. We incorporated a suitable form of these equations into
our reproduction of industry?s model.

3.2 IMT- 2000 Environmental Losses - DOD?s model assumed that 90 percent of
the power radiated from IMT- 2000 base stations would be lost to the
environment before reaching satellite orbital altitudes irrespective of the
local elevation angle; thus DOD assumed that in calculating an urban
center?s interference contribution using equation [1], 10 = E L dB in all
cases. Industry?s model assumed an elevation angle- dependent environmental
loss factor as follows:

10 ) ( = E L dB( for 0 20 ) [5A]

) 20 (60 ) (60 10 ) ( 0 0

0

ï¿½ = E L dB( for 0 0 60 20 < ) [5B]

0 ) ( = E L dB( for 0 60 > ) [5C] We incorporated a suitable form of
equation [5] into our reproduction of industry?s model. We noted that
neither DOD nor industry made conservative assumptions regarding
environmental losses. For example, during our limited review of the Air
Force?s IMT- 2000 interference model, we discussed environmental losses with
Aerospace Corporation representatives. The Air Force model assumed
environmental losses would be much less than either DOD or industry assumed.

9 Recommendation ITU- R F. 1336 with k = 0.

Appendix II: Replication of DOD?s and Industry?s IMT- 2000 Base Station
Interference Models

Page 47 GAO- 01- 795 Defense Spectrum Management

We were generally able to reproduce the value of the peak IMT- 2000
interference levels predicted by DOD?s and industry?s original models for
each orbital shell to within about 0.1 dB. In addition, the locations of
those peaks generally corresponded well with those predicted by DOD and
industry. It should be noted, however, that because DOD calculated PSD
values for every point on a 1- degree longitude by 1- degree latitude
lattice, whereas our PSD values were calculated over a 2- degree longitude
by 2- degree latitude lattice, some minor differences exist in the location
of the peak PSD value on the orbital shell. Tables 2 and 3 summarize these
results for DOD and industry, respectively.

Table 2: Comparison of DOD and Our Interference Results based upon Our
Replication of DOD?s Model

Satellite altitude (km) DOD peak PSD

(dBw/ Hz) Replication of DOD?s

peak PSD (dBW/ Hz) Difference

(dB)

250 - 161.5 (at +39 lat./+ 117 long.) - 161.5

(at +38 lat./+ 118 long.) + 0.0

833 - 166.6 (at +36 lat./+ 119 long.)

- 166.6 (at +36 lat./+ 120 long.)

+ 0.0 20,200 - 186.7

(at +72 lat./+ 48 long.) - 186.7

(at +72 lat./+ 48 long.) + 0.0

35,784 - 191.1 (at +73 lat./+ 16 long.)

- 191.0 (at +72 lat./+ 18 long.)

- 0.1

Table 3: Comparison of Industry and Our Interference Results based upon Our
Replication of Industry?s Model

Satellite altitude (km) Industry?s peak PSD

(dBw/ Hz) Replication of

Industry?s peak PSD (dBW/ Hz) Difference

(dB)

250 - 166.2 (at +52 lat./- 2 long.)

- 166.2 (at +52 lat./- 2 long.)

+ 0.0 833 - 173.2

(at +52 lat./+ 4 long.) - 173.2

(at +52 lat./+ 4 long.) + 0.0

20,200 - 191.9 (at +64 lat./- 6 long.); (at +64 lat./- 4 long.)

- 191.3 (at +64 lat./- 2 long.)

- 0.6 35,784 - 195.8

(at +54 lat./+ 8 long.); (at +56 lat./+ 4 long.); (at +62 lat./- 18 long.)

- 195.8 (at +56 lat./+ 4 long.)

+ 0.0

Results From the Replication of DOD?s and Industry?s IMT- 2000 Base Station
Interference Models

Appendix II: Replication of DOD?s and Industry?s IMT- 2000 Base Station
Interference Models

Page 48 GAO- 01- 795 Defense Spectrum Management

With respect to table 3, we did not attempt to resolve the -0.6 dB
discrepancy with industry representatives. As discussed above, we used a
slightly different approach to calculating the IMT- 2000 base station
penetration factor, which might account for part of this difference.
Moreover, we did not attempt to review, in detail, industry?s assumptions
regarding values used for fundamental physical constants, or rounding of
intermediate calculations. While additional work would likely have resolved
this discrepancy, the - 0.6 dB difference is not large enough to justify
that exercise.

Appendix III: Comments From the Department of Defense

Page 49 GAO- 01- 795 Defense Spectrum Management

Appendix III: Comments From the Department of Defense

Appendix III: Comments From the Department of Defense

Page 50 GAO- 01- 795 Defense Spectrum Management

Appendix III: Comments From the Department of Defense

Page 51 GAO- 01- 795 Defense Spectrum Management

Appendix III: Comments From the Department of Defense

Page 52 GAO- 01- 795 Defense Spectrum Management

Now on p. 24. Now on pp. 23- 24.

Now on p. 23.

Appendix III: Comments From the Department of Defense

Page 53 GAO- 01- 795 Defense Spectrum Management

Appendix III: Comments From the Department of Defense

Page 54 GAO- 01- 795 Defense Spectrum Management

Appendix IV: Comments From the Department of Commerce

Page 55 GAO- 01- 795 Defense Spectrum Management

Appendix IV: Comments From the Department of Commerce

Appendix V: Comments From the Federal Communications Commission

Page 56 GAO- 01- 795 Defense Spectrum Management

Appendix V: Comments From the Federal Communications Commission

(120017)

The first copy of each GAO report is free. Additional copies of reports are
$2 each. A check or money order should be made out to the Superintendent of
Documents. VISA and MasterCard credit cards are also accepted.

Orders for 100 or more copies to be mailed to a single address are
discounted 25 percent.

Orders by mail:

U. S. General Accounting Office P. O. Box 37050 Washington, DC 20013

Orders by visiting:

Room 1100 700 4 th St., NW (corner of 4 th and G Sts. NW) Washington, DC
20013

Orders by phone:

(202) 512- 6000 fax: (202) 512- 6061 TDD (202) 512- 2537

Each day, GAO issues a list of newly available reports and testimony. To
receive facsimile copies of the daily list or any list from the past 30
days, please call (202) 512- 6000 using a touchtone phone. A recorded menu
will provide information on how to obtain these lists.

Orders by Internet

For information on how to access GAO reports on the Internet, send an e-
mail message with ?info? in the body to:

Info@ www. gao. gov or visit GAO?s World Wide Web home page at: http:// www.
gao. gov

Contact one:

 Web site: http:// www. gao. gov/ fraudnet/ fraudnet. htm

 E- mail: fraudnet@ gao. gov

 1- 800- 424- 5454 (automated answering system) Ordering Information

To Report Fraud, Waste, and Abuse in Federal Programs
*** End of document. ***