B A L L I S T I C

M I S S I L E
D E F E N S E
Threats and Challenges

A Report by the American Physical Society

Panel on Public Affairs
January 2022
STUDY GROUP MEMBERS ABOUT APS & POPA
Frederick K. Lamb (Chair), Founded in 1899 to advance and diffuse the
University of Illinois knowledge of physics, the American Physical Society
Laura Grego (Co-Chair), Massachusetts (APS) is now the nation’s leading organization of
Institute of Technology and Union of physicists with approximately 55,000 members in
Concerned Scientists academia, national laboratories and industry. APS has
long played an active role in the federal government;
James D. Wells (Co-Chair), its members serve in Congress and have held
University of Michigan, Ann Arbor positions such as Science Advisor to the President of
David Barton, the United States, Director of the CIA, Director of the
Independent Consultant National Science Foundation and Secretary of Energy.
Philip E. Coyle, This report was overseen by the APS Panel on Public
Center for Arms Control and Affairs (POPA). POPA routinely produces reports on
Non-Proliferation* timely topics being debated in government so as to
inform the debate with the perspectives of physicists
Steve Fetter, working in the relevant issue areas.
University of Maryland
Alec Gallimore,
University of Michigan, Ann Arbor AUTHORSHIP
George N. Lewis, The American Physical Society has sole responsibility
Independent Consultant for the contents of this report, and the questions,
findings, and recommendations within.
Cynthia Nitta,
Lawrence Livermore National Laboratory
William Priedhorsky, ACKNOWLEDGEMENTS
Los Alamos National Laboratory
The Study Group is grateful to the many experts who
Jaganath Sankaran, were kind enough to brief the Study Group on issues
University of Texas at Austin related to the study or to answer our questions. We
Aric Tate, also wish to thank the technical and subject matter
University of Illinois experts who generously agreed to serve as external
reviewers and provided valuable comments and
Frank von Hippel, suggestions on the draft of the report they reviewed.
Princeton University
The Study Group wishes to express its appreciation
Institutions are listed for purposes of identification to the APS Panel on Public Affairs for commissioning
only. Any opinions, findings, conclusions, or this study, for providing advice as it developed, and
recommendations expressed in this report are those for reviewing and approving the final report. We also
of the Study Group members, and do not necessarily
represent the official views, opinions, or policies of their thank the APS Physics Policy Committee for its review
institutions. of the report and the APS Council for reviewing the
*Dr. Coyle passed away as the study was ending. report and approving its release.
We are especially grateful to the APS Office of
Governmental Affairs staff, including Chief External
Publication Date: January 2022 Affairs Officer Francis Slakey, Director of Government
Affairs Mark Elsesser, and Office of External Affairs
American Physical Society Administrator Janay Oliver, as well as APS staff
members Jorge Nicolás Hernández Charpak
This report is available under the terms of a and Mariam Mehter, for their advice and support
Creative Commons Attribution 4.0 International
License. Sharing and adapting the material for
throughout the study. We thank illustrator Janet Sinn-
any purpose, even commercial, does not require Hanlon for producing the figures in the report.
prior written permission. Further distribution Finally, we wish to express our deep gratitude for
of this work must provide appropriate credit, the unfailing courtesy, understanding, wisdom, and
provide a link to the license, and indicate if knowledge that our friend and colleague Dr. Phil
changes were made. For more information,
Coyle, who passed away during the final revision of
please visit the Creative Commons website.
the report, brought to this study and to our report.
Cover and report design by Ashley Mumford
Table of Contents

1. Introduction............................................................................................................3

2. North Korea’s ICBM Capabilities..........................................................................8

North Korea’s long-range ballistic missiles.................................................... 8
Liquid-propellant ICBMs ..................................................................... 8
Solid-propellant ICBMs ....................................................................... 9
Launch platforms for long-range ballistic missiles ........................... 9
North Korea’s nuclear weapons ...................................................................10
Missile warheads and penetration aids .......................................................10
Nuclear warheads for ICBMs ............................................................10
Countermeasures to missile defenses .............................................10
The North Korean ICBM capability we consider ........................................ 11

3. Challenges of Missile Defense ......................................................................... 12

Technical challenges .....................................................................................12
Challenges posed by the adversary’s response .........................................14
The challenge of obtaining and acting on independent evaluations........15

4. Midcourse Intercept Systems ........................................................................... 17

Appeal and challenges of midcourse intercept..........................................17
Overview ............................................................................................17
Passive countermeasures ..................................................................18
Attacking the defense as a countermeasure ...................................18
Multiple intercept attempts ..............................................................19
The Ground-based Midcourse Defense system .........................................20
Overview ............................................................................................20
Concept of operations ......................................................................22
Elements of the system .....................................................................22
Testing program ................................................................................23
 Midcourse countermeasures in flight tests ......................................24
Overall assessment ............................................................................25
Potential additional midcourse intercept layers: Aegis BMD & THAAD ......26
Wider implications of planned U.S. midcourse intercept systems ............27

5. Boost-Phase Intercept Systems ........................................................................ 28

Appeal and challenges of boost-phase intercept ......................................29
Key challenges of boost-phase intercept ........................................29
The reach-versus-time challenge .....................................................30
Kill vehicle and system requirements ..............................................35
Countermeasures ..............................................................................35
Using land- & sea-based rocket interceptors for boost-phase intercept ....36
Using drone-based rocket interceptors for boost-phase intercept ..........36
Using aircraft-based rocket interceptors for boost-phase intercept .........37
Using space-based rocket interceptors for boost-phase intercept ...........38
Required size of a space-based interceptor system .......................38
Cost of a space-based interceptor system ......................................40
Countermeasures to space-based interceptor systems .................42
Other disadvantages of space-based interceptor systems ............42
Using laser weapons for boost-phase intercept .........................................42

6. Closing Remarks ................................................................................................ 43

Biographical Sketches of Study Group Members .......................................... 45

Figures ................................................................................................................ 47

References .......................................................................................................... 48

2
1. INTRODUCTION

One of the most critical security challenges for Iraq might have, or a small accidental or unau-
humankind is the existence of nuclear weap- thorized launch by China or Russia [NMDA
ons. Nuclear-armed intercontinental ballistic 1999]. Today, Iraq and Iran have no nuclear
missiles (ICBMs) exacerbate this challenge by weapons, although there is concern that
making people vulnerable to sudden nuclear Iran might develop them in the future. North
attack—whether deliberate or mistaken—from Korea, which has tested both nuclear weap-
across the globe. The explosion of even a single ons and ICBMs capable of delivering them
nuclear warhead over a major U.S. city would [CRS 2021a; Kristensen 2021], has therefore
be an enormous disaster, potentially killing a become a primary focus of the current U.S.
million people and reducing 100 square miles ballistic missile defense (BMD) program.
to rubble [DOD 1977]. Multiple large nuclear
explosions over cities would be a catastrophe In 2002, the United States withdrew from the
for all humanity (see [OTA 1979] and the discus- U.S.-Russian ABM Treaty, which had been
sion of climatic effects in [NRC 1985]). designed to prevent missile defense efforts
from driving defense-offense arms race cycles
A natural reaction to such a threat is to between the two countries. Then, in 2016, the
consider the possibilities for intercepting U.S. Congress struck the word “limited” from
and disabling nuclear-armed ICBMs before its description of the threat the U.S. BMD
they reach their targets. The United States has program is charged with defending against
been pursuing the possibility of a defense [NDAA 2017, Sec. 1681], thereby opening the
against ballistic missiles for over 65 years. door to pursuing defenses against Russian
Missile defense efforts have so far cost Ameri- and Chinese ICBMs. Russia and China also
can taxpayers over $350 billion in 2020 dollars have missile defense programs [Baklitskiy
[BMD Expenditures 2021], most of which has 2021], although they currently have little stra-
been for intercepting ICBMs. However, as we tegic relevance to the United States.
explain below, no missile defense system thus
far developed has been shown to be effective An unusual aspect of any effort to defend
against realistic ICBM threats. against nuclear-armed ICBMs in flight is that
it provides decisive protection only once it
From 1972 to 2002, the Anti-Ballistic Missile is nearly perfect, because a successful attack
(ABM) Treaty permitted the United States by even one nuclear-armed ICBM would be
and the Soviet Union (later Russia) to deploy catastrophic, but its negative strategic and
only limited defenses against ICBMs, defined arms race implications are felt immediately.
as ballistic missiles with ranges greater than We recognize that a limited missile defense
5,500 km. Then the National Missile Defense capability may be perceived as having value
Act of 1999 restricted the United States to for deterring an attack on the United States
deploying only a system that could defend or its allies, facilitating a preventive war or a
against a limited ballistic missile attack, which pre-emptive attack by the United States, limit-
was understood to mean an attack using the ing the damage caused by a nuclear attack in
smaller number of less sophisticated missiles case of war, increasing the bargaining power
that a country such as North Korea, Iran, or of the United States vis-à-vis North Korea,

3
China, and Russia, or for other purposes. All technologies specifically designed to defeat
these considerations must be factored into current and future U.S. defenses against ballis-
national policy but are outside the scope of tic missiles, such as maneuvering warheads,
the present brief study. multiple independently targeted warheads,
and hypersonic glide weapons. They may also
In this report, we focus on questions that are include delivery systems designed to circum-
narrow enough to be answered with some vent current and future U.S. defenses against
confidence but have broad implications ballistic missiles, such as short-range ballistic
for programs and systems whose goal is to missiles launched from ships off U.S. coasts,
defend against ICBMs in flight. In particular, we nuclear weapons launched on fractional-or-
focus on the fundamental question of whether bit trajectories (sometimes referred to as
current and proposed systems intended fractional orbital bombardment systems or
to defend the United States against nucle- FOBS), nuclear-armed uncrewed underwa-
ar-armed ICBMs are now effective, or could in ter vehicles, or nuclear-armed cruise missiles.
the near future be made effective in prevent- Defending against these more numerous
ing the death and destruction that a success- and sophisticated threats is likely to be much
ful attack by North Korea on the United States more challenging than defending against
using such ICBMs would produce. As noted the numerically smaller and technologically
above, this is a primary concern of the current less sophisticated threat posed by the nucle-
U.S. missile defense program. In addressing ar-armed North Korean ICBM force that we
this question, we consider ICBMs that North focus on here.
Korea might have within the 15-year horizon
of this study. However, we do not consider A key purpose of this report is to explain why
multiple, maneuvering, or glider-like ICBM a defense against even the limited ICBM
warheads. Although North Korea has tested threat we consider is so technically challeng-
maneuvering warheads and a glider-like ing, and where the many technical difficulties
warhead on medium-range missiles and is lie. Our hope is that readers will come away
seeking to develop the capability to launch with realistic views of the current capabilities
multiple nuclear warheads on a single missile, of U.S. systems intended to defend against
it has not yet demonstrated these technolo- the nuclear-armed ICBMs North Korea may
gies on an ICBM. As we discuss in this report, have at present and an improved understand-
building a practical, effective defense against ing of the prospects for being able to defend
North Korean ICBMs that do not take advan- against the ICBMs North Korea might deploy
tage of any of these technologies is already within the next 15 years. In our view, despite
extremely challenging. We do not discuss some high-profile comments to the contrary
North Korea’s short-, medium-, and interme- [Panetta 2012; Trump 2019; Hyten 2020],
diate-range missiles, which would chiefly be the current capabilities are low and will likely
used in the Asia-Pacific region. continue to be low for the next 15 years.

We also do not consider missile defense To focus our report further, we consider what
systems intended to defend against the numer- would be required to defend against the
ically larger and technically more sophis- launch of a single ICBM from North Korea,
ticated current and future nuclear-armed or the salvo launch of 10 in rapid succession,
ICBM forces of Russia or China. These forces taking into account countermeasures North
are likely to include delivery systems that use Korea may be able to use to penetrate U.S.

4
defensive systems. While these are only two
of many possible attacks, considering them
reveals many of the technical challenges and
broader implications of any effort to defend
against nuclear-armed ICBMs.

Figure 1 illustrates three ICBM trajectories

from North Korea to the United States. The Boston
distance to Boston is about 11,000 km and an
ICBM would travel this distance in about 40
minutes.

In general, defense against an ICBM can be

Dallas
attempted during any of its three phases of
flight:
Los Angeles
• Boost phase. During its boost phase, the
ICBM’s rocket engines are burning, produc-
ing a bright exhaust plume as it lifts off and
gains altitude and speed. This phase lasts
three to five minutes for current ICBMs,
depending on their design.
Figure 1 View of Earth illustrating the trajectories
• Midcourse phase. The midcourse phase of ICBMs from North Korea to Los Angeles, Dallas,
begins when the engine of the missile’s and Boston. The trajectories shown are great circles
final stage has stopped burning. At that rather than the true trajectories, which would
point the rocket body’s final stage, one need to take into account the rotation of Earth.
or more warheads, and any other objects
that have been discarded or deployed by
the missile—such as deployment modules, The objective of a missile defense system is to
insulation, and other parts of the booster, disable the ICBM or its warhead during one of
or deliberate countermeasures to the these three phases of flight.
defensive system—begin moving along
The weapons currently being proposed to
ballistic trajectories in space. This phase
disable North Korean ICBMs during their
lasts approximately 30 to 40 minutes for
boost phase are airborne or space-based
ICBM trajectories from North Korea to the
rocket interceptors. The proposed airborne
continental United States.
interceptors would be based on long-dura-
• Terminal phase. The terminal phase begins tion, heavy-payload uncrewed aerial vehicles
once the warhead(s) and accompany- (“drones”) or aircraft positioned near or even
ing objects re-enter the atmosphere at an over North Korea, China, or Russia, close to
altitude of about 100 km, slowing due to the initial flight paths of North Korean ICBMs
air resistance as the warhead descends potentially heading toward the United States.
toward its target. This phase lasts less than
The current U.S. midcourse intercept systems
a minute.
are the Ground-based Midcourse Defense
(GMD), which currently has interceptors

5
 Boost Midcourse Phase Terminal
Phase (20–30 minutes) Phase
(2–5 minutes) (<1 minute)

Notional, not to scale.

1000
Altitude (approximate, km)

600

ICBM
200 X Burnout

Space
100
ICBM Ground-based Atmosphere
Launch Midcourse
Defense system

Aegis BMD
Boost-phase
interceptor THAAD

-
Intercontinental Ballistic Missile (ICBM) 5,500 10,000 km

Figure 2 Schematic portrayal of the layered missile defense system being proposed to defend the
United States against ICBMs launched from North Korea. An attempt can be made to intercept an
ICBM while its rocket engine is burning (its boost phase), during the flight of its warhead through the
vacuum of space (the midcourse phase), or after its warhead has re-entered the sensible atmosphere
(the terminal phase). Currently, the sole system deployed to defend the U.S. homeland from an inter-
continental ballistic missile attack is the Ground-based Midcourse Defense (GMD) system. To increase
the overall effectiveness of the system, in 2020 the Missile Defense Agency proposed the layered
system depicted here, in which intercept attempts by the GMD system would be followed by intercept
attempts by the Aegis regional midcourse defense system, and perhaps finally by a terminal defense
system based on an enhancement of the existing THAAD system. No boost-phase intercept system
currently exists. (Note that the vertical scale changes at the break in the axis shown on the left.)

deployed at Fort Greely, Alaska, and Vanden- developed to defend against the warheads
berg Air Force Base, California, and the Aegis of shorter-range ballistic missiles, but there
BMD system, which is currently being used are now proposals to upgrade it to attempt
to defend U.S. military installations and allied the much more challenging task of defending
territory but is now being proposed as an against much faster ICBM warheads.
additional midcourse intercept system to
protect U.S. territory. Figure 2 presents a pictorial overview of these
deployed and proposed system elements. We
The very short duration of the terminal phase discuss these elements in more detail below.
requires terminal interceptors to be deployed
very close to the area they are intended to The most recent U.S. Missile Defense Review
defend. The Army’s transportable Terminal [MDR 2019] calls for further development and
High-Altitude Area Defense (THAAD) was testing of all the missile defense elements

6
mentioned above and more; $10 billion or back centuries and, in the 20th century, by the
more has been allocated annually to this 1910–45 Japanese occupation and the exter-
program in recent years. While this could nally enforced division of the Korean Penin-
change as the current administration reviews sula at the end of World War II” [DIA 2021].
its nuclear and missile defense postures, The DIA report assesses that the primary moti-
current executive branch and congressio- vations that led Kim Jong Il to put the North on
nal guidance and budget allocations largely a path to a nuclear breakout in the mid-2000s
continue the direction of the existing program. were “apprehension about U.S. military inten-
tions after the 9/11 attacks and major [U.S.]
Although this report is primarily technical, operations in Afghanistan and Iraq, a contin-
it does discuss some of the wider implica- ually worsening military imbalance on the
tions of missile defenses, such as their likely [Korean] peninsula, and failure to obtain antic-
effects on the current U.S. offensive-defensive ipated energy assistance and other economic
nuclear competition with North Korea, China, concessions from international negotiations.”
and Russia. These effects include the incen- It concludes that the objectives of North
tives the deployment of defenses gives Russia Korea’s military are “to hold the United States
and China to develop and deploy additional at bay while preserving the capacity to inflict
nuclear-armed ICBMs and other, new offen- sufficient damage on the South, such that both
sive weapons as hedges against future break- countries have no choice but to respect the
throughs in U.S. missile defense capabilities North’s sovereignty and treat it as an equal.”
[Baklitskiy 2021; Cropsey 2021; Erästö 2021;
for a different perspective, see Roberts 2014; These assessments and conclusions suggest
Roberts 2020]. These developments appear that reducing the threat of North Korea’s
to be generating a new nuclear arms race to ICBMs requires careful analysis and responses
deploy more, and more sophisticated, offen- to all relevant dimensions of this problem,
sive and defensive weapons. including its political and diplomatic aspects
as well as its military dimensions. Ballis-
We do not consider many other important tic missile defense capabilities are just one
questions related to missile defenses, such component of this complex question.
as the appropriate level of funding for missile
defense relative to other priorities. The next two sections describe North Korea’s
current and possible near-term ICBMs and
Both U.S. government and nongovernmental nuclear warheads, and some of the key chal-
experts have assessed that a primary motiva- lenges that confront efforts to build a system
tion for North Korea’s nuclear weapons and that could defend against them once they are
missile programs is to deter other countries launched. The two main sections of the report
from attempting to change North Korea’s then follow. The first describes midcourse
ruling regime by force [DOD 2020a; Bennett intercept systems, including the GMD system
2021; CRS 2021a]. According to the October and potential contributions of the regional
2021 report by the U.S. Defense Intelligence Aegis BMD and THAAD systems, while the
Agency (DIA), “North Korea’s perception that second describes boost-phase intercept
the outside world is inherently hostile drives systems, including possible land-, sea-, air-,
the North’s security strategy and pursuit of and space-based rocket interceptors and
specific military developments. This percep- aircraft-based missiles and laser weapons. The
tion is informed by a history of invasion and report ends with some closing thoughts.
subjugation by stronger powers stretching

7
2. NORTH KOREA’S ICBM CAPABILITIES

This report considers the threat posed to relies on nuclear deterrence to prevent poten-
the United States by North Korea’s potential tial Russian or Chinese nuclear attacks employ-
deployment of a limited but significant number ing their large and technically sophisticated
of nuclear-armed ICBMs within the 15-year intercontinental missile systems,” and with
time horizon of this study. Focusing on this the fiscal year 2020 (FY20) National Defense
threat is consistent with previous U.S. missile Authorization Act [NDAA 2020], which says
defense policies and the 2019 Missile Defense (Sec. 1681), “It is the policy of the United States
Review [MDR 2019], which stated (p. IX) that to rely on nuclear deterrence to address more
U.S. missile defense capabilities are sized to sophisticated and larger quantity near-peer
defend the U.S. homeland against the limited intercontinental ballistic missile threats.”
offensive missile threats posed by states such
as North Korea. It is also consistent with the North Korea’s long-range ballistic missiles
assessment (see below) that North Korea is Liquid-propellant ICBMs. North Korea has
unlikely to deploy intercontinental-range successfully tested two types of liquid-propel-
submarine-based ballistic missiles that could lant ICBMs capable of striking part or all of the
strike the U.S. homeland within the time hori- continental United States. It tested its first ICBM,
zon of this study. The United States relies on the Hwasong-14 (U.S. designation KN-20),
nuclear deterrence to deter attacks from any on July 4 and July 28, 2017 [CSIS 2021]. This
source, including North Korea [MDR 2019, IX]. missile is estimated to have a full burn time of
In contrast to North Korea, Iran does not have about 375 seconds and a range of more than
nuclear weapons or ICBMs and is currently 10,000 km [CSIS 2021; DIA 2021, 24].
observing a self-imposed moratorium on test- North Korea tested a longer-range ICBM, the
ing ballistic missiles with ranges greater than Hwasong-15 (KN-22), on November 28, 2017.
2,000  km [Einhorn 2019]. While Iran likely This missile is estimated to have a full burn time
could produce a nuclear weapon and an of 290 seconds, a maximum range of about
ICBM within the 15-year time horizon of this 13,000  km  [Panda 2017; Dominguez 2019;
study [Belk 2018; Cordesman 2019; Einhorn Bennett 2021; CSIS 2021; DIA 2021, 24], and
2019; Elleman 2021a], we do not explicitly the ability to carry “penetration aids” (devices
consider this possibility. However, much of designed to enable the warheads to pene-
our discussion would be relevant to assessing trate defensive systems). It shares some design
the potential for the United States to defend features with the early Soviet UR-100/SS-11
itself against nuclear-armed Iranian ICBMs. missile and has an engine based on the Soviet
We also do not discuss the capabilities of RD-250, but its engine uses two gimbaled
missile defense systems to defend against the main chambers for steering, rather than four
ICBM or submarine-launched ballistic missile small vernier engines. Changes such as these
(SLBM) forces of Russia or China. This is consis- suggest modest indigenous North Korean
tent with previous U.S. missile defense policies missile engineering ability [Schiller 2019].
and the 2019 Missile Defense Review [MDR In its October 2020 military parade, North
2019], which states (p. 8), “The United States Korea displayed a model of a new liquid-pro-

8
pellant ICBM, the Hwasong-16 (KN-27), which 2004] ICBM programs, some of the technical
is much larger than previous North Korean challenges involved in scaling up solid rocket
ICBMs [Hansen 2020; Van Diepen 2020; motors to the sizes required may already have
Varner 2020]. A functioning missile of this size been overcome by North Korea in producing
would be able to carry multiple warheads and its current motors, but other challenges would
penetration aids. need to be surmounted to produce motors
large enough for ICBMs [Schiller 2019].
Solid-propellant ICBMs. Boost-phase defense
against solid-propellant ICBMs is more chal- It appears unlikely that North Korea can
lenging because they can be launched with develop and deploy solid-propellant SLBMs
less preparation time and have substan- with intercontinental ranges and the subma-
tially shorter burn times than liquid-propel- rines needed to carry and launch them within
lant ICBMs. North Korea has not yet built or the 15-year time horizon of this study [Kim
tested a solid-propellant ICBM, but it does 2021a; Kim 2021c]. However, it might be able
have a solid-propellant missile development to develop and deploy a solid-propellant
program [Smith 2020]. ICBM, depending on the foreign assistance it
receives and whether its current solid rocket
North Korea successfully tested a medi- motors are being manufactured within North
um-range solid-propellant ballistic missile, the Korea or elsewhere [Smith 2020].
Pukkŭksŏng-1 (KN-11), with a range of more
than 1,000 km, in August 2016 [CSIS 2021; Launch platforms for long-range ballistic
DIA 2021, 24]. This missile was presented missiles. In recent congressional testimony,
as intended to be launched from a subma- Gen. Scott Berrier, Director of the Defense
rine. North Korea has developed addi- Intelligence Agency, noted [Berrier 2021],
tional missiles in this series, including the “The October 2020 parade also featured eight
Pukkŭksŏng-2 (KN-15) [Elleman 2017; CSIS road-mobile ICBM launchers, the most North
2021], Pukkŭksŏng-3 (KN-26) [CSIS 2021], and Korea has ever displayed.” This suggests that
possibly a Pukkŭksŏng-4 [Van Diepen 2020; North Korea’s ability to deploy ICBMs may no
Varner 2020]. In January 2021, North Korea longer be constrained by its apparent former
displayed a model of a new, larger, solid-pro- inability to produce adequate transporter-erec-
pellant missile, the Pukkŭksŏng-5 (KN-15), tor-launchers, and that it may now be able to
which appears similar in size and shape to deploy 10 or more ICBMs on mobile launchers.
the U.S. Polaris A3 SLBM [Elleman 2021b].
It continues to test these solid-propellant North Korea is also working to develop the
missiles, with the most recent test occurring ability to launch missiles from trains [Van
on October 18, 2021 [Choe 2021b]. Although Diepen 2021a; Smith 2021], which if success-
advertised as SLBMs, these missiles could also ful could allow it to launch heavier missiles
be launched from land. such as ICBMs from a larger number and a
wider range of locations.
The Pukkŭksŏng missiles are estimated to
have ranges of 1,000 - 2,000 km [CSIS 2021]. Finally, North Korea is continuing its program
To produce a solid-propellant ballistic missile to develop small, diesel-powered subma-
with intercontinental range, North Korea rines that could carry and launch up to three
would have to develop significantly larger Pukkŭksŏng ballistic missiles [Sutton 2019;
solid rocket motors. Based on the history of Cha 2020; Nikkei 2021]. However, these
the U.S. [Caveny 2003] and Soviet [Podvig submarines are vulnerable because they are

9
slow, cannot remain submerged for long peri- light enough to be carried by the Hwasong-15
ods, and have relatively loud acoustic signa- and a re-entry vehicle robust enough to
tures [Kim 2021a]. Consequently, we do not survive the rigors of launch and re-entry into
consider them further in this report. the atmosphere after a full-range ICBM flight.
Two reports requested by the U.S. govern-
North Korea’s nuclear weapons ment assessed that as of 2017 North Korea
North Korea apparently already had a nuclear had developed a nuclear warhead that could
weapons program and had fabricated two or be mounted on its ICBMs [Bennett 2021;
three nuclear devices by the late 1990s (see CRS 2021a]. A careful independent assess-
[Kristensen 2021] for a detailed review of what ment [Wright 2017] concluded that “North
is known and surmised about North Korea’s Korea has not yet demonstrated a working
nuclear weapons program). It has so far tested re-entry vehicle (RV) on a trajectory that its
six nuclear devices, two with yields estimated missiles would fly if used against the United
to be in the range of 10 to 15 kilotons and one States. However, there doesn’t appear to be
with a much larger yield estimated to be in the a technical barrier to building a working RV,
range of 140 to 250 kilotons. Due to the opac- and doing so is not likely to be a significant
ity of North Korea’s nuclear program, U.S. and challenge compared to what North Korea has
international officials, experts, and agencies already accomplished in its missile program.
have had difficulty assessing the program’s … While the United States put very significant
purposes and accomplishments. resources into developing sophisticated RVs
and heatshields … that effort was to develop
Knowledgeable observers estimate that North highly accurate missiles and is not indica-
Korea might have produced enough fissile tive of the effort required by North Korea to
material (plutonium and highly enriched develop an adequate RV to deliver a nuclear
uranium) to construct 20 to 60 nuclear weap- weapon to a city.”
ons but may have assembled fewer [DOD
2020a; CRS 2021a; Hecker 2021; Kristensen Countermeasures to missile defenses. In
2021]. Most of these weapons would likely be 1999, the U.S. national intelligence commu-
single-stage fission weapons with possible nity assessed that Russia and China have both
yields of 10 to 20 kilotons with at most only developed numerous countermeasures to
a few thermonuclear weapons  [Kristensen missile defense and probably are willing to sell
2021]. Some have estimated that North the requisite technologies, and that emerging
Korea may be able to produce enough fissile missile states such as North Korea would likely
material to construct about 3 to 7 additional have developed countermeasures by the time
weapons per year [DOD 2020a; CRS 2021a; they flight-test their missiles [NIC 1999].
Kristensen 2021]. If so, North Korea could For some years, North Korea has been
produce enough fissile material to make 50 developing technologies designed to give
to 100 additional nuclear weapons within the its warheads greater ability to penetrate
15-year time horizon of this study. missile defense systems. At least two of the
Missile warheads and penetration aids short-range missiles it introduced in 2019,
the KN-23 and KN-24, are reported to have
Nuclear warheads for ICBMs. North Korea warheads that can perform low-altitude
is likely to have already developed, or could maneuvers, making them harder to intercept
develop soon, a nuclear weapon small and [Choe 2021a]. In October 2021, North Korea

10
test-launched what it called its first hyper- and, given the assessments cited above, will
sonic missile, the Hwasong-8, which appears have enough nuclear weapons to mount them
to have a boost-glide warhead [Choe 2021a; on its ICBMs. North Korea has not yet demon-
Gallo 2021; Panda 2021;Trevithick 2021]. In strated a working re-entry vehicle on a trajec-
an official statement, the South Korean Joint tory its missiles would fly if used against the
Chiefs of Staff said that this missile “appears United States, but there appears to be no
to be at an early stage of development that technical barrier to its building them. The
would require considerable time for actual accuracy of these missiles is likely to be low,
deployment” [Choi 2021; see also Van and they would therefore probably be used
Diepen 2021b]. In January 2022, North Korea against relatively large targets, such as cities,
launched two different, apparently improved rather than against hardened military targets.
hypersonic boost-glide vehicles that could
be advanced MaRVs [Rogoway 2022; Smith The U.S. intelligence community has assessed
2022]. We assess that North Korea has that North Korea has likely developed counter-
devoted substantial efforts to developing measures to missile defenses. It is equipping
countermeasures to missile defenses and is its shorter-range missiles with maneuvering
continuing to do so. re-entry vehicles and is actively working on
more advanced countermeasures, such as
North Korean ICBM capability we consider a possible glider-like warhead. It has not yet
demonstrated these countermeasures in tests
Based on the information just described, of long-range missiles.
North Korea probably has a few liquid-pro-
pellant ICBMs that could strike the continen- Based on these assessments of North Korea’s
tal United States and may be able to deploy current nuclear-armed ICBM capabilities and
10 or more within the 15-year time horizon of those it may be able to develop within the
this study. Publicly available information indi- 15-year time horizon of this report, the follow-
cates that it probably has transporter-erec- ing chapters focus on the performance that a
tor-launchers for these missiles and is working missile defense system would need to have
on being able to launch large missiles from to successfully defend the continental United
trains. North Korea is developing solid-pro- States against the baseline threat represented
pellant missiles and might be able to develop by the launch of a single liquid-propellant
and deploy a few solid-propellant ICBMs ICBM like the Hwasong-15 or a salvo launch of
within 15 years. The reliability of these long- 10 such ICBMs at intervals of less than a minute.
range missiles has not been demonstrated. Although it would be challenging for North
Korea to deploy solid-propellant ICBMs within
North Korea has probably assembled several the 15-year time horizon of this study, it might
nuclear weapons and may have several dozen do so, and the consequences for any boost-
within the 15-year time horizon of this study. phase defense system would be profound. We
According to the assessments cited above, therefore consider this possibility in our report.
most are probably fission devices with yields As we show, the missile defense systems that
in the 10 - 15 kiloton range, but a few may be would be needed to defend against these
thermonuclear weapons with yields of about threats are technically very challenging and
200 kilotons. Numerous sources assess that illustrate the difficulty of providing decisive
North Korea has developed nuclear devices protection against even limited threats.
small enough to be launched by its ICBMs

11
3. CHALLENGES OF MISSILE DEFENSE

Intercepting even a single nuclear-armed tive range of 4 - 70 km [Lister 2021]. They do
intercontinental-range ballistic missile or not strike the incoming missiles but instead try
its warhead(s) in flight under the conditions to approach them and then explode, sending
expected during a nuclear attack is extremely out shrapnel that can disable the home-made
challenging. The ability of any missile defense rockets if the interceptor is approaching the
system to do this reliably has not been rocket from the right direction and gets close
demonstrated. enough [Postol 2014]. The Iron Dome system
has been greatly improved over the decade
Here we briefly mention some of the import- it has been in use. It now engages about 50%
ant challenges faced by any program to of the rockets launched against the area it is
develop and deploy an effective missile defending and is claimed to destroy about
defense system. These include technical 80% - 90% of the rockets it engages [Bartels
challenges and challenges created by the 2017; Hambling 2021; Lister 2021].
adversary’s ability to respond to defensive
measures. We also call attention to the diffi- The U.S. Patriot system was originally designed
culties encountered in using the results to defend against aircraft, but at the outset of
of independent evaluations effectively to the 1991 Gulf War it was rushed to the Gulf to
remedy problems identified in this large and try to defend the Israeli population and U.S.
complex defense program. military forces against attacks by Iraq’s Al-Hus-
sein missiles, a variant of the Scud missile with
Technical challenges. The argument is some- a range of about 600 km. But the Patriot system
times made that missile defense must be feasi- almost completely failed to do this. A subse-
ble because of the reputed successes of Isra- quent investigation by the House Committee
el’s “Iron Dome” system and the U.S. Patriot on Government Operations found, “There
system. But the challenges faced by these is little evidence to prove that the Patriot hit
systems are far less than those confronting any more than a few Scud missiles launched by
system attempting to defend against ICBMs. Iraq during the Gulf War” and added, “There
Moreover, neither the Iron Dome system nor are some doubts about even these engage-
the Patriot system is fully successful against ments” [Congress 1992]. (For further details,
the much-less-capable missiles it is designed see [Lewis 1993; Sullivan 1999].)
to defend against.
More recently, the United States supplied
The Iron Dome system was developed about Patriot Advanced Capability-3 systems to
a decade ago to defend against rockets, artil- Saudi Arabia to help it defend against missiles
lery, mortar shells, and simple, very short- launched by Houthi forces. On November 4,
range, highly inaccurate home-made rock- 2017, Houthis attacked the airport in the Saudi
ets that travel at speeds of only about 1 km/s capital, Riyadh, using a Burqan-2 [Williams
over distances of only about 7 - 70 km and 2020], a variant of the Scud with a reported
carry warheads with an explosive power of range of about 1,000 km [Savelsberg 2018].
about 10 kg of TNT [Bartels 2017; Hambling According to evidence collected during and
2021]. Iron Dome interceptors have an effec- after the attack, the relevant Patriot defen-

12
sive battery fired five interceptors at the The boost phases of current ICBMs last three
missile, but its warhead flew unimpeded over to five minutes, depending on their design.
the interceptors and detonated on Riyadh’s Hence, as will be explained in the boost-phase
airport, indicating that the defense failed intercept section, for a land-, sea-, or air-based
when confronting a missile much less capable interceptor rocket to intercept an ICBM during
than an ICBM [Fisher 2017]. its boost phase, the interceptor must typically
be based within about 500 km of the intended
The interceptors of Israel’s Iron Dome system intercept point, have a speed of 5  km/s or
and the interceptors of the U.S. Patriot more, and be fired less than a minute after the
system are both designed to explode near launch of a potentially threatening missile has
the warheads of missiles traveling within the been detected. Interceptor bases and aircraft
atmosphere, and these systems therefore must be positioned 100 to 200 km from the
cannot be fooled by lightweight decoys. In borders of potentially hostile countries, or, in
contrast, U.S. GMD interceptors must strike the case of sea-based interceptors, at least
directly the warheads of ICBMs while they are 100 km from the coasts of potentially hostile
traveling far above the atmosphere, where countries, so that the ships that are carrying
the GMD system could be fooled by light- the interceptors are beyond the horizons of
weight decoys and other penetration aids land-based radars and have adequate room
(see below). Moreover, the warheads the for maneuvering (see [APS 2003, S66]). As
GMD system would have to engage would discussed below, these requirements severely
be traversing distances of 12,000 km or more restrict the ability of a system of land-, sea-, or
at speeds of more than 7  km/s, distances air-based rockets to intercept an ICBM during
100 times greater and speeds seven times its boost phase. If a large enough number
faster than the missiles engaged by Israel’s of rocket interceptors were instead placed
Iron Dome, and distances 10 times greater in appropriate low-Earth orbits, a sufficient
and speeds more than two times faster than number would be within range of any attack-
the warhead the Patriot system missed. If the ing ICBM during its boost phase to attempt an
GMD system misses the nuclear warhead intercept. But as discussed below, a constel-
it is seeking to destroy, the warhead could lation of many hundreds of interceptors in
explode on its target with a power a million low-Earth orbit would be required for one to
times greater than the warheads that the Iron be within range at all times to defend against
Dome and Patriot systems sometimes miss, even a single ICBM launched from a single
utterly destroying its target and the surround- site. As also discussed below, there are a vari-
ing area. ety of potentially effective countermeasures
For systems intended to defend against against boost-phase intercept, such as launch-
ICBMs, the brevity of the boost and re-entry ing several ICBMs nearly simultaneously
phases of these missiles and the lack of air (a “salvo launch”) or  programming evasive
resistance during the midcourse phase pose maneuvers by the ICBM.
severe technical challenges for the defense. The midcourse phase, during which nuclear
Moreover, “to be credible and effective, a warheads follow ballistic trajectories, lasts
ballistic missile defense system must be about 30 to 40 minutes, but the absence of
robust even if any of its elements fail to work air drag during this phase means that launch
as planned” (see [NRC 2012], Major Finding debris, such as spent upper stages, deployment
6, S-9). modules or attitude control modules, separa-

13
tion debris, debris from unburned fuel, insula- Challenges posed by the adversary’s
tion, and other parts of the booster, as well as response. Unlike civilian research and devel-
deliberately generated missile fragments, light- opment programs, which typically address
weight decoys, and other penetration aids, will fixed challenges, a missile defense program
all follow the same trajectory as a warhead. This confronts intelligent and adaptable human
makes it difficult for the defense to discriminate adversaries who can devise approaches to
the warhead from other objects in this “threat disable, penetrate, or circumvent the defen-
cloud,” so that it can target the warhead. The sive system. This can result in a costly arms
radar and infrared sensors required for track- race. Which side holds the advantage at any
ing, discrimination, and homing are vulnerable particular moment depends on the relative
to the effects of high-altitude nuclear detona- costs of the defensive system, the offensive
tions, which may be preplanned or result from a system adaptations required to defeat it, and
successful intercept. the resources each side is prepared to devote
to the competition.
The terminal phase, during which the nuclear
warhead re-enters the atmosphere, lasts only During the Cold War, the United States and
about a minute. As a result, only very high- the Soviet Union each deployed more than
speed rocket interceptors launched from 10,000 megaton-class strategic nuclear
bases close to the warhead’s target could warheads [Kristensen 2013]. A number of
reach and destroy a warhead during the factors contributed to the deployment of
terminal phase of its flight before it detonates. such irrationally large forces, but an import-
Furthermore, lightweight decoys would be ant one was the concern that nuclear-armed
stripped away by the atmosphere only during ballistic missiles might be countered, at least
the final 10 seconds or so before the warhead in part, by defensive systems. Because it takes
explodes. Terminal-phase defenses can there- a decade or more to develop and deploy
fore potentially defend only limited areas, major weapons systems and designers hope
such as a metropolitan area or a critical mili- they will be able to cope with the evolving
tary facility or command post. They are also situation for at least a decade after they are
vulnerable to the blinding effects of nuclear deployed, it is necessary to project the quan-
explosions in the atmosphere. titative and qualitative evolution of weap-
ons systems 20 years or more into the future.
Given all these challenges to ballistic missile These projections are, of course, uncertain,
defense, it is easy to understand why, when and because “it is better to be overprepared
engineers have been under intense polit- than underprepared,” there is a tendency for
ical pressure to deploy a system, the United planners to make worst-case assumptions,
States has repeatedly started costly programs which accelerate the defense-offense arms
that proved unable to deal with key technical race cycle.
challenges and were eventually abandoned
as their inadequacies became apparent. As The open-ended nature of the current U.S.
noted in the Introduction, the United States missile defense program has stimulated anxi-
has spent more than $350 billion in 2020 ety in both Moscow and Beijing. President
dollars [BMD Expenditures 2021] since 1957 Vladimir Putin has announced a variety of new
on research and development and deploy- nuclear-weapon delivery systems designed to
ment of ballistic missile defense systems, counter U.S. missile defenses. These include
none of which have proven effective. hypersonic boost-glide re-entry vehicles; the

14
Sarmat, a new, larger ICBM capable of carry- requirements to “fly before you buy” and to
ing many warheads and a wide variety of achieve positive evaluations by DOD’s Direc-
devices to aid its warheads in penetrating U.S. tor for Operational Test and Evaluation of their
missile defense systems; the Poseidon long- effectiveness under battlefield conditions.
range, nuclear-powered uncrewed underwa-
ter vehicle; and the Burevestnik nuclear-pow- One way to ensure that the missile defense
ered long-range cruise missile. program does not commit itself to ineffec-
tive or impractical approaches is to obtain
As for China, the DOD assesses that “The PLA independent reviews of all missile defense
[China’s People’s Liberation Army] justifies approaches and then act on them. For more
developing a range of technologies China than two decades, the U.S. missile defense
perceives are necessary to counter U.S. and program has solicited or been given reviews
other countries’ ballistic missile defense and reports that have pointed to serious prob-
systems, including MaRV [maneuvering reen- lems with the program. For example, in 1998,
try vehicles], MIRVs [multiple independent a panel commissioned by the Ballistic Missile
reentry vehicles], decoys, chaff, jamming, ther- Defense Organization and led by General
mal shielding, and hypersonic glide vehicles” Larry Welch found that the program was in a
[DOD 2019]. In summer 2021, China report- “rush to failure” because it lacked coherence
edly tested a system that launched a maneu- and a realistic plan. The panel recommended
vering glide vehicle onto an orbital trajectory that the program be fundamentally restruc-
[Rogoway 2021]. And China now appears to tured [Cerniello 1998; Boese 1999].
be building hundreds of new silos that could
hold ICBMs [Warrick 2021]. In 2010, Congress instructed the Secre-
tary of Defense to arrange for the JASON
The challenge of obtaining and acting on Defense Advisory Panel to study the discrim-
independent evaluations. It is important to ination capabilities and limitations of the
ensure that the missile defense program does U.S. ballistic missile defense system [NDAA
not commit itself to technical approaches 2010, Sec. 237]. Seven years later the Missile
that are impractical or easy to defeat. One Defense Agency (MDA) released an unclassi-
reason so much money has been spent on fied summary of the JASON report [JASON
U.S. ballistic missile defense efforts with little 2010]. Among its recommendations were
to show for it is that many of these efforts that “MDA should consider adjusting its
have been initiated in response to presiden- priorities to establish alliances with U.S.
tial advocacy, highly charged political argu- government-sponsored laboratories and
ments, or the perceived urgency of near- academic groups. These bodies [could be
term threats [Mosher 2000]. “In this climate, given] full inside knowledge of relevant MDA
ideas and programs are not fully conceived or programs and funding to carry out challeng-
vetted by the Pentagon bureaucracy and the ing reviews and simulations as well as to
budget process before they are pushed into propose alternative concepts. When justi-
the spotlight, contributing to poor program fied and with the cooperation and support
design, inaccurate initial cost estimates, and of MDA, these bodies should be involved
subsequent increases” [Mosher 2000]. As a in testing programs. Their role would be to
result, missile defense programs have often give independent and authoritative critical
neglected the difficulties and risks involved reviews of MDA programs; to formulate and
and bypassed normal safeguards, such as the simulate alternative concepts and strategies;

15
and to supply Red Team challenges to the material accompanying them in the vacuum
missile defense system.” of space. This discrimination is not the only
challenge for midcourse defense, but it is
In 2011, the Defense Science Board warned the most formidable one, and the midcourse
that “successful operations [sic] of [the discrimination problem must be addressed
system’s] components is predicated on an far more seriously if reasonable confidence
ability to discriminate (in the exo atmosphere) is to be achieved” (p. 10). It went on to say,
the missile warhead(s) from other pieces of “The midcourse discrimination problem must
the offensive missile complex, such as rocket be addressed far more seriously if reason-
bodies, miscellaneous hardware, and inten- able confidence is to be achieved” (p. 11). In
tional countermeasures. The importance of conclusion, the National Academies report
achieving reliable midcourse discrimination found that “the current GMD system has been
cannot be overemphasized” [DSB 2011]. developed in an environment of limited objec-
In 2012, Congress mandated a comprehen- tives (e.g., dealing with an early-generation
sive, independent review of the U.S. missile North Korean threat of very limited numbers
defense program by the National Acade- and capability) and under conditions where
mies [NRC 2012]. The 2012 National Acade- a high value was placed on getting some
mies report found that the GMD system “lacks defense fielded as quickly as possible, even if
fundamental features long known to maxi- its capability was limited and the system less
mize the effectiveness of a midcourse hit-to- than fully tested” (p. 13).
kill defense capability against even limited As we explain in the following chapters, some
threats.” The report stated: “The hard fact is of the challenging problems with the missile
that no practical missile defense system can defense program that were identified in the
avoid the need for midcourse discrimination— reports quoted above and in other reports
that is, the requirement to identify the actual have been addressed, but they have not been
threat objects (warheads) amid the cloud of solved.

16
4. MIDCOURSE INTERCEPT SYSTEMS

The United States has for many decades emies study concluded that “the GMD inter-
been pursuing defensive systems to intercept ceptors, architecture, and doctrine have short-
warheads in midcourse. Currently, the sole comings that limit their effectiveness against
system deployed to defend the U.S. home- even modestly improved threats and threats
land against an ICBM attack is the Ground- from countries other than North Korea” and
based Midcourse Defense (GMD) system. deemed the system “deficient with respect to
To increase the overall effectiveness of the all its fundamental precepts of a cost-effective
system, in 2020 the Missile Defense Agency defense” [NRC 2012].
proposed a “layered” approach in which
attempts to intercept ICBM warheads during We now provide an overview of midcourse
the midcourse phase using the GMD system intercept systems, including potential counter-
would be followed by further attempts to measures and their possible remedies, and the
intercept them using two systems not origi- three elements of the layered approach that is
nally designed for defending against ICBMs: currently being proposed.
the Navy’s Aegis BMD system during the Appeal and challenges of midcourse
midcourse phase and, perhaps finally, a intercept
system based on the Army’s THAAD system
during the terminal phase (see Figure 2). Overview. The midcourse phase of flight,
which begins when the ICBM’s final boost
The development of a U.S. homeland missile stage has burned out and it and the missile’s
defense has been contentious politically and warhead(s) have separated and are moving
difficult technically. Independent assess- ballistically above Earth’s atmosphere (see
ments are routinely commissioned to report Figure 2), presents both advantages and
on these efforts and provide public informa- special challenges for the defense. While in
tion on the challenges and prospects of the the past some midcourse intercept systems
U.S. midcourse intercept systems. Since 2002, were designed to use nuclear weapons to
Congress has mandated that the Govern- destroy incoming nuclear warheads, today’s
ment Accountability Office (GAO) produce systems seek to disable or destroy warheads
annual reports on the Missile Defense Agen- by firing an interceptor with a kill vehicle that
cy’s progress toward its acquisition goals, and will home in on and collide with them at a
the Defense Department’s Director of Oper- velocity high enough to cause them to fail.
ational Test and Evaluation issues annual
reports on the status of the missile defense For a warhead launched from North Korea to
test programs. Congress has also commis- the continental United States, the midcourse
sioned studies such as the 2012 study by phase lasts 30 to 40 minutes, long enough
the National Academies [NRC 2012], which that more than one intercept attempt may
assessed the GMD system. As discussed be possible. But the warhead is only about a
below, these reports paint a picture of a meter in length and can appear to radar and
program beset by poor management and infrared sensors as similar to the final stage
poor congressional oversight that struggles and other objects that have been discarded or
to make progress. The 2012 National Acad- deployed by the missile. Since these objects

17
are traveling in a near-vacuum, relatively China’s programs to develop countermea-
simple, lightweight decoys would follow the sures against ballistic missile defenses were
same trajectory as the warhead and could decades old, suggested that these countries
therefore confuse or overwhelm the defense. were probably willing to sell the technolo-
gies, and concluded that emerging missile
Passive countermeasures. To be successful, a states would likely have developed their own
midcourse intercept system must adequately countermeasures—based, for example, on
address the discrimination problem—identify- radar-absorbing materials, booster fragmen-
ing the nuclear warheads in the presence of tation and chaff, jammers, and simple balloon
other objects, such as the rocket’s final stage, decoys—by the time they flight-tested ICBMs
possibly deliberately broken into pieces, and [NIC 1999]. North Korea has demonstrated
other intentional penetration aids, such as a number of relevant technologies, includ-
radar-interfering chaff or decoys, about which ing the capability to deliberately break up a
the defense is unlikely to have detailed prior rocket stage, which if applied to the final stage
information. of an ICBM could create debris with radar
Decoys, such as aluminized mylar balloons, cross-sections similar to that of the re-entry
can be built to effectively mimic the radar, vehicle [Talmadge 2016].
infrared, and visible signatures the warhead In its tests of shorter-range missiles, North
presents to the defense’s sensors [Sessler Korea has demonstrated the ability to launch
2000]. Many such lightweight decoys could multiple missiles simultaneously and to deploy
be deployed with the warhead. The defense a maneuvering re-entry vehicle, indicating
would need to engage all objects that could investment in strategies to defeat missile
be warheads, potentially depleting its inven- defenses by saturating or evading them [UN
tory of interceptors. 2017, Item 12; Gallo 2021]. Some techniques,
Instead of building lookalike decoys, the such as the use of lookalike decoys, might
adversary could disperse objects with a range need to be flight-tested to provide assurance
of radar cross sections, apparent tempera- that they work, while others, such as anti-sim-
tures, and flight characteristics by altering ulation balloons (balloons that enclose
their shapes, coatings, and moments of iner- warheads to camouflage them), might be
tia (which affect their in-flight movement). tested adequately unobserved in ground
The adversary could also alter the observable facilities.
characteristics of the warhead or enclose it in a Attacking the defense as a countermeasure.
balloon large enough to make it difficult for the Rather than confusing the defensive system’s
interceptor’s kill vehicle to strike the enclosed sensors, an adversary could instead attack or
warhead directly enough to disable it. interfere with them. Long-range midcourse
While the details of which countermeasure intercept of warheads depends on a geograph-
strategies North Korea and other states have ically spread chain of sensors, primarily radars,
developed are not in the public domain, the for tracking and discrimination. Continuous
physics and engineering of the techniques observation of the threat cloud is important
involved are well established, and effec- both to prevent tracking errors from growing
tive countermeasures are likely to be widely and to attempt to identify the warhead within
available. In 1999, the U.S. national intelli- the threat cloud. An adversary could try to
gence community assessed that Russia and disable key sensors, especially forward-based

18
radars that are within the reach of short- and uation and be better able to track objects
intermediate-range missiles. in the threat cloud [Canavan 2003, Fig. D.1].
However, fluctuations in the radar signatures
The adversary could also confound sensors of the warhead and other objects would make
without attacking them directly by creat- discrimination significantly more difficult.
ing radar and infrared blackout effects with
high-altitude nuclear detonations [Garwin Less well studied are the high and spatially
1968]. Incoming warheads could be designed variable infrared backgrounds that nuclear
to detonate before an interceptor reaches detonations would produce over similarly
them, using the long-established technol- large areas. The infrared homing sensors of
ogy of proximity fuzes, or the warheads could the midcourse systems’ kill vehicles may find it
detonate, either intentionally or accidentally, impossible to detect incoming warheads and
when struck by an interceptor. A nuclear deto- associated objects against such a background
nation at an altitude of 100 to 1,000 kilome- [Stair 1993].
ters would create a large volume of ionized
gas that would attenuate radar signals pass- In summary, nuclear weapons detonated at high
ing through it. For example, a 1 megaton altitudes are countermeasures within reach of
detonation at 400 km would create a cylindri- North Korea that could make midcourse track-
cal ionized region more than 400 km in diam- ing and discrimination extremely challenging
eter, extending within 15 minutes from below and could potentially defeat any current or
300 km to nearly 1,000 km altitude. Radars planned midcourse defense.
would have difficulty tracking any targets Multiple intercept attempts. Theoretically, the
behind this ionized region [Dolan 1972, Fig. defense’s effectiveness could be increased
8-6]. Variations of the ionization density would by making multiple intercept attempts, if fail-
refract radar signals and create directional ure modes are independent. But using multi-
errors. ple interceptors will not improve the system’s
Department of Defense research in 1963 performance if the failures are due to a
investigated the effects of a high-altitude common design flaw or an inability to discrim-
ionized region on radar tracking of warheads inate the warhead.
and found that even ionized regions one Also, this strategy would rapidly deplete the
to two orders of magnitude less dense and interceptor inventory—especially if warheads
much smaller than expected from a nuclear cannot be discriminated from decoys. The
detonation produced ultra-high frequency defense could conserve interceptors with a
(UHF, 0.3–3 GHz) radar tracking errors aver- “shoot-look-shoot” strategy, in which inter-
aging 4 km and variations in the apparent cept attempts are sequential and cease
radar cross-section of a factor of 10,000 [DNA upon confirmation that the target has been
1963]. UHF tracking radars, such as those the destroyed. However, the current GMD system
GMD system relies on, would therefore be has a relatively small number of interceptors
unable to accurately track objects in or behind and has never been tested in shoot-look-
such an ionized cloud. As attenuation scales shoot mode. Nor does it appear to have a
with the inverse square of the frequency, the sensor system that could effectively distin-
higher-frequency S- and X-band (2–4 and guish a warhead from credible decoys or
8–12 GHz) radars fielded for the current GMD reliably confirm the warhead’s destruction. A
system would experience much less atten- new GMD interceptor design with multiple kill

19
vehicles could increase the number of targets and image the threat cloud, though discrimi-
that could be intercepted, as could inclusion nating objects on the basis of their appearance
of Aegis Standard Missile-3 (SM-3) IIA inter- in visible light is unlikely to be effective against
ceptors in the homeland defense system. But anti-simulation countermeasures and such a
a shoot-look-shoot strategy provides little system would be operationally complex to
advantage if the warhead cannot be discrimi- field. This program’s funding was zeroed out in
nated from numerous decoys. the FY22 budget request [DOD 2021].

Proposed midcourse intercept systems that While including multiple kill vehicles on an
could better distinguish warheads from interceptor in place of a single, larger kill vehi-
decoys and execute a shoot-look-shoot strat- cle does not help discriminate warheads from
egy, such as the GMD-E system [NRC 2012], decoys, this strategy makes more kill vehicles
would rely on concurrent, long-duration available to intercept more targets, potentially
observations by X-band radars and infrared improving the system’s effectiveness when its
sensors. However, the MDA instead plans ability to discriminate is poor. The multi-object
to rely on the S-band Long-Range Discrim- kill vehicle project was canceled in 2009 and
ination Radar under construction in Clear, resurrected in 2015, and it again lost its fund-
Alaska, and has fielded an experimental kill ing in 2019. This feature may be included in
assessment system based on commercial the next-generation GMD interceptor that is
satellite-hosted infrared detectors. The latter, being developed.
Space-based Kill Assessment (SKA) system’s
22 sensor payloads are sets of three passively Ground-based Midcourse Defense system
cooled single-pixel photodiodes [Sherman Overview. The Ground-based Midcourse
2019]. They have no tracking capability, but Defense (GMD) system (see Figure  3) is
instead detect flashes for analysis. This system designed to destroy warheads above the
was not designed to determine whether atmosphere using the force of impact of a kill
the intercepted object was a warhead or a vehicle. It comprises 40 interceptors based
decoy. While it might be able to distinguish in underground silos at Fort Greeley, Alaska,
the destruction of a massive re-entry vehicle and four at Vandenberg Air Force Base, Cali-
from a light balloon decoy, it is less clear that it fornia; a suite of space-based sensors and
could tell if the destroyed object was a re-en- ground-based radars; and a command,
try vehicle or part of a rocket booster. A recent control, and communications system. Consid-
GAO report raised several concerns about erable resources have been expended on
the system and noted that missile defense this system. It is expected to cost around $90
commanders did not regard “SKA—and its billion, one of the most expensive Pentagon
intended design—as a proven, operationally systems ever developed. (The GAO’s estimate
sustainable solution” [GAO 2017, 59]. The in 2018 was $67 billion in 2017 dollars [GAO
success of such an approach requires North 2018, 70], which does not include the expan-
Korea to make only limited progress fielding sions proposed in the 2019 Missile Defense
countermeasures. Review, estimated to cost $9 billion [CBO
Other initiatives to increase the U.S. midcourse 2021], or a new interceptor effort, estimated
intercept systems’ ability to discriminate to cost $18 billion [Judson 2021a].)
warheads from other objects include a program The system’s technical roots are in the national
to use lasers hosted on drones to illuminate missile defense (NMD) research efforts of the

20
 9 Intercept
attempt SKA
Threat Cloud

Satellite Sensor

SKA, LRDR 8 Kill vehicle views

Missile releases attempt to assess kill 10
threat cloud
warhead and
decoys (threat
cloud) 6 Kill vehicle
separates from
3 7 SBX, LRDR track threat cloud, interceptor
attempt discrimination

Ground-based
radars track
threat cloud

2 Satellite and radars

detect threat launch Ground-Based
Tracking Radar
1 Forward- 5
Based Radar
Threat GMD
missile interceptor
launch LRDR launch
Sea-Based
X-Band Radar

Figure 3 Sequence of events in an attempted intercept by the GMD system. The launch of a threatening ICBM
from North Korea (1) is detected within a minute by forward-based radars and satellite-based infrared sensors
(2). At the end of the boost phase, the ICBM deploys its warhead and decoys (3). In this example the decoys
are balloons, and a balloon encloses the warhead. The warhead, decoys, and any other accompanying
objects that must be discriminated from the warhead are referred to as the “threat cloud.” Long-range ground-
based tracking radars begin to track the threat cloud (4). Based on this information, the GMD system launches
one or more interceptors from Alaska and/or California (5), each of which launches a kill vehicle (6) toward the
predicted intercept point (9). If a discrimination radar, such as the Sea Based X-band Radar or the Long-Range
Discrimination Radar (LRDR), is in place, it will observe the threat cloud (7) to try to determine which object
is the warhead and will pass this information to the kill vehicle. The kill vehicle also uses its own, onboard
infrared sensor to observe the threat cloud (8) and attempt to determine which object is the warhead. The kill
vehicle then steers itself into the path of the chosen object and attempts to destroy it by the force of impact
(9). The GMD system attempts to confirm the destruction of the chosen object using ground-based radar
(LRDR) and Space-based Kill Assessment (SKA) infrared observations (10). Adapted from [Grego 2016].

1990s. In 2002, the George W. Bush admin- streamlined development process exempted
istration withdrew the United States from the from the usual Pentagon “fly before you buy”
U.S.-Soviet/Russian ABM Treaty that limited system was created, allowing the GMD to be
the two countries’ missile defenses, announc- fielded with minimal oversight and account-
ing that the United States must urgently ability. The MDA used existing technology and
field a system to be able to defend against designs, much of which existed only as proto-
missiles that North Korea, Iran, and Iraq might types, and cut short engineering processes
field [Bush 2002]. It accelerated the deploy- [Grego 2016].
ment of the GMD system to meet a presiden-
tially mandated 2004 deadline. To do so, a Defense Department officials acknowledged
that a development schedule that was driven

21
by externally imposed timelines rather than discrimination and a shoot-look-shoot capa-
technical readiness, and the lack of rigorous bility are untested aspirations.
oversight were sources of significant design
and reliability problems [Butler 2014]. Most Elements of the system. The GMD system’s
interceptors were fielded before interceptors interceptors, which cost about $70 million
with their design had completed even one each, use powerful multi-stage boosters
successful intercept test, and since they were to accelerate the kill vehicle to a speed of
fielded, testing has proceeded at a slow pace, about 7.2 km/sec, permitting it to travel long
with repeated failures. Two decades later, the distances (see [Grego 2016], Appendix 6 and
testing program remains plagued by delays references therein). These boosters carry
and reduced test objectives [GAO 2020]. one of three types of kill vehicles, each with
a different test success rate (see [GMD Tests
Concept of operations. The GMD system’s 2021]). These kill vehicles are complex and
sensors and interceptors are positioned along time consuming to build and to repair, leav-
the northerly trajectories of land-based ICBMs ing them prone to quality control failures
from potential adversaries—North Korea in [DOD 2014]. The MDA has made seven major
particular. Notice of a missile launch would attempts to fix the ground-based intercep-
come within a minute from space-based infra- tor (GBI) kill vehicle in the past 15 years. The
red early-warning sensors and forward-based most recent attempt, the Redesigned Kill Vehi-
radars, and these data would be used to cue cle (RKV), was canceled in August 2019 due
tracking and discrimination radars. to significant technical issues and a tripling of
the cost [GAO 2019a].
Based on the sensor data, the fire control
centers would attempt to discriminate the The current initiative, the Next Generation Inter-
warhead from other objects, including decoys, ceptor (NGI), has two competing bidders who
and launch one or more interceptors toward were selected to develop and build prototype
potential intercept points. Each interceptor’s interceptors, with final selection scheduled for
booster would deploy a 1.4-meter-long kill 2026. Boeing, the prime contractor responsi-
vehicle. The kill vehicle’s onboard computer ble for the GMD system, competed but was not
would choose a target using data from the selected. The Pentagon estimates an $18 billion
kill vehicle’s cooled charge-coupled device lifetime cost for the NGI, including 21 intercep-
(CCD) sensors, which observe long-wave- tors for deployment and 10 for testing, so each
length infrared (LWIR) emissions from the will cost more than half a billion dollars [Judson
threat cloud and compare them with pre-pro- 2021a]. These interceptors will supplement, not
grammed information about the warhead’s replace, the 44 existing GBIs, starting in 2027
expected appearance, adding any informa- at the earliest. Importantly, few spares of the
tion it receives via its limited communica- currently deployed interceptors are available for
tions from the ground. The kill vehicle would tests, and no further intercept tests that could be
maneuver using divert thrusters to collide at used to better understand the existing system’s
a high relative velocity with its chosen target. capabilities are currently scheduled.
(See [Grego 2016] and references therein.) To
improve effectiveness, four or five intercep- The sensors supporting the GBIs include
tors would be launched at each undiscrimi- infrared early-warning satellite sensors and
nated object, which could be the warhead, forward-based radars, two TPY-2 X-band radars
a decoy, or debris. Currently, effective target in Japan, and any Aegis ship-based radars in
the vicinity when the GMD system is used. U.S.

22
Aegis ships deploy SPY-1 S-band radars, some be able to provide long-duration radar obser-
of which will be upgraded to SPY-6 S-band vations of multiple missiles, but at a longer
radars. These radars cue large UHF tracking radar wavelength and hence with less angu-
radars in Alaska, California, Massachusetts, the lar and range resolution. The system is there-
United Kingdom, and Greenland. In addition, fore optimized for less sophisticated threats
there are two radars for discriminating targets: than those assumed in independent studies
the Sea-based X-band radar (SBX), based [Sessler 2000; NRC 2012], which analyzed
on a floating platform that is home-ported in the performance of countermeasures against
Hawai'i, and the S-band Long-Range Discrim- larger numbers of X-band radars.
ination Radar (LRDR) in Clear, Alaska, planned
to begin operations in 2021. Japan planned Proposed sensor improvements include a
to field two Aegis Ashore sites with SPY-7 constellation of low-Earth orbiting satellites
radars built with the same technology as the hosting infrared sensors to track missiles and
LRDR, but recently canceled these land sites possibly discriminate warheads from decoys
in favor of sea-based platforms [Abott 2021]. [Cohen 2019; Insinna 2019]. However, the last
If properly placed and incorporated into the major effort to build such a system, the Preci-
U.S. BMD system, those radars could provide sion Tracking Space System, was terminated
S-band coverage of North Korean missiles in 2013 because it was “too far away from
early in their flight. the threat to provide useful discrimination
data, does not avoid the need for overhead
The GMD system’s current and planned sensor persistent infrared cueing, and is very expen-
architecture is not well suited for successfully sive” [NRC 2012].
discriminating complex countermeasures
from warheads. The warhead and any asso- FY20 plans included two large S-band radars
ciated objects become visible as point-like similar to the LRDR, one to be sited in Hawai'i
objects in the field of view of the kill vehi- and one somewhere else in the Pacific.
cle’s infrared sensors only about one minute However, MDA has decided to reassess the
before the kill vehicle’s projected impact with sensor architecture and has put the additional
the target and cannot be resolved until a few sensors on hold [Judson 2020].
seconds before impact [Grego 2016, Appen- Testing program. To incorporate the system
dix 10]. Once deployed from the interceptor’s into war plans or to decide how to use it
boosters, current kill vehicles have limited under conditions that could include a nuclear
ability to receive and analyze radar and infra- attack, decision makers must have reliable
red data from other sensors in the system. This evidence of the system’s actual effectiveness,
limitation is likely to be mitigated in the new but the 20 years of past GMD tests have been
interceptor design. conducted under scripted conditions and
The SBX can provide X-band observations designed for success. Even so, the system has
over long parts of expected ICBM warhead failed as often as it has succeeded: Of the 19
trajectories from North Korea, but only if it has tests conducted since 1999, the interceptors
been moved in advance to the required loca- successfully destroyed their targets 10 times
tion. Even so, the SBX’s limited “soda straw” [GMD Tests 2021].
field of view makes it unsuitable for observ- The Pentagon has consistently rated the
ing multiple ICBM launches in flight at the GMD tests as low in operational realism; real-
same time [Willman 2015]. The LRDR should ism would require testing against threat-rep-

23
resentative targets that include complex increasing the GMD test tempo would require
countermeasures and with unannounced more trained staff and expanded test infra-
target launch times [DOT&E 2015]. Only structure [Gilmore 2015].
the last two tests have used the warheads
of ICBM-range missiles as targets, and in all There are disincentives, however, to more
the successful intercept tests, the time of the frequent testing or making the tests more
test was chosen so that the kill vehicle would challenging. Since the tests are the most visi-
see the target brightly lit by the sun against a ble indicator of the system’s capability, a high
dark background. And the GMD system has value is placed on succeeding. The MDA’s
yet to be tested against a salvo of attacking position on testing is that “[It] also contributes
missiles. This is a critical test, because a deter- to U.S. non-proliferation goals by sending a
mined adversary could launch several missiles very credible message to the international
at once. community on our ability to defeat ballistic
missiles in flight, thus reducing their value
Midcourse countermeasures in flight tests. to potential adversaries” [MDA 2020b]. The
Critically, as of 2021 no GMD flight test had tests are also expensive, costing $200 million
included complex countermeasures, defined - $300 million each.
by the Director, Operational Test and Eval-
uation (DOT&E) and the MDA as the “use Modeling and simulation (M&S) are critical
of target dynamics and penetration aids” for the GMD program because of the limited
[DOT&E 2015, 38]. When tests have included number of tests and because range safety
decoys, the decoys have been intention- limitations prohibit end-to-end tests over the
ally designed to be much brighter or much expected paths of adversary ICBMs using the
dimmer than the target and the interceptor system’s operational sensors. M&S routinely
has been programmed in advance to use this uses optimistic models of the performance
difference to discriminate the target from the of the GMD system and simplistic representa-
decoys [Wright 2019]. It is not publicly known tion of the operational environment for oper-
whether any test has included a tumbling ational assessments [GAO 2018, 34]. Its threat
warhead, the likely outcome if a warhead models have been developed in-house and
has not been intentionally spin-stabilized. A have not been validated by the Defense Intel-
tumbling warhead would present a challeng- ligence Agency or accredited by the testing
ing time-varying brightness to the midcourse authority [GAO 2018, 32].
intercept system’s sensors. Close coordination between the MDA and the
The GMD’s slow pace of testing—only 19 inter- intelligence agencies to assess threats was a
cept tests in 20 years—and the limited real- key recommendation of the JASON report on
ism of the tests is a serious weakness. Other countermeasures [JASON 2010]. Because of
systems deemed important to national secu- the MDA’s special acquisition arrangements, it
rity are tested much more frequently. The is not required to seek input from the defense
Trident II submarine-launched ballistic missile, intelligence community, and the defense intel-
for example, was tested dozens of times ligence community is struggling to provide
before deployment in 1990 and continues to the MDA timely and detailed information,
be tested about five times per year (see the though efforts are underway to improve this
Trident II table in [McDowell 2021]). The MDA situation [GAO 2019b]. The Pentagon’s oper-
and the Pentagon testing authority state that ational testing office’s current assessment is
that the M&S effort “lags behind operationally

24
realistic threats with respect to countermea- of new interceptors, and adding a new S-band
sures, debris, raid sizes, and electronic attack,” radar (the LRDR) in Alaska. At present, the GMD
and that it “remains insufficient to support system still does not have continuous X-band
quantitative effectiveness and lethality assess- radar coverage, nor the ability to fuse data
ments” [DOT&E 2021]. on the threat cloud obtained using the infra-
red sensors on board the kill vehicle with data
Overall assessment. Despite significant invest- obtained using off-board radar observations.
ment of resources and decades of effort, the The GAO continues to warn that the MDA is
GMD system has not been shown to be reli- developing next-generation systems (in partic-
ably effective even in carefully scripted tests, ular the LRDR, the SKA, and the now-canceled
and its effectiveness in battlefield situations RKV) by making “tradeoffs that favor field-
is likely to be low. If rigorous engineering ing capabilities sooner and less expensively”
procedures are followed in developing a new and which DOD officials are concerned “will
interceptor, some of the previous design and compromise performance and reliability” and
reliability problems should be addressed. may end up being insufficient against current
However, even if those improvements are and anticipated threats [GAO 2017, 59].
made, the issue of effectively discriminating
warheads from decoys will remain unsolved. For most of the next decade, therefore, the
The MDA has made little progress in this area, core of the GMD system will be 44 low-re-
and to assess the system as designed as likely liability interceptors that would need to be
to be successful, optimistic assumptions must launched in salvos against each credible target
be made about the adversary’s ability to field (though the system has been tested in a salvo
countermeasures. The system sensors also mode only once, using a salvo of only two inter-
are not robust against direct attack or high-al- ceptors). Sometime near the end of the current
titude nuclear detonations. decade, an additional 21 newly designed
interceptors are projected to be fielded. For
The National Academies report [NRC 2012] the simplest of threats, such as a single missile
therefore recommended a complete over- or a few with the type of simple countermea-
haul, including redesigning the system with sures the system is designed to handle, this full
new interceptors and sensors, and with multi- system may provide some capability. As the
ple X-band radars to cover the likely paths Director of Operational Test and Evaluation
of missiles from North Korea and Iran to the concluded, when the GMD system can use its
United States to make the system more robust complete, proposed architecture of sensors
to sensor outages. It proposed a concept of and command-and-control systems, it “has
operations that relied on a shoot-look-shoot demonstrated capability” to defend the United
strategy, simultaneous observations of the States against a small number of intermediate
threat cloud using infrared and visible light range ballistic missiles or ICBMs “with simple
sensors and X-band radars over long periods, countermeasures” [DOT&E 2021].
ongoing communications between off-board
sensors and the kill vehicle, and fusing this data However, because the system is not designed
to improve the system’s ability to discriminate. to reliably discriminate a warhead from decoys,
it is likely to quickly exhaust its inventory of
The DOD apparently judged it infeasible to interceptors when faced with an attack that
start over and instead continues to plan incre- includes more missiles and better countermea-
mental improvements, such as refurbishing sures, such as the baseline threat considered in
existing interceptors, building a limited number

25
this study. Moreover, this system, which relies The Aegis BMD system was originally designed
on a small number of large radars and satel- to defend aircraft carrier battle groups from
lites with limited redundancy, is not resilient to short- to intermediate-range ballistic missiles.
direct attacks on these sensors. It is becoming increasingly capable as it is
upgraded with faster and more sophisticated
Due to its fragility to countermeasures, and interceptors; soon, it will also be equipped with
the inability to expand it readily or cost-effec- more capable shipboard radars. The newest
tively, the current midcourse intercept system SM-3 Block IIA interceptor may be fast enough
cannot be expected to provide a robust or reli- to potentially defend large areas of U.S. terri-
able capability against more than the simplest tory against ICBMs if launched from a site near
attacks by a small number of relatively unso- a U.S. coast. However, it is not clear how well
phisticated missiles within the 15-year time suited the system is for this task, given that inter-
horizon of this report. cepting ICBM warheads was not its intended
Potential additional midcourse intercept purpose and neither its sensors nor its inter-
layers: Aegis BMD and THAAD ceptors were designed for this task. Congress
therefore mandated a test of the Aegis system
The Donald Trump administration proposed against an ICBM-range missile.
using the Navy’s ship- and shore-based Aegis
BMD system and an upgraded version of The test was conducted in November 2020.
the THAAD system to augment the defense An Aegis ship stationed northeast of Hawai'i
provided by the GMD system (see Figure destroyed the warhead launched by an ICBM-
2). While no proposed locations for these range missile using an SM-3 IIA interceptor
systems have been specified, the MDA esti- [DOD 2020b]. Despite being executed under
mates that a single Aegis site could defend highly favorable conditions [GAO 2021], the
an area one-fourteenth the size of the area test stressed the system. At a press event, the
the GMD is designed to defend (which is the Director of the Missile Defense Agency, Vice
United States) [Hill 2020a]. Some analysts esti- Admiral Jon Hill, stated that, to intercept the
mate that an Aegis site could defend an even target, the ship had to maneuver to a better
larger area, based purely on the speed of the location and the interceptor had to use “the
Aegis interceptor (see, e.g., [Butt 2011]). A highest divert” of any test [Eckstein 2021b]. The
single THAAD system is designed to defend GAO states that “several challenges” remain to
a much smaller area yet, so many THAAD sites be overcome to make the Aegis system a work-
would be needed for a layered defense of the able defense against realistic ICBM threats, and
entire United States. notes that some elements of the SM-3 IIA inter-
ceptor may prove to be unsuited to the longer-
The Aegis BMD system is currently hosted on U.S. range ICBM mission [GAO 2021]. One critical
Navy cruisers and destroyers and at Aegis Ashore issue among many is whether Aegis intercep-
ground sites (one in Romania, one in Poland, and tors can reliably be launched and guided to an
a test site in Hawai'i). Each system includes a four- ICBM warhead by offboard radars, which would
faced S-band phased-array SPY-1 radar, dozens be necessary for the system to potentially cover
of vertical launch tubes that can launch SM-3 enough territory to make a meaningful contri-
exoatmospheric hit-to-kill interceptors, and a bution to defending the U.S. homeland against
command-and-control system that can provide ICBMs. The Aegis system is of course suscep-
target information based on tracking from radars tible to the same midcourse countermeasures
in other locations [CRS 2021b]. as the GMD system. Additionally, some Navy

26
officials have expressed frustration that when tions for the strategic relationships between
performing missile defense duties to protect the United States and China and Russia (see
land areas, the very sophisticated and capable also the discussion in Section 3: “Challenges
Aegis ships are pinned down in geographically of Missile Defense”; [Baklitskiy 2021], 16 ff;
small areas and are unable to perform other [Erästö 2021]).
missions (see [CRS 2021b, 16–19]).
The United States plans to have 60 Aegis
THAAD was designed to defend areas the size BMD-capable ships by the end of FY23 [MDR
of military bases against the warheads of short- 2019, 48] that will host scores to hundreds of
to intermediate-range missiles and can attempt SM-3 IIA interceptors. The GMD and Aegis
hit-to-kill intercepts of warheads at altitudes of interceptor inventory will then be much
40 - 150 km (within and just above the atmo- larger than the expected numbers of Chinese
sphere) and ranges of up to 200 km [Reuters missiles that could survive a U.S. first strike.
2017]. The suitability of the THAAD system for The anticipated deployment of these inter-
a local defense against ICBM warheads has ceptors is giving China incentives to increase
not been established or tested. The THAAD and diversify its offensive nuclear capabili-
system’s X-band radar provides better range ties and disincentives to engage in nuclear
resolution and discrimination capability than arms reductions. China currently has only 72
the existing Aegis radars, but before initial tests mobile ICBMs [Kristensen 2020], but it may
can be conducted against ICBM warheads (in now be building several hundred new ICBM
2023 at the earliest; see [Sherman 2020]), the silos that could be intended to make a U.S.
system will need crucial upgrades that, among disarming first strike more difficult [Kristensen
other things, would significantly increase the 2021]. As James Miller, a former Undersecre-
speed of its interceptor [MDA 2020a]. While tary of Defense for Policy during the Barack
THAAD interceptors can intercept within the Obama administration, has noted, the objec-
atmosphere, the system could still be deceived tive “to bring the SM-3 IIA missile into the
by lightweight midcourse countermeasures national defense architecture . . . means that
until the last minute of the warhead’s flight. China and Russia must expect the United
States by 2025–2030 to have many hundreds
Wider implications of planned U.S. of available interceptors for national missile
midcourse intercept systems defense.” He warned, “We should expect the
Given the technical realities of the existing Chinese nuclear arsenal to grow substantially
U.S. midcourse intercept systems and the and Russia to resist reductions below the 2010
limits imposed on their future effectiveness New Strategic Arms Reduction Treaty—and to
by countermeasures, the enormous planned prepare seriously to break out” [Reif 2019].
investments in these systems are likely to A clear-headed assessment of the economic
provide only incremental rather than compre- and security costs of pursuing midcourse
hensive improvements in their capability. defense, together with a careful assessment
But the unbounded nature of the U.S. missile of its possible benefits, is critical for U.S. secu-
defense enterprise and the planned dramatic rity. Given the information presented in this
expansion of the Aegis BMD system—even section, it has become increasingly appar-
if developed primarily to counter existing ent that the drawbacks of the current U.S.
threats from North Korea and potential future midcourse defense program outweigh its
threats from Iran—has important implica- potential benefits.

27
5. BOOST-PHASE INTERCEPT SYSTEMS

Systems that would disable attacking ICBMs 2019, Secs. 1676 and 1680; MDR 2019; MDA
during their boost phase—while their rocket 2019, Sec. PE 0604115C; NDAA 2020, Sec.
engines are still burning and before they 1682; NDAA 2022, Sec. 1664]. To be reli-
have deployed their nuclear warheads— able and effective, a boost-phase intercept
first attracted significant interest in the early system must have operational capabilities
1980s, but no effective system was developed that are not just marginal when used for the
then. Such systems again attracted interest in intended mission, but sufficient to deal with
the early 2000s, as the difficulty of midcourse unexpected events and contingencies. In this
intercept became increasingly obvious [APS chapter we reexamine these types of systems
2003, S2], but careful analyses showed that and assess whether anything has changed in
such systems were still not feasible [APS 2003; the past decade that would alter the conclu-
NRC 2012]. sions of the National Academies study
regarding boost-phase defenses against
For example, the 2012 National Academies North Korean ICBMs.
report concluded, “With one or two minor
exceptions, land-, sea-, or air-based boost- As we explain, the situation regarding boost-
phase defense is not feasible when time- phase rocket interceptors based on land has
line, range, geographical/geo-political, or not changed. Unless they could be coopera-
cost constraints are taken into account” [NRC tively based in China or Russia, they would not
2012, S-6]. It also found that the total life-cy- be feasible [NRC 2012, S-12, footnote 13]. Nor
cle cost of placing and sustaining the number has the situation regarding sea- or air-based
of space-based interceptors required for a rocket interceptors for boost-phase intercept
boost-phase defense system was at least an fundamentally changed; neither approach
order of magnitude greater than that of any could protect the entire continental United
other alternative, making the project imprac- States; at most they would cover only some
tical for that reason alone [NRC 2012, S-7]. limited regions.
Consequently, its first major recommenda-
tion was, “The Department of Defense should As we describe, the number of interceptors
not invest any more money or resources in required for a space-based interceptor system
systems for boost-phase missile defense. to be able to defend in principle against
Boost-phase missile defense is not practical North Korea’s Hwasong-15 liquid-propellant
or cost effective under real-world conditions ICBM is at least 400, and about 4,000 would
for the foreseeable future” [NRC 2012, 4-13]. be required to defend against a salvo launch
of 10 such ICBMs. Many times more inter-
However, boost-phase systems that would ceptors would be required to defend against
disable attacking ICBMs using rocket inter- solid-propellant ICBMs, should North Korea
ceptors or laser weapons carried by fighter acquire them. (We note that Iran is assessed
aircraft or drones, or similar systems based to have the technical and industrial capacity
on platforms in low-Earth orbit are again needed to develop ICBMs and in April 2020
being proposed [Abott 2018; NDAA 2018, launched a satellite using its three-stage
Secs. 1685 and 1688; Cohen 2019; NDAA solid-propellant Qased rocket, which could

28
probably be transformed into a long-range Key challenges of boost-phase intercept. A
ballistic missile [Elleman 2021a].) This situ- boost-phase intercept system must success-
ation remains very unfavorable for a space- fully and simultaneously deal with a number
based defense. In principle, the financial costs of challenging problems for which solutions
of building and launching commercial space- have not yet been demonstrated. Its inter-
based systems have decreased dramatically, ceptors must be based on platforms in loca-
but as we discuss, whether these economies tions that are geographically and geopolit-
could be captured by a space-based inter- ically feasible and secure, which generally
ceptor system is unclear. The weaponization limits their performance, yet be able to reach
of space and arms race instability that would the target ICBM within about two to four
be caused by testing and deploying a constel- minutes after it has been launched. Conse-
lation of space-based interceptors are signif- quently, the launch of any threatening ICBM
icant issues in addition to its technical chal- must be detected, its trajectory estimated, a
lenges and cost. firing solution for the interceptors computed,
and interceptors fired less than a minute after
We also note that ICBMs launched from North the launch of the ICBM has been confirmed
Korea would need to be intercepted over by remote sensors. The performance required
Chinese territory, hundreds of kilometers to intercept North Korean ICBMs during their
inside China’s borders. Hence, to respond boost phase would be much less if intercep-
effectively to a suspected ICBM attack by tors could be based in China or Russia (see
North Korea, a boost-phase intercept system Figure 4).
would have to launch at least several, and
perhaps dozens of interceptor missiles over An interceptor rocket that strikes an ICBM
Chinese territory, and their final stages would while it is in powered flight will damage it suffi-
come down in China or Russia. The conse- ciently to terminate its thrust, though perhaps
quences of firing such a system by mistake not immediately. The collision may be violent
could be very serious. Such a boost-phase enough to cause the warhead to explode,
system would therefore have to be able to either because it has not been constructed to
reliably identify the launch of a threatening remain safe if struck, or because it has been
missile and distinguish it from other events designed to explode if it is struck (“salvage
with very high confidence. fuzing”). If the warhead explodes when the
ICBM is hit, the explosion could blind the
Appeal and challenges of boost-phase defensive system’s sensors, interfering with
intercept its ability to intercept other ICBMs launched
Boost-phase intercept systems have attracted at nearly the same time (a “salvo launch”). If
attention for several reasons: Intercepting an the intercept does not cause the warhead to
ICBM during its boost phase could prevent explode, the warhead may remain attached
any of its warheads from striking their targets, to the ICBM’s final stage, in which case both
so a single, effective boost-phase intercept will re-enter the atmosphere together, or the
system could in principle defend a very large warhead may separate from the ICBM’s final
area; and intercepting ICBMs during their stage, in which case it will re-enter the atmo-
boost phase has sometimes been portrayed sphere separately. Either way, it may detonate
as easier than intercepting warheads during when or before it hits the ground [APS 2003,
the midcourse or terminal phases of flight. Sec. 13]. We do not consider what would be
needed to disable or destroy the warhead,

29
which is a much more demanding task [APS ble technical problem and may not be possi-
2003, Sec. 13.2]. This report instead discusses ble. The seriousness of this problem is miti-
what would be needed to prevent an ICBM’s gated by the context: such a shortfall would
warhead from reaching various parts of the occur during a nuclear war and the warhead
continental United States by terminating the would likely explode on an area with a rela-
ICBM’s thrust sufficiently early. tively low population. We do not attempt
to address this problem in our report (for a
According to our modeling, preventing a detailed discussion of this problem, see [APS
warhead launched by a Hwasong-15 from 2003, Sec. 5.8]).
reaching any part of the continental United
States would require intercepting it no later While a boost-phase defense could poten-
than 260 seconds after it is launched, and tially reduce the number of missiles that the
preventing a warhead launched by our midcourse defense would face, it could also
notional solid-propellant ICBM (which is make midcourse defense more difficult. For
based on the S1 ICBM in the 2003 APS report) example, if a boost-phase intercept destroys
from reaching any part of the continental the booster but not the warhead, the intact
United States would require intercepting it warhead may then be accompanied by the
no later than 145 seconds after it is launched. booster debris, or the warhead may be set
Intercepting these ICBMs this early would tumbling or spinning in ways that the defense
protect cities in Alaska as well as cities on has not anticipated (see [APS 2003, Sec. 13.3]).
the U.S. East and West Coasts. It would also
eliminate the efficacy of a late-stage dog-leg The reach-versus-time challenge. Boost-phase
maneuver that would sacrifice range (e.g., intercept systems face a severe reach-versus-
hit Alaska instead of Boston) in order to evade time challenge because their interceptors
the boost-phase defense (see [APS 2003, Sec. must be based in safe or defendable loca-
15.2]). (“Dog-legs” are maneuvers in which tions, which are typically 500 km or more from
the missile starts out in one direction and then the location where the intercept occurs; their
veers off in another, making it difficult for the interceptors cannot be fired until the ICBM’s
defense to anticipate the eventual missile direction of flight has been determined; and
trajectory.) they must reach the ICBM early enough to
prevent it from delivering its warhead to a
As noted above, if the intercept is otherwise target in the United States. It is difficult even
successful but does not disable the warhead, for fast interceptors to achieve this. Whether it
the warhead will fall short of its intended is possible depends on many factors, includ-
target but may detonate when or before it hits ing details of the offense—where the ICBM is
the ground. For ICBMs launched from North based and how long its powered flight lasts
Korea, the resulting nuclear detonation would (the “burn time” of its boost phase), which
not occur over North Korea, but over China, depends strongly on whether it is a liquid-
Russia, Canada, or locations within the United or solid-propellant missile and its intended
States that are closer than the intended target. target—and the detailed performance capabil-
This poses a complex political and human- ities of the defense—the speed of the intercep-
itarian problem, called “the shortfall prob- tor and whether it is fired almost automatically
lem” (see [APS 2003]). Timing an intercept to or some decision time is allowed, and whether
prevent a live warhead from falling on other the system is expected to defend all or only
countries and exploding presents a formida- part of the United States. (We use the term

30
“decision time” in the same way as [APS 2003],
to refer to any additional time after the ICBM’s Russia
trajectory is first estimated that can be used for

ks
n
communication between system elements to

rba
co

Boston

llas
evaluate whether a reported launch detection is

Fa i
China nc

Da
a
is an ICBM or a spoof; to resolve any uncer-
n Fr
tainties about the performance of the defen- Sa
sive system; and to better identify the type of
missile detected, its likely performance char-
acteristics, and its trajectory [APS 2003, xxiii,
S70].) N. Korea Hawaii

Figure 4 provides a map of North Korea and

the adjacent parts of China and Russia with
the initial ground tracks of ICBMs launched Japan
from north-central North Korea to five cities in S. Korea
the United States. (The initial azimuths of these
ground tracks are about 10° farther north than 500 km
the initial azimuths of the great circles connect-
ing the launch site to the targets because of Figure 4 Map showing North Korea and adjacent
the effects of Earth’s rotation.) countries and the initial ground tracks of ICBMs
launched from north-central North Korea to five
The kinematically allowed basing area for a cities in the United States. ICBM ground tracks
given interceptor, decision time, and inter- differ from great circles connecting the launch
cept time is the circular area on the ground site to the target because of Earth’s rotation.
centered directly under the point on the
ICBM’s trajectory where it will be at the
moment when it is intercepted. The radius where interceptors can be safely positioned
of this area is approximately equal to the or defended.
distance the interceptor can travel from the
Figure 5 illustrates these challenges using a
time it is fired until the time it intercepts the
model of North Korea’s Hwasong-15 and our
ICBM (see [APS 2003, Sec. 4.6] for a more
notional model of a solid-propellant ICBM
precise definition and a more detailed discus-
launched from a site that favors the offense,
sion). The kinematically allowed basing area is
but with other assumptions that favor the
larger the later the ICBM can be intercepted
defense (see below).
and is largest if the ICBM can be intercepted
just before it gives its warhead sufficient veloc- Boost-phase intercept of ICBMs launched
ity to reach the intended target [APS 2003, Ch. from even a small country like North Korea is
5]. Note, however, that the intended target very challenging. As a result, whether an ICBM
is generally not known in advance by the can be intercepted before it gives its warhead
defense. Also, some portions of the basing sufficient velocity to reach the intended target
area determined in this way may be unavail- depends on the type of ICBM, its target, and
able or unsafe places to base interceptors. the performance of the interceptor, as well as
The possible basing area is that portion of the decision time needed by the defense. This
the kinematically allowed basing area, if any, is illustrated by Figures 5(a) and 5(b).

31
 at 260 s at 145 s

Boston
s eles
ele ng

n
ng

Bosto
A sA
os Lo
L

0 200 400 600km 0 200 400 600km

Figure 5: Possible basing areas for the interceptors discussed in the text to be able to reach liquid- and
solid-propellant ICBMs launched from North Korea in time to prevent them from delivering their warheads
to the indicated target cities. (a) Initial ground tracks of Hwasong-15s launched toward Boston and Los
Angeles. Our model of the Hwasong-15 would have to be intercepted no later than 260 seconds after
launch to defend all of the continental United States. The positions of the two Hwasong-15s at this time
are indicated by the crosses on their trajectories. The colored circular areas show where 5 km/s intercep-
tors could be based and intercept them at this time, assuming the interceptors are fired without allowing
any time for a decision whether to fire (zero decision time). The two dashed lines indicate distances of 100
and 200 km from the East Coast of North Korea beyond which interceptors would have to be based to be
safe from North Korean defenses, depending on the performance of these defenses (see text). There is
no safe basing area that would allow interceptors to reach Hwasong-15s aimed in the direction of Boston.
(b) As in (a), but for our notional solid-propellant ICBM. It would have to be intercepted no later than 145
seconds after launch to defend all of the continental United States. The positions of the two solid-propel-
lant ICBMs at this time are indicated by the crosses on their trajectories. The colored circular areas show
where 5 km/s interceptors could be based and intercept them at this time, assuming the interceptors are
fired with zero decision time. The two dashed lines again indicate distances of 100 and 200 km from the
East Coast of North Korea. There are no safe basing areas that would allow these interceptors to reach
such ICBMs headed to any place in the continental United States. The limited basing areas shown in this
figure reflect the severe reach-versus-time challenge of boost-phase intercept. After [APS 2003, Fig 5.9].

Figure 5(a) shows the kinematically allowed than 260 seconds after they were launched
basing areas from which interceptors with the would be significantly smaller. Importantly, for
ability to rapidly accelerate to 5 km/s and fired Hwasong-15s launched from sites in north-cen-
with zero decision time could reach our model tral North Korea in the direction of Boston—or in
of the Hwasong-15s launched from north-cen- the direction of other cities in the northeastern
tral North Korea in the direction of Boston or United States, such as New York—these possible
in the direction of Los Angeles, 260 seconds basing areas do not extend outside the terri-
after they were launched. tory of North Korea, Russia, and China. Conse-
quently, basing interceptors in them would not
The basing areas that would make it possible currently be politically realistic. The same is true
to intercept Hwasong-15s significantly earlier

32
for the basing areas from which Hwasong-15s intercepted earlier than 285 seconds after
launched from sites in northwest North Korea launch. Figure 5(a) shows that if the defense
could be intercepted. For intercepts signifi- planned to intercept an ICBM in the direc-
cantly later than 260 seconds after launch, the tion of Los Angeles as early as 260 seconds
kinematically allowed basing areas would have after launch, there would be a portion of the
larger radii, but would be further north and kinematically allowed basing area where the
again would not extend outside North Korea, interceptors we are discussing could be safely
Russia, and China. based. If the defense planned to intercept an
ICBM aimed in the direction of Los Angeles
As we now explain, even if the intercept could later than 260 seconds after launch, it could
be timed to occur at 260 seconds after the make use of a larger safe basing area or allow
Hwasong-15 was launched, which would some decision time before firing its intercep-
require precise knowledge of its perfor- tors. Similarly, there could also be some safe
mance and intended target, there would be basing locations from which interceptors
no portion of the kinematically allowed basing could intercept Hwasong-15s from northern
area where these interceptors could safely be North Korea early enough to prevent them
positioned. from striking cities in the Midwest or on the
The absence of any safe basing area is shown West Coast of the United States, though not
by the two dashed lines in Figure 5(a), which cities elsewhere in the United States.
indicate the distances of 100 and 200 km off The examples shown in Figure 5(a) illustrate
the east coast of North Korea beyond which several key considerations of interceptor
interceptors would be safe while on-station, basing. First, if the interceptors being consid-
depending on the assumed capabilities of ered are not based in China or Russia, in most
North Korea’s sea and air defenses. The latter cases they could not reach ICBMs launched
may include as many as six batteries of older toward targets in the continental United States
S-200 surface-to-air missile systems, which until they are over Chinese territory. Second,
have a maximum range of 250 to 400 km, to be able to reach Hwasong-15s aimed in the
depending on the type, and an unknown direction of cities on the U.S. East Coast early
number of more modern KN-06 systems that enough to defend these cities, these intercep-
resemble the Russian S-300 or Chinese HQ-9 tors would have to be based too close to the
and are claimed to have a range of 160 km east coast of North Korea to be safe from North
[Yeo 2017]. North Korea has recently tested Korean air and sea defenses, even if they were
what it says is a newly developed surface-to- fired with zero decision time. Third, there are
air missile system called the Pon’gae-6 [Kim safe locations where interceptors like these
2021b; Rahmat 2021]. No portion of the kine- could be safely based and reach Hwasong-15s
matically allowed basing area for intercept- aimed in the direction of cities in the Midwest
ing a Hwasong-15 aimed in the direction of or on the U.S. West Coast early enough to
Boston extends beyond the dashed lines. prevent them from delivering warheads to
A Hwasong-15 launched from northern those targets, even allowing some decision
North Korea and aimed in the direction of time. However, it is not to be expected that
Los Angeles could strike targets in the Aleu- North Korea would choose to launch ICBMs in
tian Islands unless it is intercepted earlier directions it knows would make them vulner-
than 260 seconds after launch, but could be able to the defense. A boost-phase defense
prevented from striking Los Angeles if it is would be much easier kinematically if inter-

33
ceptors could be based in China or Russia; after the launch is first detected by remote
however, short of extensive cooperation in sensors. For a typical solid-propellant ICBM,
such a defense, the United States cannot real- a firing solution is expected to be available
istically or prudently expect that interceptors about 45 seconds after it was launched, about
intended for defense against North Korean 15 seconds after the launch has been detected
ICBMs can be stationed in Chinese or Russian [APS 2003, Fig. 2.2; NRC 2012, Figure 2-3]. The
territory or airspace [NRC 2012, S-12]. 2012 National Academies study found that it
is counterproductive to commit an intercep-
Deployment of a boost-phase defense by the tor earlier than these times [NRC 2012, 2–27].
United States would give North Korea a strong Committing interceptors this early means they
incentive to develop and deploy solid-propel- must be fired almost automatically, i.e., with no
lant ICBMs, because they have much shorter decision time. Allowing 30 seconds of deci-
burn times than liquid-propellant ICBMs and sion time would reduce the radii of the kine-
the time available to intercept them is there- matically allowed basing areas by about 150
fore much reduced. This is illustrated by Figure km for a 5 km/s interceptor, making intercept-
5(b), which shows there is no safe basing ing even a Hwasong-15 aimed in the direction
area from which a 5 km/s interceptor could of Los Angeles substantially more challeng-
reach a solid-propellant ICBM launched from ing. That all these times are very short reflects
northwest or north-central North Korea early the severe reach-versus-time challenge of
enough to prevent its warhead from striking boost-phase intercept [NRC 2012, S-11]. The
any part of the United States, even if the inter- time available might be increased if distrib-
ceptor were fired with zero decision time. uted sensors and machine learning allow
The kinematically allowed basing areas shown as-yet-unquantified improvements in estimat-
in Figure 5 make the same assumptions that ing the trajectory of the target ICBM quickly
were made in the 2003 APS study, which and deciding whether to fire interceptors.
generally favor the defense [APS 2003, xxvi]. The kinematically allowed basing areas shown
For example, they make optimistic assump- in Figure 5 also assume the interceptor has a
tions about the missile detection and track- burnout velocity of 5 km/s, which is slightly
ing capabilities available to the defense. They higher than the maximum velocity of a rocket
also assume that the interceptor is fired at interceptor based on current technology that
the earliest moment a firing solution can be could fit in an Aegis vertical launch system
constructed. tube [NRC 2012, 2-17] and the highest veloc-
A firing solution cannot be constructed as ity that has been proposed for airborne rocket
soon as an ICBM is launched. When it would interceptors [Garwin 2018a; Garwin 2018b].
become available depends on a number The calculated basing areas further assume
of factors, including the type of ICBM, what a very fast-burning interceptor, which burns
remote sensors are available, meteorological out after only 25 seconds, about half the burn
conditions at the time, and the capabilities of time of the 5  km/s interceptor used in the
the interceptor (see [APS 2003], Section 2.4.1, 2003 APS study [APS 2003, Table 5.3]. While
for a detailed analysis and explanation). For a shortening its burn time increases the reach
liquid-propellant ICBM like the Hwasong-15, of the interceptor by about 60 km, it lengthens
this is expected to be possible about 65 the already long duration of the interceptor’s
seconds after it is launched, about 20 seconds coasting phase, when the rocket motor of the
booster has burned out but the rocket motor

34
of the kill vehicle has not yet begun firing. also not accounted for possible operational
During this phase, the interceptor cannot delays in processing and transmitting infor-
adjust its trajectory to compensate for delib- mation. All of these factors must be included
erate or unexpected incidental accelerations when assessing possible boost-phase inter-
of the target ICBM. Lengthening the duration cept defenses against ICBMs.
of this phase decreases the likelihood that the
interceptor will be able to hit its target [APS Kill vehicle and system requirements. The kill
2003, Sec. 2.2, 12, and 14, and Appendices B vehicles carried by boost-phase interceptors
and C]. Finally, these basing areas do not take must reach the ICBM early enough to prevent
into account the possibility that the attack- its warhead from reaching the target and must
ing ICBM could be programmed to use any have the sensors and cumulative divert veloc-
energy it has beyond the minimum needed to ity required to be able to home in on and hit
reach its target to fly a dog-leg trajectory to the dim missile body rather than its bright
the target that keeps it farther from potential exhaust plume while the missile is moving at a
safe interceptor basing areas. velocity of about 6 km/s and accelerating and
maneuvering somewhat unpredictably [NRC
On the other hand, the ICBM launch site 2012, 2-31]. This requires a sensor such as a
assumed in Figures 5(a) and 5(b) is one of light-detection-and-ranging (LIDAR) system
the most challenging launch sites for a boost- and a kill vehicle with a cumulative divert
phase defense, in the sense that trajecto- velocity of at least 2.5 km/s [APS 2003, Sec.
ries to the United States from northwest and 12.3.2]. The system’s sensors and kill vehicles
north-central North Korea would be hardest must not be confused, misled, or distracted by
for interceptors not based in China or Russia to countermeasures the attacker could employ
reach before the ICBM has achieved the veloc- (see below). Finally, the system must be able
ity needed to deliver its warhead to a target in to handle the complex battle management
the United States. Boost-phase intercept would task of assigning multiple interceptors to
be easier for ICBMs launched from some other multiple attacking ICBMs and guiding the kill
sites in North Korea, but it is not expected that vehicles to their targets. It must be able to do
North Korea would choose to launch its ICBMs this even if confronted with debris created by
from such sites knowing that they would then successful previous intercepts.
be more vulnerable to intercept.
Countermeasures. Although a boost-phase
This discussion also does not account for many defense would not be susceptible to some of
of the real-world factors that would have to be the countermeasures to midcourse defense
considered to realistically assess the capability that have been proposed, it would face
of a proposed boost-phase intercept system countermeasures [APS 2003, Ch. 9; NRC 2012,
[APS 2003, xxvi and Sec. 5.1.3]. These include 2-31]. In order to avoid arguments about what
lack of knowledge of the locations of the countermeasures to boost-phase intercept
adversary’s ICBM launch sites, uncertainties are or are not feasible, the 2003 APS report
about the performance of its missiles and their considered only techniques that have actu-
possible maneuvers during flight, ignorance of ally been employed in operational systems
a missile’s intended target, the unpredictable over the past 60 years and that North Korea
nature of the variations in any missile’s flight, is therefore likely to be able to implement.
and uncertainties in how quickly an intercept Examples include (a) launching several ICBMs
would terminate the ICBM’s thrust. We have nearly simultaneously (salvo or staggered

35
launch); (b)  launching smaller decoy rock- Coast and then shifting its trajectory to strike
ets simultaneously with the ICBM, to confuse cities in Alaska or the U.S. Northwest. Inter-
the defense; (c) deploying solid-propellant ceptors launched from ship-based VLS tubes
ICBMs, with their much shorter burn times; could intercept long-burning liquid-propel-
(d) deploying the ICBM’s warhead (re-entry lant ICBMs aimed in the direction of cities
vehicle) while its final stage is still burning; on the U.S. West Coast. In order to be able to
(e) deploying rocket-propelled decoys and attempt intercept of a salvo of 10 such ICBMs,
jammers during the flight of the ICBM’s upper VLS tubes would have to be preloaded with
stages; (f) programming the upper stages to 10 or 20 interceptors, depending on their
fly evasive maneuvers, possibly in conjunction expected effectiveness against ICBMs and
with deployment of decoys and jammers; and the expected countermeasures. Interceptors
(g) deploying short-burn boosters with multi- launched from VLS tubes would be unable to
ple upper stages, each with its own warhead. intercept solid-propellant ICBMs headed to
Each of these countermeasures constitutes the United States.
an independent and important challenge to
boost-phase intercept. Aegis ships are being considered for use as
platforms for rocket interceptors that would
Using land- and sea-based rocket be used for intercepting ICBM warheads late
interceptors for boost-phase intercept in their midcourse flight, but this has been crit-
icized as an inefficient use of these expensive,
As discussed in the previous section, land- very capable ships [CRS 2021b, 16-19]. The
based rocket interceptors would have to be same criticism could be made of continuously
based in China or Russia, north of poten- basing Aegis ships off the coast of North Korea
tial launch sites in North Korea, to be able to and Russia, as part of a boost-phase defense.
intercept even a long-burning, liquid-pro- This criticism would be less relevant if the plan
pellant ICBM like the Hwasong-15 launched were to surge ships to positions off North
from northwest or north-central North Korea Korea and Russia in case of high tensions or
toward the U.S. East Coast, in time to prevent a crisis [MDR 2019, XV and 56]. Interceptors
its warhead from striking the United States. launched from Aegis ships could not intercept
Sea-based rocket interceptors small enough solid-propellant ICBMs headed to any targets
to fit in Aegis vertical launch system (VLS) in the continental United States.
tubes would have maximum burnout veloc- Using drone-based rocket interceptors for
ities slightly lower than the 5 km/s burnout boost-phase intercept
velocity assumed in the previous discussion.
They therefore also could not reach an ICBM A system of drone-based rocket intercep-
like the Hwasong-15 launched from northwest tors for a boost-phase defense against ICBMs
or north-central North Korea toward the U.S. launched from North Korea would avoid the
East Coast in time, even if the Aegis ships were discrimination problem faced by all midcourse
positioned within 200 km of the east coast of intercept systems and could be designed not
North Korea or Russia and the interceptors to threaten current Russian or Chinese ICBMs
were fired with no decision time. Depend- [Garwin 2017; Goodby 2018]. It would require
ing on its maximum range, an ICBM like the high-altitude, long-duration drones able to
Hwasong-15 could also evade intercept by carry high-speed rocket interceptors with kill
starting on a trajectory toward the U.S. East vehicles capable of intercepting a maneuver-

36
ing ICBM (for further information, see [Garwin fraction of the time the ICBM is in powered
2017; Garwin 2018a; Garwin 2018b; Goodby flight preceding the intercept attempt, reduc-
2018; Postol 2018]).The drones would need to ing the chance that they would be able to
loiter on-station for tens of hours and, as noted intercept the ICBM reliably (see [APS 2003],
above, might have to stay 100 to 200 km off Sections 2.2, 12, and 14, and Appendices B
the coast of North Korea to be safe from North and C). If a decision were made to develop
Korean air defenses. a defensive system that would use drone-
based rocket interceptors, it would probably
As noted previously, such a system armed be necessary to use more capable intercep-
with 5  km/s interceptors could not defend tors and kill vehicles designed specifically
against even long-burning, liquid-propellent for this mission. It would likely also be desir-
ICBMs like the Hwasong-15, if the ICBMs were able to develop and deploy drones with flight
launched from sites in northwest or north-cen- times longer than current drones and capable
tral North Korea toward cities on the U.S. East of carrying heavier interceptors.
Coast, even if the interceptors were fired with
zero decision time. Depending on their range, Interceptors, kill vehicles, and drones opti-
ICBMs like the Hwasong-15 could also avoid mized for this purpose could be developed
such a defense and strike cities in Alaska or and deployed within the time horizon of
the U.S. Northwest by starting on a trajectory this study, if a decision were made to do so.
toward the U.S. East Coast to avoid intercept Concepts of operation, basing locations, and
and then changing their trajectories. Drone- the number of drones that would be required
based rocket interceptors like these could to defend against a single North Korean ICBM
defend against ICBMs launched from north- or a salvo launch of 10 have not been studied.
west or north-central North Korea, if the ICBMs
were aimed in the direction of targets on the The strategic and arms race implications
U.S. West Coast or were launched from other of developing, testing, and deploying a
sites in North Korea that would make them large system of transportable, high-altitude,
vulnerable to such a defense. long-duration drones armed with high-ve-
locity, highly capable rocket interceptors
If North Korea were eventually to deploy could be profound, unless agreed confi-
solid-propellant ICBMs, which typically have dence-building measures could be devel-
full burn times of 180 seconds or less, they oped and adopted to reassure Russia and
could be launched from many sites in North China that these weapons could only be used
Korea that would prevent them from being to defend against ballistic missiles launched
intercepted by 5 km/s interceptors (see Figure by North Korea.
5(b) above and [APS 2003, Fig. 5.9]).
Using aircraft-based rocket interceptors for
Some proposals for developing and deploy- boost-phase intercept
ing drone-based rocket interceptors quickly
and cheaply have advocated using already A system for boost-phase intercept of North
available, off-the-shelf parts [Garwin 2017; Korean ICBMs that uses fighter aircraft (e.g.,
Garwin 2018a; Garwin 2018b]. However, the F-16s or F-35s) armed with endoatmospheric
particular boosters and kill vehicles that have missiles such as the AIM-260 that can steer
been proposed have burn times so short that only within the atmosphere would require
they could be steered during only a small fighters to operate within 100 to 200 kilo-
meters of the ICBM launch site, hence over

37
North Korean territory, for the missile to reach be needed as well as their construction and
the ICBM before it reaches altitudes greater launch costs. On the other hand, both U.S.
than 30 km, where such interceptors cannot government and nongovernmental sources
operate. Operations with piloted aircraft over assess that North Korea now has, or could field
unfriendly territory inevitably risk pilot capture within the 15-year time horizon of this study,
and serious geopolitical consequences. 10 or more nuclear-armed ICBMs (see Section
Aircraft could be used safely for this purpose 2, North Korea’s ICBM Capabilities). Having
only if the United States has suppressed North to defend against 10 or more nuclear-armed
Korean air defenses. ICBMs is much more challenging than defend-
ing against a single ICBM, which was the
Using space-based rocket interceptors for potential threat considered by the 2003 APS
boost-phase intercept and 2012 National Academies studies [APS
The limitations imposed on the performance 2003; NRC 2012].
of boost-phase intercept systems that use Required size of a space-based intercep-
surface-based interceptors by geographi- tor system. As noted above, to be effective,
cal and geopolitical constraints on intercep- a system of space-based interceptors must
tor basing locations could be sidestepped ensure that at least one would be in range
by placing the interceptors in low Earth orbit. at all times to intercept any ICBM launched
For such a system to be potentially effective, against the United States. We emphasize that
at least one interceptor must be in position to the assumptions used to design such a system
intercept every ICBM that is launched before would need to be conservative, in the sense
the ICBM can give its warhead the velocity that it would need to anticipate the possi-
needed to reach the intended target. But any ble types and performance of North Korea’s
space-based interceptors would continuously ICBMs a decade or more in the future, because
orbit Earth, Earth would be rotating beneath it would take a decade or more to design and
its orbit, and an adversary could launch multi- construct such a system of space-based rocket
ple ICBMs at times of its choosing. There must interceptors and a similar time to significantly
therefore be many interceptors in any such increase its capabilities. One would not want to
system for it to be effective. deploy a system that turns out to be ineffective
In this section we explore the implications of the day it becomes operational.
the differences between the current situation Assuming that the system would not attempt
for space-based interceptors and the situa- to defend any cities in Alaska or in the north-
tion considered by the 2003 APS and 2012 ern parts of the U.S. East and West Coasts or
National Academies studies. On the one hand, in the Midwest, making several other optimis-
North Korea’s current ICBM, the Hwasong-15, tic and simplifying assumptions (see below),
has a full burn time of 290 seconds, signifi- and using the methodology of the 2003 APS
cantly longer than the 240-second burn time Study [APS 2003, Ch. 6], we estimate that if
of the liquid-propellant model ICBM consid- a system were constructed assuming that
ered by these studies. The longer burn time interceptors would be fired almost automat-
reduces the reach-versus-time challenge ically, i.e., with no time allowed for a decision
somewhat, making boost-phase intercept whether to fire once the initial trajectory of the
easier. Also, advances in technology since ICBM has been estimated, a constellation of
those studies were performed have reduced about 1,600 space-based interceptors would
the masses of the interceptors that would

38
need to be deployed to ensure that at least
one would be in position to intercept each of
a rapid “salvo” of four liquid-propellant ICBMs
like the Hwasong-15 launched within three
minutes or so (see Figure 6), and 4,000 would
need to be deployed to attempt to counter a
salvo of 10 such ICBMs. If instead the system
were designed to allow 30 seconds to decide
whether to fire its interceptors, about 2,200
interceptors would be needed to attempt to
counter a rapid salvo of four ICBMs and about
5,500 would be needed to attempt to counter
a rapid salvo of 10.

Orbital motion of the interceptors would

repopulate the coverage that such a constel-
lation would provide on a timescale of
about 200 seconds, so if the defense could Figure 6 View of Earth showing the constellation
be certain that all ICBM launches would be of 1,600 space-based interceptors that would be
spaced at intervals greater than 200 seconds, required to ensure that one is available to intercept
a rapid salvo launch of four Hwasong-15 ICBMs
it could treat multiple launches as a series of
from North Korea, if the system was designed to fire
single launches. A system designed to be able interceptors almost automatically, i.e., if no time is
to defend against only one Hwasong-15, with- allowed to decide whether to fire them. If instead
out allowing any decision time, would need at the system was designed to allow 30 seconds to
least 400 interceptors. At least 500 intercep- decide whether to fire interceptors, about 2,200
tors would be needed if it was constructed to interceptors would be needed to ensure that
enough are available to intercept such a salvo.
allow 30 seconds of decision time.
See text for details. After [NRC 2012, Fig. 2-20].
If North Korea were to eventually deploy
solid-propellent ICBMs, which typically
have burn times of only about 170 seconds, have chosen a final “flyout” velocity of 4 km/s
and a space-based interceptor system was because the 2003 APS study found that for its
constructed to allow no decision time before baseline system, a two-stage interceptor with
its interceptors are fired, a constellation of a flyout velocity of 4 km/s minimized the total
about 16,000 interceptors would be required system mass for a kill vehicle with a 2.5 km/s
to defend against a salvo of 10 ICBMs. In order cumulative divert capability that is capable
to allow 30 seconds of decision time, about of a 15 g acceleration in the endgame of the
36,000 interceptors would be required. intercept and has an interceptor with a total
lag in its response of less than 0.1 seconds
These estimates assume that all interceptors [APS 2003, Sec. 6.5].
are in orbits inclined 45° relative to Earth’s
rotation axis, are distributed roughly uniformly Just like our estimates for sea-, land-, or
over the portion of Earth’s surface that they aircraft-based interceptors defending against
cover, and would have an average accelera- the Hwasong-15, these estimates assume the
tion of 10 g to a final velocity of 4 km/s. We Hwasong-15 could be detected with confi-

39
dence 45 seconds after it was launched. They late as 165 seconds after launch, 10 seconds
also assume that the ICBM’s trajectory would less would be available to defend cities in the
be sufficiently well understood within another northern United States, and 20 seconds less to
20 seconds that a firing solution could be defend cities in Alaska.
constructed, enabling space-based intercep-
tors to be fired 65 seconds after the launch If any additional time is allowed to assess
of the ICBM if the constellation of intercep- whether a launch has occurred, determine
tors is designed to have no decision time whether it is a spoof, better determine the
before interceptors are fired. If the intercep- type of missile, or correct any operational
tors can be fired 65 seconds after the ICBM is errors, the number of interceptors needed
launched, they would be able to reach ICBMs would be correspondingly larger. Additional
about 1,000 km from the position where their interceptors would also be required if they are
launch platform would be 285 seconds after not perfectly reliable or could be defeated by
launch. This is the latest time at which the any of the countermeasures against boost-
ICBM could be intercepted during its longest phase intercept described earlier. The meth-
burn-time trajectories, the trajectories that are odology of the 2003 APS report guarantees
the most favorable for the defense. that there is at least one interceptor in range
for every ICBM at any given time, although
As noted above, these estimates assume that often there is more than one.
the system would not attempt to defend any
cities in Alaska or in the northern parts of the While in orbit, each interceptor would need a
U.S. East and West Coasts or in the Midwest. “lifejacket” or “garage” to provide necessary
To defend cities in the northern United States, services (such as electrical power and commu-
the system would have to be designed to be nications); this would stay behind when the
able to intercept the Hwasong-15 no later interceptor flies out. It may be advantageous
than 275 seconds after it was launched, which to place two interceptors on each orbit-
is 10 seconds earlier than we have assumed in ing platform (“satellite”) to reduce costs and
the estimates cited above. To defend cities in provide some redundancy [APS 2003, Sec.
Alaska, the system would have to be designed 6.3]. If the interceptors are placed in orbits that
to be able to intercept the Hwasong-15 no are only slightly more inclined than the lati-
later than 260 seconds after it was launched. tudes of the required ICBM intercept points,
Constructing a system that could defend the concentration of satellites at latitudes
these targets would require many more inter- close to the orbital inclination [Washburn
ceptors than the estimates provided above. 2013] could in principle allow a reduction in
the number of interceptors required, perhaps
When defending against our model solid-pro- by as much as a factor of two. However, the
pellant ICBM, these estimates assume the substantial spread in the latitudes of the inter-
system could detect the ICBM 30 seconds cept points for ICBMs aimed at different parts
after it was launched and a firing solution of the United States and the inability of the
constructed during the next 15 seconds, so defense to determine the intercept points in
that interceptors could be fired 45 seconds advance may limit this reduction in practice.
after the ICBM was launched if the constel-
lation were constructed to have no decision Cost of a space-based interceptor system. The
time before it fired its interceptors. While we 2003 APS study (see [APS 2003, Table 14.2])
assumed above that intercept is possible as estimated an interceptor mass of 549 kg for

40
an interceptor with a performance compara- tional threat. If North Korea were to deploy
ble to that assumed above using technology solid-propellant ICBMs, the number of inter-
it projected would be available by 2015 [APS ceptors required, and the cost of the system,
2003, Sec. 6.9]. Further advances in electron- would become four times larger.
ics and sensors would almost certainly allow
them to be made even less massive today. Commercial entities have built and launched
Garwin and Postol [Garwin 2017; Garwin space hardware at costs dramatically lower
2018a; Garwin 2018b] have suggested that than those assumed in the 2012 National Acad-
this mass could be reduced by 50% using emies report. However, there is no instance of
current technology. For this report, we assume a DOD procurement taking advantage of such
a more modest 30% mass reduction and economies at the systems level, which would
hence an interceptor mass of about 400 kg, require substantial reductions in the cost of
plus a garage with a mass equal to 50% of space hardware as well as launch costs.
the interceptor mass. Defending against the Commercial launch services have reduced the
launch of a single Hwasong-15 liquid-propel- cost to LEO by a factor of 20, and costs are
lant ICBM using a 400-interceptor constella- expected to continue to decline [Jones 2018].
tion would then require placing about 240 The current cost for launching 23 tonnes
tonnes in low Earth orbit (LEO), while defend- into LEO using a fully expendable Falcon 9
ing against a salvo of 10 Hwasong-15s would rocket is $2,700 per kg, whereas launching
require about 2,400 tonnes in LEO. 63 tonnes into LEO using a fully expendable
Using NRC cost estimates [NRC 2012], the Falcon Heavy rocket costs about $1,400 per
major costs for an initial deployment would kg [Jones 2018]. The latter cost per tonne is
be $19 million to $32 million per tonne for 9 to 16 times smaller than that assumed in
on-orbit hardware and $13 million to $22 the 2012 National Academies study. The cost
million per tonne for launch. This implies per kg using a reusable Falcon Heavy rocket
an initial cost of $8 billion to $13 billion for would undoubtedly be significantly less. The
a system of 400 interceptors designed to Starlink program proposes to launch 12,000
defend against a single Hwasong-15, if the satellites totaling 3,000 tonnes into orbit for
system is designed without allowing any a cost of approximately $10 billion, or about
time to decide whether to fire interceptors, $3 million per tonne for both hardware and
or $100 billion to $180 billion for a system launch costs [Najjar 2020]. Elon Musk states
to defend against a salvo of 10 Hwasong- that each SpaceX Starship rocket will be able
15s, if the system is designed to allow 30 to place 100 tonnes in LEO at an operational
seconds to decide whether to fire intercep- cost of $20 per kg [Bender 2021]. Reductions
tors. There would be additional costs as plat- in launch costs by such large factors could
forms are replaced over the lifetime of the drive down the costs of space-based inter-
system. If this estimate holds, even to within ceptors by an order of magnitude or more.
a factor of 10, the cost of space-based inter- However, in commercial space activities such
ceptors is highly unfavorable to the defense. economies of scale often come with built-in
The offense can add one more ICBM to a reduced reliability, and if so it is not clear that
salvo launch, at about $20 million in 2021 this increased risk would be acceptable for a
dollars (based on U.S. Minuteman III costs missile defense system that must work with
[MMIII Costs 2015]), driving the defense to extremely high reliability.
spend 1,000 times more to match the addi-

41
Countermeasures to space-based intercep- ICBM’s trajectory. According to the 2003 APS
tor systems. Besides the countermeasures study, a properly focused 3 MW laser weapon
to boost-phase missile defense already illuminating an ICBM at an altitude greater
described, a space-based system would likely than 60 km for 5 - 20 seconds could disable a
be vulnerable to interference, damage, or liquid-propellant ICBM at a range up to about
destruction by anti-satellite weapons, and 600 km and a solid-propellant ICBM at a range
might be attacked or sabotaged when inter- up to about 300 km. These ranges could allow
ceptors are first orbited, to prevent an effec- an aircraft carrying the laser to operate 100
tive system from being assembled. km outside North Korean airspace [APS 2003,
Sec. 7.3]. This is the performance that was
Other disadvantages of space-based inter- planned for the laser and optics carried by the
ceptor systems. The large constellation of YAL-1 Airborne Laser aircraft [APS 2003, Sec.
orbiting satellites required for a space-based 21; NRC 2012, 2-20]. According to Depart-
interceptor system may be threatening in and ment of Defense officials, current lasers are
of itself, since these weapons would essen- very far from meeting these performance
tially blanket the sky (see Figure 6). A system requirements [Hill 2020b; Mehta 2020].
designed to defend against ICBMs launched
from North Korea would also threaten China’s Efforts to develop and deploy destructive laser
strategic nuclear forces. If all the intercep- weapons are advancing slowly. While MDA has
tors were in orbits with inclinations less than backed away from developing defensive laser
45°, they would not threaten ICBMs launched weapons, various branches of the U.S. mili-
from Russia’s current launch sites, but such a tary have continued to pursue this technology
system could readily be expanded to cover [Judson 2021b]. In 2021, the U.S. army demon-
them. With their high burnout speeds and strated a 50-kilowatt laser on a combat vehi-
ability to maneuver, space-based interceptors cle intended for short-range air defense, and
would be potent anti-satellite weapons that in 2022 it is expected to acquire a 300-kilowatt
could potentially reach all satellites, includ- technology demonstrator to explore using a
ing those in geosynchronous orbits [Wright laser to defend fixed and semi-fixed sites
2002]. Fielding space-based interceptors— against cruise missiles, unmanned aircraft and
even just a few in the guise of a testbed—could rockets, as well as artillery and mortars. The
drive a significant weaponization of space U.S. navy is currently deploying the HELIOS
and threaten potential adversaries’ sensitive system on some U.S. destroyers, but it is only
national security satellites. Developing and destructive at short ranges against relatively
testing such a system, let alone deploying it, soft targets, such as rubber dinghies [Eckstein
would therefore have major negative strategic 2021a; Kubovich 2020]. Israel is developing a
and arms race implications. 100-kilowatt ground-based laser system that it
hopes will be able to destroy targets at a range
Using laser weapons for boost-phase of eight to 10 kilometers [Ahronheim 2021].
intercept
There is widespread agreement that laser
Practical laser weapons for boost-phase inter- weapons that could disable or destroy ICBMs
cept would require laser weapons systems during their boost-phase, whether based on
compact and light enough to be carried on aircraft, drones, or space platforms, will not
an aircraft or drone, but powerful enough and be technically feasible within the 15-year time
well enough focused to be able to disable an horizon of this study (see [Hill 2020b]).
ICBM at a realistic standoff distance from the

42
6. CLOSING REMARKS

This report has used publicly available There are also important non-technical ques-
information to consider whether currently tions that we have only been able to touch on
deployed and proposed future U.S. missile briefly but deserve more extensive consid-
defense systems could successfully defend eration and assessment. These include the
the continental United States against an attack strategic costs and benefits of deploying a
by a limited threat: North Korea’s current and missile defense system that is only partially
near-term nuclear-armed ICBM force. Consid- effective against nuclear-armed ICBMs; the
ering these systems in the context of this very security costs and benefits of pursuing missile
limited threat has revealed not only the key defense efforts relative to pursuing diplo-
technical challenges that would have to be matic and arms control efforts; the effects of
surmounted to address this particular threat, the U.S. missile defense program on the like-
but also the technical challenges that would lihood that potential adversaries will develop
have to be overcome to address any other more numerous and advanced offensive
possible limited ICBM threats that may arise nuclear weapons and defensive systems; and
in the future. Considering the limited threat the economic and social costs of devoting
posed by North Korea’s ICBMs now and in the very large resources to missile defense
the near term has also brought out several that would be required to continue, let alone
broader questions that arise whenever efforts expand, the current program.
to create a defense against nuclear-armed
ICBMs are examined. Nevertheless, there Rather than addressing these and other
are many technical and non-technical ques- important but very broad questions, this brief
tions about missile defense systems that are report focused on the fundamental question of
outside the primary focus of this study. whether current or proposed missile defense
systems could defend the continental United
On the technical side, we have not discussed States against a baseline threat consisting of
how North Korea’s nuclear-armed ICBM capa- a single nuclear-armed ICBM launched from
bility might evolve beyond the 15-year time North Korea, or a salvo of 10 ICBMs launched
horizon of this study, or whether other coun- in rapid succession (see “North Korea’s ICBM
tries might develop a similar ICBM capability Capabilities” section above), once they are
in the future. One would need accurate fore- launched. We discussed the myriad chal-
casts of the longer-term evolution of these lenges involved in defending against even one
and other possible nuclear-weapon capabil- ICBM, challenges that include various possi-
ities and the longer-term evolution of missile ble countermeasures to the defensive system
defense technologies to be able to judge that North Korea could employ (see the “Chal-
whether defensive systems could meaning- lenges of Missile Defense” section and the
fully defend against these potential future more detailed discussions in the “Midcourse
threats. We have also not considered what Intercept Systems” and “Boost-Phase Intercept
defensive systems, if any, could meaningfully Systems” sections that follow it).
defend against the much more numerous and
sophisticated nuclear-armed ICBMs and other We described the U.S. missile defense
nuclear forces of China and Russia. systems that have already been deployed,

43
are currently being considered, or have been What we found is that creating a reliable
proposed to defend against nuclear-armed and effective defense against even the small
ICBMs. These systems fall into two main cate- number of relatively unsophisticated nucle-
gories: midcourse intercept systems and ar-armed ICBMs that we considered remains
boost-phase intercept systems. The two main a daunting challenge. The difficulties are
sections of the report—“Midcourse Inter- numerous, ranging from the unresolved
cept Systems” and “Boost-Phase Intercept countermeasures problem for midcourse
Systems”—summarize what is publicly known intercept to the severe reach vs. time problem
about the current status, hoped-for capabili- of boost-phase intercept. In addition to many
ties, and future prospects of these two types shared challenges, each system has its own
of systems. Examples of these systems include unique difficulties that must be overcome. We
the GMD midcourse intercept system, the have detailed these in the “Midcourse Inter-
Aegis BMD system when used for midcourse cept” and “Boost-Phase Intercept” sections of
intercept, and the drone-based rocket-in- the report.
terceptor system that has been proposed
for boost-phase intercept. We explained Our survey of the literature and our analysis
the current and near-term abilities of these of published work has led us to conclude that
systems to address the baseline threat and few of the main challenges involved in devel-
the increased threat that can reasonably be oping and deploying a reliable and effective
expected within the 15-year time horizon of ballistic missile defense have been solved,
this report. and that many of the hard problems we have
identified are likely to remain unsolved during,
and probably beyond, the 15-year time hori-
zon we considered.

44
STUDY GROUP MEMBERS

Frederick K. Lamb (Study Group Chair) is he co-led the 2018 POPA study “Neutrons for
Research Professor of Physics, Brand and the Nation: Discovery and Applications while
Monica Fortner Chair of Theoretical Astro- Minimizing the Risk of Nuclear Proliferation”.
physics emeritus, and a core faculty member
in the Program in Arms Control & Domestic David K. Barton is an independent consul-
and International Security at the University tant. A member of the National Academy of
of Illinois. A member of the American Acad- Engineering, he is also a Fellow of the Insti-
emy of Arts and Sciences and a Fellow of the tute of Electrical and Electronics Engineers
American Physical Society (APS), his scientific (IEEE). He is an expert on the use of radars for
research has focused on high-energy and rela- missile launch detection, sensor cuing, missile
tivistic astrophysics. He has been a consultant tracking, and target discrimination, and has
to the Defense Department, national labora- served as a member of the NRC’s Air Force
tories, Congressional committees, and the Studies Board, as chair of the Committee on
Institute for Defense Analyses on defense and the E-3A Radar, and as chair of the Committee
security matters, including space policy, ballis- on Advanced Airborne Surveillance Radar. He
tic missiles and missile defenses, and techni- was a member of the 2003 American Physical
cal aspects of nuclear test bans. He co-chaired Society and the 2012 National Academy of
the 2003 APS study of boost-phase missile Sciences panels on missile defense.
defense and shared the American Physical Philip E. Coyle was a Senior Science Fellow
Society’s 2005 Leo Szilard Lectureship Award. at the Center for Arms Control and Non-Pro-
Laura Grego (Study Group Co-chair) is a Stan- liferation. A renowned expert on nuclear
ton Nuclear Security Fellow at MIT’s Labora- weapons, missile defense, national secu-
tory for Nuclear Security and Policy, on leave rity, and defense spending, his recent work
from  the Union of Concerned Scientists’ focused heavily on analyses of ground-based
Global Security Program, where she is senior midcourse defense. He spent more than
scientist and research director. A physicist three decades working on nuclear weapons
by training, she works at the intersection of and related programs at Lawrence Livermore
science and policy on the topics of nuclear National Laboratory, becoming Deputy Labo-
weapons, missile defense, and space secu- ratory Director, before serving as Assistant
rity. She is an associate editor for the journal Secretary of Defense and Director, Opera-
Science and Global Security, and a Fellow of tional Test & Evaluation and subsequently as
the American Physical Society and a member Associate Director for National Security and
of its Panel on Public Affairs. International Affairs in the White House Office
of Science and Technology Policy.
James D. Wells (Study Group Co-chair) is a
Professor of Physics at the University of Michi- Steve Fetter is a professor in the School of
gan. A Fellow of the American Association for Public Policy at the University of Maryland. He
the Advancement of Science and the Amer- is a Fellow of the American Physical Society,
ican Physical Society, his research interests a current member of the APS Panel on Public
include theoretical high-energy physics and Affairs, and a recipient of the APS Leo Szilard
the intersections of science and national secu- Lectureship and Joseph A. Burton Forum
rity policy. A former chair of the American Phys- awards. He has served in the departments of
ical Society’s Panel on Public Affairs (POPA), State and Defense and led the national secu-

45
rity and international affairs division in the tist of Los Alamos’ Threat Reduction director-
White House Office of Science and Technology ate, was a member of the 2003 APS study of
Policy. He is a member of the National Acad- boost-phase missile defense, and shared the
emy of Sciences’ Committee on International American Physical Society’s 2005 Leo Szilard
Security and Arms Control and has contrib- Lectureship Award.
uted to 10 reports by the National Academies.
Jaganath Sankaran is an assistant professor
Alec Gallimore is an Arthur F. Thurnau Profes- in the Lyndon B. Johnson School of Public
sor in the Department of Aerospace Engi- Affairs at The University of Texas at Austin.
neering, the Richard F. and Eleanor A. Towner Formerly a defense scientist with the Indian
Professor of Engineering, and the Robert J. Missile R&D establishment, his current
Vlasic Dean of Engineering at the University research focuses on missile defenses, space
of Michigan. A member of the National Acad- weapons, military net assessment, and arms
emy of Engineering and a Fellow of the Amer- control. He has held fellowships at the Los
ican Institute of Aeronautics and Astronautics, Alamos National Laboratory, the Belfer
he is an expert on electric propulsion, rock- Center for Science and International Affairs
etry, plasma physics and advanced plasma at Harvard University, and the RAND Corpo-
diagnostics. He is a past member of the U.S. ration, and has served on study groups of
Air Force Scientific Advisory Board and has the National Academy of Sciences and the
served on numerous NASA and DOD boards American Physical Society, examining missile
and studies. defenses and strategic stability.
George N. Lewis is a physicist and an inde- Aric Tate is a physicist who is currently a PhD
pendent consultant. He is an expert on tech- student in Nuclear, Plasma, and Radiological
nical aspects of nuclear weapons, delivery Engineering at the University of Illinois. His
systems, missile defense, and arms control. research includes developing instruments to
After 16 years in the Technology and Secu- detect and track nuclear explosive material
rity Group of MIT’s Security Studies Program, as well as radiation hard detectors for use at
he became a senior research associate at the particle accelerators. For the last five years
Judith Reppy Institute for Peace and Conflict he has helped teach the University of Illinois
Studies at Cornell University. He is a Fellow of physics course “Nuclear Weapons and Arms
the American Physical Society and a recipient Control”.
of its Joseph A. Burton Forum Award.
Frank von Hippel is Professor of Public and
Cynthia K. Nitta is the Chief Program Advisor International Affairs emeritus at Princeton
for the Future Deterrent at Lawrence Liver- University, co-director emeritus of Prince-
more National Laboratory. She has an exten- ton’s Program in Science & Global Security,
sive background in weapon physics and and a member and founding co-chair of the
design and strong interest in nuclear weap- International Panel on Fissile Materials. A
ons policy. former Assistant Director for National Secu-
rity in the White House Office of Science
William Priedhorsky is Director of the Labo- and Technology, his areas of policy research
ratory-Directed (internal) R&D Program at Los focus on nuclear arms control and nonpro-
Alamos National Laboratory. A Los Alamos liferation. Prior to joining the faculty at Princ-
and American Physical Society (APS) Fellow, eton, he worked for 10 years in the field of
he has a broad range of expertise, including theoretical elementary-particle physics. He
on space hardware and space-based sensors. was awarded a MacArthur Fellowship in 1993
He was previously program manager for prolif- and the American Physical Society’s 2010 Leo
eration detection, later served as chief scien- Szilard Lectureship Award.

46
FIGURES

Figure 1............................................................................................................................. page 5

View of Earth showing three illustrative ICBM trajectories from North Korea to targets in
the United States.

Figure 2............................................................................................................................. page 6

Schematic portrayal of the missile defense system elements being proposed to defend
the United States against ICBMs launched from North Korea.

Figure 3.......................................................................................................................... page 21

Sequence of events in an intercept attempt by the GMD system.

Figure 4.......................................................................................................................... page 31

Map showing North Korea and adjacent countries and the initial ground tracks of ICBMs
launched from north-central North Korea to five cities in the United States.

Figure 5.......................................................................................................................... page 32

Possible basing areas for the interceptors discussed in the text to be able to reach liq-
uid- and solid-propellant ICBMs launched from North Korea in time to prevent them from
delivering their warhead to the indicated target cities.

Figure 6.......................................................................................................................... page 39

View of Earth showing the constellation of 1,600 space-based interceptors that would be
required to ensure that at least one is available to intercept each Hwasong-15 ICBM in a
rapid salvo launch of four from North Korea, if the system is designed to fire interceptors
almost automatically.

47
REFERENCES

[Abott 2018] Abott, Richard. 2018. “Final NDAA Keeps Annual%20Threat%20Assessment%20Statement%20

Space and Boost-Phase Missile Defenses.” Defense for%20the%20Record.pdf.
Daily, July 27. https://www.defensedaily.com/final-ndaa-
keeps-space-boost-phase-missile-defenses/budget/. [BMD Expenditures 2021] U.S. Expenditures on Ballistic
Missile Defense Through Fiscal Year 2021. Prepared for
[Abott 2021] Abott, Richard. 2021.“MDA Demos Jap- the APS Study Group on Ballistic Missile Defense, Octo-
anese Aegis Software on SPY-7 Radar.” Defense Daily, ber 26, 2021. https://doi.org/10.5281/zenodo.5748223
February 26. https://www.defensedaily.com/mda-dem-
os-japanese-aegis-software-spy-7-radar-aegis-ashore- [Boese 1999] Boese, Wade. 1999. “NMD Program Again
cancelled/navy-usmc/ Criticized by Independent Review Panel.” Arms Control
Association, November. https://www.armscontrol.org/
[Ahronheim 2021] Ahronheim, A., “Lockheed Martin, act/1999-11/press-releases/nmd-program-again-criti-
Rafael Advanced Defense Systems to make new laser cized-independent-review-panel.
weapon”, The Jerusalem Post, July 27, 2021. https://
www.jpost.com/israel-news/lockheed-martin-rafa- [Bush 2002] Bush, George W. 2002. “National Policy on
el-advanced-defense-systems-create-new-laser-sys- Ballistic Missile Defense.” National Security Presidential
tem-675034. Directive/NSPD-23. The White House. December 16.
http://fas.org/irp/offdocs/nspd/nspd-23.htm.
[APS 2003] Barton, David K., et al. 2003. “Report of the
American Physical Society Study Group on Boost-Phase [Butler 2014] Butler, Amy. 2014. “Fiscal ’15 Funds to
Intercept Systems for National Missile Defense: Scien- Counter GMD ‘Bad Engineering.’” Aviation Week &
tific and Technical Issues.” 2003 APS report. Reviews Space Technology, February 25.
of Modern Physics 76, no. S1. https://doi.org/10.1103/
RevModPhys.76.S1. [Butt 2011] Butt, Yousaf, and Theodore Postol. 2011.
“Upsetting the Reset: The Technical Basis of Russian
[Baklitskiy 2021] Baklitskiy, Andrey, James Cameron, Concern Over NATO Missile Defense.” Federation of
and Steven Pifer. 2021. “Missile Defense and the Of- American Scientists, September 1. https://www.jstor.
fense-Defense Relationship.” Deep Cuts Working Paper org/stable/resrep18923.
no. 14, October 2021. https://deepcuts.org/images/
PDF/DeepCuts_WP14.pdf. [Canavan 2003] Canavan, Gregory H. 2003. Missile
Defense for the 21st Century. The Heritage Foundation.
[Bartels 2017] Bartels, Elizabeth M. 2017. “Is Iron Dome http://highfrontier.org/wp-content/uploads/2013/07/
a Poisoned Chalice? Strategic Risks from Tactical Suc- Canavan-Heritage.pdf.
cess.” TheRANDBlog, November 29. https://www.rand.
org/blog/2017/11/is-iron-dome-a-poisoned-chalice- [Caveny 2003] Caveny, Leonard H., Robert L. Geisler,
strategic-risks-from.html. Russell A. Ellis, and Thomas L. Moore. 2003. “Solid
Rocket Enabling Technologies and Milestones in the
[Belk 2018] Belk, Clara. 2018. “Iran’s Ballistic Missile United States.” Journal of Propulsion and Power 19, no.
Inventory.” Atlantic Council, IranSource, December 6. https://doi.org/10.2514/2.6944.
3. https://www.atlanticcouncil.org/blogs/iransource/
iran-s-ballistic-missile-inventory/. [CBO 2021] Congressional Budget Office. 2021. “Costs
of Implementing Recommendations of the 2019 Missile
[Bender 2021] Bender, Maddie. 2021. “SpaceX’s Defense Review.” January 13. https://www.cbo.gov/
Starship Could Rocket-Boost Research in Space.” Sci- publication/56949.
entific American, September 16. https://www.scientifi-
camerican.com/article/spacexs-starship-could-rocket- [Cerniello 1998] Cerniello, Craig. 1998. “Panel Criti-
boost-research-in-space/. cizes U.S. ‘Rush’ to Deploy Missile Defense Systems.”
Arms Control Today, March. https://www.armscontrol.
[Bennett 2021] Bennett, Bruce W., Kang Choi, Myong- org/act/1998-03/arms-control-today/panel-criticiz-
Hyun Go, Bruce E. Bechtol Jr., Jiyoung Park, Bruce es-us-rush-deploy-missile-defense-systems.
Klingner, and Du-Hyeogn Cha. 2021. “Countering the
Risks of North Korean Nuclear Weapons.” RAND Report [Cha 2020] Cha, Sangmi. 2020. “North Korea Building
PEA1015-1, The RAND Corporation, April 2021. https:// Two Submarines, One Capable of Firing Ballistic Mis-
www.rand.org/pubs/perspectives/PEA1015-1.html. siles - Lawmaker.” Reuters, November 1. https://www.
reuters.com/article/us-northkorea-missiles/north-ko-
[Berrier 2021] Berrier, Scott. 2021. Worldwide Threat rea-building-two-submarines-one-capable-of-firing-bal-
Assessment. Armed Services Committee. United States listic-missiles-lawmaker-idUSKBN27J0MC.
Senate. 117th Cong. (statement of Scott Berrier, director,
Defense Intelligence Agency). https://www.armed-ser- [Choe 2021a] Choe, Sang-Hun. 2021a. “North Korea’s
vices.senate.gov/imo/media/doc/2021%20DIA%20 Arsenal Has Grown Rapidly. Here’s What’s in It.” New

48
York Times, October 18. https://www.nytimes.com/arti- exoatmospheric-kill-vehicle-quality-assurance-and-reli-
cle/north-korea-arsenal-nukes.html. ability-assessment-part-a/.

[Choe 2021b] Choe, Sang-Hun. 2021b. “North Korea [DOD 2019] Department of Defense. 2019b. Military
Tests First Submarine-Launched Missile in Two Years.” and Security Developments Involving the People’s
New York Times, October 18. https://www.nytimes. Republic of China 2019, Annual Report to Congress.
com/2021/10/18/world/asia/north-korea-ballistic-mis- Washington, DC: Office of the Secretary of Defense: 67.
sile-test-washington.html. https://media.defense.gov/2019/May/02/2002127082/-
1/-1/1/2019_CHINA_MILITARY_POWER_REPORT.pdf.
[Choi 2021] Choi Soo-hyang. 2021. “N. Korea’s ‘Hyper-
sonic Missile’ Appears to Be at Early Stage of Devel- [DOD 2020a] Department of Defense. 2020a. North Ko-
opment: JCS.” Yonhap News Agency, September 29. rean Tactics. Publication ATP 7-100.2. U.S. Department
https://en.yna.co.kr/view/AEN20210929005552325?- of Defense, Department of the Army, July 24: 1-11,
section=nk/nk. paragraph 1-45. https://irp.fas.org/doddir/army/atp7-
100-2.pdf
[Cohen 2019] Cohen, Rachel S. 2019. “SDA’s first five-
year plan looks to spend $525 million on LEO layer.” Air [DOD 2020b] U.S. Department of Defense. 2020b.
Force Magazine. March 19. https://www.airforcemag. “U.S. Successfully Conducts SM-3 Block IIA Intercept
com/sdas-first-five-year-plan-looks-to-spend-525-mil- Test Against an Intercontinental Ballistic Missile Target”
lion-on-leo-layer. (press release). November 17. https://www.defense.
gov/Newsroom/Releases/Release/Article/2417334/
[Cordesman 2019] Cordesman, Anthony H. 2019. “The us-successfully-conducts-sm-3-block-iia-inter-
Iranian Missile Challenge.” Center for Strategic and cept-test-against-an-intercontinen/.
International Studies, May 30. https://www.csis.org/anal-
ysis/iranian-missile-threat. [DOD 2021] Department of Defense. 2021. Department
of Defense Fiscal Year (FY) 2022 Budget Estimates: De-
[Cropsey 2021] Cropsey, Seth. 2021. “Hypersonic fense-Wide Justification Book Volume 2a of 5. Depart-
Weapons Could Tilt War in Favor of Russia, China.” The ment of Defense, May. https://comptroller.defense.gov/
Hill, August 5. https://thehill.com/opinion/national-se- Portals/45/Documents/defbudget/fy2022/budget_jus-
curity/566534-hypersonic-weapons-could-tilt-war-in-fa- tification/pdfs/03_RDT_and_E/RDTE_Vol2_MDA_RDTE_
vor-of-russia-china. PB22_Justification_Book.pdf.
[CRS 2021a] Congressional Research Service. 2021. North [Dolan 1972] Dolan, Philip J. 1972. Capabilities of Nu-
Korea’s Nuclear Weapons and Missile Programs. IF10472, clear Weapons, Part One: Phenomenology. Washington,
April 14. https://sgp.fas.org/crs/nuke/IF10472.pdf. D.C.: Defense Nuclear Agency, July.
[CRS 2021b] Congressional Research Service. 2021. [Dominguez 2019] Dominguez, Gabriel. 2019. “USFK
Navy Aegis Ballistic Missile Defense (BMD) Program: confirms North Korea's Hwaseong-15 ICBM can target
Background and Issues for Congress. RL33745, October all of US mainland.” Janes, July 11. https://www.janes.
20, 2021. https://sgp.fas.org/crs/weapons/RL33745.pdf. com/defence-news/news-detail/usfk-confirms-north-ko-
reas-hwaseong-15-icbm-can-target-all-of-us-mainland.
[CSIS 2021] Center for Strategic and International Stud-
ies. 2021. “Missiles of North Korea.” June 14. https:// [DOT&E 2015] Director, Operational Test and Evaluation
missilethreat.csis.org/country/dprk/. (DOT&E). 2015. 2014 Assessment of the Ballistic Missile
Defense System (BMDS). Washington, DC: Department
[DIA 2021] Defense Intelligence Agency. 2021. North of Defense, March 2015. https://apps.dtic.mil/sti/pdfs/
Korea Military Power. U.S. Defense Intelligence Agency ADA617330.pdf
report, October 15, 2021. https://www.dia.mil/Por-
tals/110/Documents/News/North_Korea_Military_Pow- [DOT&E 2021] Director, Operational Test and Evalua-
er.pdf. tion (DOT&E). 2021. Ground-based Midcourse Defense
(GMD). Washington, DC: Department of Defense.
[DNA 1963] Defense Nuclear Agency. 1963.“High https://www.dote.osd.mil/Portals/97/pub/reports/
Altitude Nuclear Weapon Effects, Part Two: Systems FY2020/bmds/2020gmd.pdf?ver=PrJwoGPMLLrtA9dk-
Interference.” Fort Belvoir, VA: Defense Nuclear Agency. kEQU-g%3d%3d.
https://www.youtube.com/watch?v=makck0k9zgo.
[DSB 2011] Defense Science Board. 2011. Defense
[DOD 1977] U.S. Department of Defense and Depart- Science Board Task Force Report on Science and Tech-
ment of Energy; Samuel Glasstone and Philip J. Dolan. nology Issues of Early Intercept Ballistic Missile De-
1977. The Effects of Nuclear Weapons. Washington, DC: fense Feasibility. Washington, DC: Office of the Under
U.S. Department of Defense. https://www.dtra.mil/Por- Secretary of Defense for Acquisition, Technology, and
tals/61/Documents/NTPR/4-Rad_Exp_Rpts/36_The_Ef- Logistics, September. https://apps.dtic.mil/sti/pdfs/
fects_of_Nuclear_Weapons.pdf. ADA552472.pdf.
[DOD 2014] Inspector General, Department of Defense. [Eckstein 2021a] Eckstein, Megan. 2021a. “Navy
2014. “Exoatmospheric Kill Vehicle Quality Assurance Installing More Directed Energy Weapons on DDGs,
and Reliability Assessment—Part A.” DODIG-2014-111. Conducting Land-Based Laser Testing This Year.” USNI
September 8. https://www.dodig.mil/Reports/Compen- News, April 7. https://news.usni.org/2021/04/07/navy-
dium-of-Open-Recommendations/Article/1119057/

49
installing-more-directed-energy-weapons-on-ddgs-con- ment Accountability Office, June. https://www.gao.
ducting-land-based-laser-testing-this-year. gov/assets/700/699584.pdf.

[Eckstein 2021b] Eckstein, Megan. 2021b. “MDA: Test [GAO 2019b] Government Accountability Office.
of DDG, Standard Missile-3 IIA a Good Start, But More 2019b. Missile Defense: Further Collaboration with the
Work Needed on Homeland Defense Mission.” USNI Intelligence Community Would Help MDA Keep Pace
News, May 13. https://news.usni.org/2021/05/13/mda- with Emerging Threats. GAO-20-177. Washington, DC:
test-of-ddg-standard-missile-3-iia-a-good-start-but- Government Accountability Office, December. https://
more-work-needed-on-homeland-defense-mission. www.gao.gov/assets/710/706999.pdf.

[Einhorn 2019] Einhorn, Robert, and Vann H. Van [GAO 2020] Government Accountability Office. 2020.
Diepen. 2019. Constraining Iran’s Missile Capabilities. Missile Defense: Assessment of Testing Approach
Brookings Institution, March. https://www.brookings. Needed as Delays and Changes Persist. GAO-20-432.
edu/wp-content/uploads/2019/03/FP_20190321_mis- Washington, DC: Government Accountability Office,
sile_program_WEB.pdf. July. https://www.gao.gov/products/gao-20-432.

[Elleman 2017] Elleman, Michael. 2017. “The Pukguk- [GAO 2021] Government Accountability Office. 2021.
song-2: Lowering the Bar on Combat Readiness?” 38 Missile Defense. Fiscal Year 2020 Delivery and Testing
North, May 25. https://www.38north.org/2017/05/puk- Progressed, but Annual Goals Unmet. Washington, DC:
guksong2_052517/. Government Accountability Office, April. https://www.
gao.gov/assets/gao-21-314.pdf.
[Elleman 2021a] Elleman, Michael. 2021a.“Iran’s Ballistic
Missile Program.” The Iran Primer, January 13. https:// [Garwin 1968] Garwin, Richard L., and Hans A. Bethe.
iranprimer.usip.org/resource/irans-ballistic-missile-pro- 1968. “Anti-Ballistic-Missile Systems.” Scientific
gram. American 218, no. 3 (March). https://www.jstor.org/
stable/24925996?refreqid=excelsior%3A8fcd2cd4fdc-
[Elleman 2021b] Elleman, Michael. 2021b. “North Ko- 6cfd4ddcb20407aa4c4b0.
rea’s Newest Submarine-Launched Ballistic Missile.” 38
North, January 15. https://www.38north.org/2021/01/ [Garwin 2017] Garwin, Richard L., and Theodore A.
north-koreas-newest-submarine-launched-ballistic-mis- Postol. 2017. “Airborne Patrol to Destroy DPRK ICBMs in
sile-same-as-the-old-one/. Powered Flight.” Presented in Washington, DC, Novem-
ber 27 - 29. https://fas.org/rlg/airborne.pdf.
[Erästö 2021] Erästö, Tytti, and Matt Korda. 2021. “Time
to Factor Missile Defence into Nuclear Arms Control [Garwin 2018a] Garwin, Richard L. and Theodore A. Pos-
Talks.” SIPRI, September 30. https://www.sipri.org/com- tol. 2018a. “Technical Refinements in Design Features
mentary/topical-backgrounder/2021/time-factor-mis- of the Airborne Patrol Against North Korean ICBMs.”
sile-defence-nuclear-arms-control-talks. Cambridge, MA: Massachusetts Institute of Technology,
May 10. https://fas.org/rlg/refine.pdf.
[Fisher 2017] Fisher, Max, Eric Schmitt, Audrey Carlsen,
and Malachy Browne. 2017. “Did American Missile De- [Garwin 2018b] Garwin, Richard L. and Theodore A. Pos-
fense Fail in Saudi Arabia?” New York Times, December tol. 2018b. “New Interceptor Options for Consideration
4. https://www.nytimes.com/interactive/2017/12/04/ in the Airborne Patrol Against DPRK ICBMs.” June 25.
world/middleeast/saudi-missile-defense.html.
[Gilmore 2015] Gilmore, J. Michael. 2015. Hearing of
[Gallo 2021] Gallo, William. 2021.“North Korea Tested the Strategic Forces Subcommittee, Senate Armed
a Hypersonic Missile. Here’s Why It Matters.” Voice of Services Committee. 114th Cong. (statement of J.
America, September 29. https://www.voanews.com/a/ Michael Gilmore, director, Operational Test and Eval-
north-korea-tested-a-hypersonic-missile-here-s-why-it- uation, Office of the Secretary of Defense). https://
matters-/6249922.html. www.armed-services.senate.gov/imo/media/doc/Gilm-
ore_03-25-15.pdf.
[GAO 2017] Government Accountability Office. 2017.
Missile Defense: Some Progress Delivering Capabilities, [GMD Tests 2021] “Ground-Based Intercept (GBI)
but Challenges with Testing Transparency and Require- Intercept Tests,” prepared by George Lewis and Laura
ments Development Need to Be Addressed. GAO-17- Grego for the APS Study Group on Ballistic Missile
381. Washington, DC: Government Accountability Office, Defense, October 26, 2021. https://doi.org/10.5281/
May. https://www.gao.gov/assets/gao-17-381.pdf. zenodo.5748229.

[GAO 2018] Government Accountability Office. 2018. [Goodby 2018] Goodby, James E., and Theodore A.
Missile Defense: The Warfighter and Decision Makers Postol. 2018. A new boost-phase missile defense sys-
Would Benefit from Better Communication about the tem—and its diplomatic uses in the North Korea dispute,
System’s Capabilities and Limitations. GAO-18-324. Bulletin of the Atomic Scientists 74, no. 4: 210 - 219.
Washington, DC: Government Accountability Office, https://doi.org/10.1080/00963402.2018.1486578.
May. https://www.gao.gov/assets/gao-18-324.pdf.
[Grego 2016] Grego, Laura, George N. Lewis, and David
[GAO 2019a] Government Accountability Office. Wright. 2016 Shielded from Oversight: The Disastrous
2019a. Missile Defense: Delivery Delays Provide Op- US Approach to Strategic Missile Defense. Union of
portunity for Increased Testing to Better Understand Concerned Scientists, July. http://ucsusa.org/shielded-
Capability. GAO-19-387. Washington, DC: Govern- fromoversight.

50
[Hambling 2021] Hambling, David. 2021. “Under The cific-missile-defense-radars-nixed-mda-to-study-sensor-
Iron Dome: The Problem With Israel’s Rocket Shield.” architecture-in-region/.
Forbes, May 12. https://www.forbes.com/sites/da-
vidhambling/2021/05/12/under-the-iron-dome-the- [Judson 2021a] Judson, Jen. 2021. “Next-Gen Inter-
problem-with-israels-rocket-shield/?sh=200543db59b8. continental Ballistic Missile Interceptor Estimated Cost?
Nearly $18B.” Defense News, April 27. https://www.
[Hansen 2020] Hansen, Nick, and Nathan Hunt. 2020. defensenews.com/pentagon/2021/04/27/next-gen-in-
“North Korea Showcases World’s Largest Mobile ICBM.” tercontinental-ballistic-missile-interceptor-estimat-
Janes, December 3, 2020. https://www.janes.com/de- ed-to-cost-nearly-18-billion/.
fence-news/news-detail/north-korea-showcases-worlds-
largest-mobile-icbm. [Judson 2021b] Judson, Jen. 2021. “Congress Gives
Missile Defense Agency Authority to Research and
[Hecker 2021] Hecker, Siegfried. 2021. “Estimating Develop Laser Tech for Missile Defense.” Defense
North Korea’s Nuclear Stockpiles: An Interview With News, December 30. https://www.defensenews.com/
Siegfried Hecker.” Interview, 38 North, April 30, 2021. pentagon/2021/12/30/congress-gives-missile-defense-
https://www.38north.org/2021/04/estimating-north-ko- agency-authority-to-research-and-develop-laser-tech-
reas-nuclear-stockpiles-an-interview-with-sieg- for-missile-defense/
fried-hecker/.
[Kim 2021a] Kim, Lami. 2021a. “A Race for Nuclear-Pow-
[Hill 2020a] Hill, Jon. 2020a. Keynote address in “Army ered Submarines on the Korean Peninsula.” National
Space and AMD: Protection of the U.S. Homeland, Forc- Bureau of Asian Research, March 31. https://www.nbr.
es Abroad, Allies and Partners,” Arlington, VA, March org/publication/a-race-for-nuclear-powered-subma-
5 (slide in video at 20 minutes). https://www.youtube. rines-on-the-korean-peninsula/.
com/watch?v=6qgjMka8FMA&t=133s.
[Kim 2021b] Kim, Hyung-Jin. 2021b. “N. Korea Says It
[Hill 2020b] Hill, Jon A. 2020b. “Virtual Event: Homeland Fired Anti-Aircraft Missile, 4th Recent Test.” Washington
Missile Defense: Plotting a Clear Path Forward.” The Post, October 2. https://www.washingtonpost.com/
Heritage Foundation, August 18. https://www.heritage. world/north-korea-says-it-has-test-fired-an-anti-air-
org/missile-defense/event/virtual-event-homeland-mis- craft-missile/2021/09/30/ce705300-223a-11ec-a8d9-
sile-defense-plotting-clear-path-forward. 0827a2a4b915_story.html.

[Hyten 2020] “I have 100 percent confidence. Literally – I [Kim 2021c] Kim, Tong-Hyung, Hyung-Jin Kim, and
don’t say 100 percent very often. I have 100 percent Mari Yamaguchi. 2021. “North Korea Tests Possible
confidence in those capabilities against North Korea.” Submarine Missile Amid Tensions,” CBS News, October
Answer by General John E. Hyten, Vice Chairman of 19. https://apnews.com/article/donald-trump-seoul-
the Joint Chiefs of Staff, at the Center for Strategic and south-korea-north-korea-joint-chiefs-of-staff-62456fab-
International Studies, in response to the question “Do 450ce5bba766a8307cf06af4.
you think current missile defense systems of the United
States are capable of protecting the United States from [Kristensen 2013] Kristensen, Hans M., and Robert
North Koreans – North Korea’s new missiles?” S. Norris. 2013. “Global Nuclear Weapons Invento-
Transcript online at: https://csis-prod.s3.amazonaws. ries, 1945-2013.” Bulletin of the Atomic Scientists
com/s3fs- public/event/200121_AConversationWith- 69, no. 5:75 - 81. https://www.tandfonline.com/doi/
GeneralJohnHyten.pdf?8vMO1PZ_loYtR2V0zWY85Nc- full/10.1177/0096340213501363.
qzm845PHY.
[Kristensen 2020] Kristensen, Hans M., and Matt Korda.
[Insinna 2019] Insinna, Valerie. 2019. “Space agency has 2020. “Chinese Nuclear Forces, 2019.” Bulletin of the
an ambitious plan to launch ‘hundreds’ of small satel- Atomic Scientists 76, no. 6:443 - 457. https://www.tand-
lites. Can it get off the ground?” Defense News, April 10. fonline.com/doi/full/10.1080/00963402.2020.1846432.
https://www.defensenews.com/space/2019/04/10/sda-
has-an-ambitious-plan-to-launch-hundreds-of-small-sat- [Kristensen 2021] Kristensen, Hans M., and Matt Korda.
ellites-can-it-get-off-the-ground. 2021. “North Korean nuclear weapons.” Bulletin of the
Atomic Scientists 77, no. 4:222 - 236. https://www.tand-
[JASON 2010] JASON. 2010. MDA Discrimination. fonline.com/doi/full/10.1080/00963402.2021.1940803.
JASON Defense Advisory Panel Report JSR-10-620.
McLean, VA: The MITRE Corporation, August 3. https:// [Kubovich 2020] Kubovich, Yaniv. 2020. “Israel unveils
irp.fas.org/agency/dod/jason/mda-dis.pdf. laser-based missile interception system.” Haaretz, Jan-
uary 8. https://www.haaretz.com/israel-news/.premi-
[Jones 2018] Jones, Harry W. 2018. “The Recent Large um-israel-unveils-laser-based-missile-interception-sys-
Reduction in Space Launch Cost.” Paper presented at tem-1.8374802.
48th International Conference on Environmental Sys-
tems, Albuquerque, NM, July 8 - 12. https://ttu-ir.tdl.org/ [Lewis 1993] Lewis, George N. and Theodore A.
bitstream/handle/2346/74082/ICES_2018_81.pdf. Postol. 1993. “Video Evidence on the Effectiveness of
Patriot during the 1991 Gulf War.” Science & Global
[Judson 2020] Judson, Jen. 2020. “Pacific Missile De- Security 4:1 - 63. https://www.tandfonline.com/doi/
fense Radars Nixed, MDA to Study Sensor Architecture pdf/10.1080/08929889308426392?casa_token=R-
in Regions.” Defense News, February 11. https://www. FRgqWDbc6QAAAAA:NzzvgLRB3vI5Zo63KBtFSGVn-
defensenews.com/smr/federal-budget/2020/02/11/pa- MDLdBJ5ATGDNyNAATUwazpvVRj1VSsJ4le0k1aDv_
qMrD0IeO6Jfe2s.

51
[Lister 2021] Lister, Tim. 2021. “Israel’s Iron Dome [NDAA 2018] National Defense Authorization Act for
Doesn’t Chase Every Rocket It Sees.” CNN, May 18. Fiscal Year 2018, Pub. Law No. 115-91 (Dec. 12, 2017).
https://www.cnn.com/2021/05/18/middleeast/israel- https://www.congress.gov/115/plaws/publ91/PLAW-
iron-dome-defense-gaza-rockets-intl-cmd/index.html. 115publ91.pdf

[McDowell 2021] McDowell, Jonathan C. 2021. “GCAT: [NDAA 2019] John S. McCain National Defense Autho-
General Catalog of Artificial Space Objects: Launches.” rization Act for Fiscal Year 2019, Pub. Law No. 115-232
planet4589.com. http://www.gsfc.gsfc.planet4589.com/ (Aug. 13, 2018). https://www.congress.gov/115/plaws/
space/gcat/web/launch/ldir.html. publ232/PLAW-115publ232.pdf

[MDA 2019] Missile Defense Agency. 2019. Department [NDAA 2020] National Defense Authorization Act
of Defense Fiscal Year (FY) 2020 Budget Estimates, De- for Fiscal Year 2020, Pub. Law No. 116-92 (Dec. 20,
fense-Wide Justification Book Volume 2a of 5: Research, 2019). https://www.govinfo.gov/content/pkg/PLAW-
Development Test & Evaluation, Defense-Wide. March. 116publ92/html/PLAW-116publ92.htm
https://comptroller.defense.gov/Portals/45/Documents/
defbudget/fy2020/budget_justification/pdfs/03_RDT_ [NDAA 2022] National Defense Authorization Act for
and_E/RDTE_Vol2_MDA%20RDTE_PB20_Justification_ Fiscal Year 2022, Pub. Law No. 117-81 (Dec. 27, 2021).
Book.pdf. https://www.congress.gov/117/bills/s1605/BILLS-
117s1605enr.pdf
[MDA 2020a] Missile Defense Agency. 2020a. Terminal
High Altitude Area Defense (THAAD) Request for Infor- [NIC 1999] National Intelligence Council (1999). Foreign
mation for Booster Development and Demonstration Missile Developments and the Ballistic Missile Threat to
(Revision 2). April 29. https://govtribe.com/file/govern- the United States Through 2015. National Intelligence
ment-file/mda20thrfi01-mda20thrfi01-rfi-thaad-booster- Council, September 9. https://www.dni.gov/files/docu-
revision-2-dot-pdf. ments/Foreign%20Missile%20Developments_1999.pdf.

[MDA 2020b] Missile Defense Agency. 2020b. Ballistic [Nikkei 2021] “North Korea Completes Building New
Missile Defense Testing (fact sheet). 20-MDA-10590. 3,000-Ton Submarine: Report.” Nikkei Asia, April 11,
September 30. https://www.mda.mil/global/docu- 2021. https://asia.nikkei.com/Spotlight/N-Korea-at-
ments/pdf/testprogram.pdf. crossroads/North-Korea-completes-building-new-3-
000-ton-submarine-report.
[MDR 2019] Office of the Secretary of Defense 2019a.
2019 Missile Defense Review. Washington, DC: De- [NMDA 1999] National Missile Defense Act of 1999,
partment of Defense, January. https://media.defense. Pub. Law No. 106-38 (Jul. 22, 1999). https://www.con-
gov/2019/Jan/17/2002080666/-1/-1/1/2019-MISSILE- gress.gov/106/plaws/publ38/PLAW-106publ38.pdf.
DEFENSE-REVIEW-little.pdf.
[NRC 1985] National Research Council. 1985. The
[Mehta 2020] Mehta, Aaron. 2020. “Griffin ‘Extreme- Effects on the Atmosphere of a Major Nuclear Exchange.
ly Skeptical’ of Airborne Lasers for Missile Defense.” Washington, DC: Committee on the Atmospheric
Defense News, May 20. https://www.defensenews. Effects of Nuclear Explosions, National Academies of
com/2020/05/20/griffin-extremely-skeptical-of-air- Science Press. https://www.nap.edu/catalog/540/the-ef-
borne-lasers-for-missile-defense/. fects-on-the-atmosphere-of-a-major-nuclear-exchange.

[MMIII Costs 2015] “LGM-30G Minuteman III,” in IHS [NRC 2012] National Research Council. 2012. Making
Jane’s Weapons: Strategic 2015 - 2016, ed. James C. Sense of Ballistic Missile Defense: An Assessment of
O’Halloran (United Kingdom: IHS, 2015), 105 - 107. Concepts and Systems for U.S. Boost-Phase Missile De-
fense in Comparison to Other Alternatives. Washington,
[Mosher 2000] Mosher, David. 2000. “Understanding DC: National Academies of Science Press.
the Extraordinary Cost of Missile Defense.” RAND Cor- https://doi.org/10.17226/13189.
poration (reprinted with permission from Arms Control
Today, December). https://www.rand.org/natsec_area/ [OTA 1979] “The Effects of Nuclear War”, Office of Tech-
products/missiledefense.html. nology Assessment, Congress of the United States, May
1979. https://ota.fas.org/reports/7906.pdf.
[Najjar 2020] Najjar, Alexandre. 2020. “Can SpaceX
Profit on Certain Starlink Launches?” SpaceNews, Sep- [Panda 2017] Panda, Ankit. 2017. “The Hwasong-15: The
tember 14. https://spacenews.com/op-ed-can-spacex- Anatomy of North Korea’s New ICBM.” The Diplomat,
profit-on-certain-starlink-launches/. December 6. https://thediplomat.com/2017/12/the-
hwasong-15-the-anatomy-of-north-koreas-new-icbm/.
[NDAA 2010] National Defense Authorization Act for
Fiscal Year 2010, Pub. Law No. 111-84 (Oct. 28, 2009). [Panda 2021] Panda, Ankit. 2021. “North Korea’s New
https://www.congress.gov/111/plaws/publ84/PLAW- ‘Hypersonic Missile’ Isn’t a Game-Changer Yet.” Business
111publ84.pdf. Insider, October 7. https://www.businessinsider.com/
north-koreas-new-hypersonic-missile-isnt-a-game-
[NDAA 2017] National Defense Authorization Act for changer-yet-2021-10.
Fiscal Year 2017, Pub. Law No. 114-328 (Dec. 23, 2016).
https://www.congress.gov/114/plaws/publ328/PLAW- [Panetta 2012] “I’m very confident that American de-
114publ328.pdf. fense capabilities are able, no problem, to block a rock-
et like this one.” Answer by U.S. Defense Secretary Leon

52
Panetta to a question from CNN on the capability of U.S com/2018/05/houthi-missiles-the-iran-connection-
missile defenses, December 12, 2012, quoted in “U.S. scuds-are-not-dead-yet/.
Hesitant in Condemning North Korean Launch,” Bradley
Clapper, The Associated Press, December 13, 2012. [Schiller 2019] Schiller, Markus. “The Scope of Foreign
Assistance to North Korea’s Missile Program.” Science
[Podvig 2004] Podvig, Pavel. 2004. Russian Strategic Nu- & Global Security 27, no. 1, 29–72, July 3, 2019. https://
clear Forces. Cambridge, MA: The MIT Press. https://mit- www.tandfonline.com/doi/full/10.1080/08929882.2019
press.mit.edu/books/russian-strategic-nuclear-forces. .1613805.

[Postol 2014] Postol, Theodore A. 2014. “The Evidence [Sessler 2000] Sessler, Andrew M., John M. Cornwall,
That Shows Iron Dome Is Not Working.” Bulletin of Bob Dietz, Steve Fetter, Sherman Frankel, Richard L.
the Atomic Scientists, July 19. https://thebulletin. Garwin, Kurt Gottfried, Lisbeth Gronlund, George
org/2014/07/the-evidence-that-shows-iron-dome-is- N. Lewis, Theodore A. Postol, David C. Wright. 2000.
not-working/. Countermeasures: A Technical Evaluation of the Oper-
ational Effectiveness of the Planned US National Missile
[Postol 2018] Postol, Theodore A. 2018. “North Korean Defense System. Cambridge, MA: Union of Concerned
Ballistic Missiles and US Missile Defense.” Physics and Scientists and MIT Security Studies Program, April: 64.
Society 47, no. 2 (March 3). https://higherlogicdown- https://drum.lib.umd.edu/handle/1903/4333.
load.s3.amazonaws.com/APS/a05ec1cf-2e34-4fb3-
816e-ea1497930d75/UploadedImages/Newsletter_ [Sherman 2019] Sherman, Jason. 2019. “Raytheon Con-
PDF/April18.pdf. fident SM-3 Block IIA Can Intercept ICBM Warheads.” In-
side Defense SITREP, October 7. https://www.proquest.
[Rahmat 2021] Rahmat, Ridzwan. 2021. “North Korea com/docview/2301450989.
Conducts Test-Firing of New Anti-Air Missile.” Janes,
October 1. https://www.janes.com/defence-news/ [Sherman 2020] Sherman, Jason. 2020. “MDA Seeks
news-detail/north-korea-conducts-test-firing-of-new-an- to Double Range of THAAD Interceptor for Homeland
ti-air-missile. Defense Mission,” Inside Defense SITREP, May 4, 2020.
https://insidedefense.com/daily-news/mda-seeks-dou-
[Reif 2019] Reif, Kingston. 2019. “Trump Seeks Missile ble-range-thaad-interceptor-homeland-defense-mis-
Defense Buildup.” Arms Control Today, March. https:// sion.
www.armscontrol.org/act/2019-03/news/trump-seeks-
missile-defense-buildup. [Smith 2020] Smith, Josh. 2020. “Faster Rockets, More
Warheads: What North Korea Could Gain from New
[Reuters 2017] Thomson Reuters 2017. “THAAD Missile Weapons Development.” Reuters, January 2. https://
Defense System.” May 2. http://fingfx.thomsonreuters. www.reuters.com/article/us-northkorea-missiles-tech-
com/gfx/rngs/SOUTHKOREA-THAAD/0100403L07J/ nology-analys/faster-rockets-more-warheads-what-
index.html. north-korea-could-gain-from-new-weapons-develop-
ment-idUSKBN1Z10BQ.
[Roberts 2014] Roberts, Brad. 2014. On the Strategic
Value of Missile Defense. Proliferation Papers 50. Paris: [Smith 2021] Smith, Josh. 2021. “N.Korea Says It Tested
Ifri Security Studies Center, June. https://www.ifri.org/ New Railway-Borne Missile System to Strike ‘Threaten-
sites/default/files/atoms/files/pp50roberts.pdf. ing Forces.’” Reuters, September 16. https://www.reu-
ters.com/world/asia-pacific/nkorea-launches-were-test-
[Roberts 2020] Roberts, Brad. 2020. “Missile Defense: new-railway-borne-missile-system-kcna-2021-09-15/.
Fit for What Purpose in 2030?” In Fit for Purpose? The
U.S. Strategic Posture in 2030 and Beyond, Brad Roberts [Smith 2022] Smith, Josh, and Shin, Hyonhee. 2022.
(ed.). Center for Global Security Research, Lawrence “North Korea launches ‘more advanced’ missile after hy-
Livermore National Laboratory, October. https://cgsr. personic test.” Reuters, January 11. https://www.reuters.
llnl.gov/content/assets/docs/The-US-Strategic-Posture- com/world/asia-pacific/nkorea-launches-possible-ballis-
in-2030-and-Beyond.pdf. tic-missile-japan-says-2022-01-10/.
[Rogoway 2021] Rogoway, Tyler. 2021. “China Tested [Sonne 2019] Sonne, Paul. 2019. “Trump: US Will Be
a Fractional Orbital Bombardment System That Uses a Able to Shoot Down Missiles ‘Anytime, Anyplace.’”
Hypersonic Glide Vehicle: Report.” The Drive, October Mercury News, January 17. https://www.mercurynews.
16. https://www.thedrive.com/the-war-zone/42772/ com/2019/01/17/trump-us-will-be-able-to-shoot-down-
china-tested-a-fractional-orbital-bombardment-sys- missiles-anytime-anyplace/.
tem-that-uses-a-hypersonic-glide-vehicle-report.
[Stair 1993] Stair, A.T., and Randall E. Murphy. 1993.
[Rogoway 2022] Rogoway, Tyler. 2022. “North Korea Background Assessment and Sensor Study-BASS, Bed-
Claims It Successfully Tested A New ‘Hypersonic Gliding ford, MA: AT Stair Associates, October.
Warhead.’” The Drive, January 5. https://www.thedrive.
com/the-war-zone/43764/north-korea-releases-new-hy- [Sullivan 1999] Sullivan, Jeremiah D., Dan Fenster-
personic-glide-vehicle-test-launch-image. macher, Daniel Fisher, Ruth Howes, O’Dean Judd, and
Roger Speed. 1999. “Technical Debate over Patriot
[Savelsberg 2018] Savelsberg, Ralph. 2018. “Houthi Performance in the Gulf War.” Science & Global Security
Missiles: The Iran Connection; Scuds Are Not Dead Yet.” 8:41–98. https://scienceandglobalsecurity.org/archive/
Breaking Defense, May 17. https://breakingdefense. sgs08sullivan.pdf.

53
[Sutton 2019] Sutton, H.I. 2019. “North Korea Appears [Varner 2020] Varner, Joe. 2020. “North Korea’s Two
to Have Built Its First Real Ballistic Missile Submarine.” New Strategic Missiles: What Do We Know and What Do
Forbes, August 13. https://www.forbes.com/sites/hisut- They Mean for U.S. Deterrence?” Modern War Institute,
ton/2019/08/13/north-korea-appears-to-have-built-its- October 14. https://mwi.usma.edu/north-koreas-two-
first-real-ballistic-missile-submarine/?sh=ad173bf14e20. new-strategic-missiles-what-do-we-know-and-what-do-
they-mean-for-us-deterrence/.
[Talmadge 2016] Talmadge, Eric. 2016. “S. Korea:
North Blew Up Rocket Stage to Confound Analysts.” [Warrick 2021] Warrick, Joby. 2021. “China Building
Associated Press, February 9. https://apnews.com/arti- More Than 100 New Missile Silos in Its Western Desert,
cle/5151700021ec4ff88f113013696a46ca. Analysts Say.” Washington Post, June 20. https://www.
washingtonpost.com/national-security/china-nuclear-
[Trevithick 2021] Trevithick, Joseph. 2021. “North Korea missile-silos/2021/06/30/0fa8debc-d9c2-11eb-bb9e-
Claims to Have Tested a Hypersonic Missile (Updated).” 70fda8c37057_story.html.
The Drive, September 28. https://www.thedrive.com/
the-war-zone/42.560/north-korea-claims-to-have-test- [Washburn 2013] Washburn, Alan R. 2013. Earth Cover-
ed-a-hypersonic-missile. age by Satellites in Circular Orbit. Calhoun Institutional
Archive of the Naval Postgraduate School. https://cal-
[Trump 2019] “Our goal is simple: to ensure that we can houn.nps.edu/handle/10945/38386.
detect and destroy any missile launched against the
United States —anywhere, anytime, anyplace.” President [Williams 2020] Williams, Ian, and Shaan Shaikh. 2020.
Donald Trump’s remarks on the occasion of the release The Missile War in Yemen. Center for Strategic and
of the 2019 Missile Defense Review. See transcript at International Studies, June. https://csis-website-prod.
12:55: https://www.c-span.org/video/?457050-1/pres- s3.amazonaws.com/s3fs-public/publication/Williams_
ident-trump-delivers-remarks-2019-missile-defense-re- MissileWarYemen_WEB_FINAL_v2.pdf.
view.
[Willman 2015] Willman, David. 2015. “Failed Missile
[UN 2017] United Nations. 2017. Midterm report of Defense Programs Cost the U.S. Nearly $10 Billion,” Los
the Panel of Experts established pursuant to reso- Angeles Times, April 4. https://www.latimes.com/nation/
lution 1874 (2009). August 4. https://www.undocs. la-na-missile-defense-20150405-story.html.
org/S/2017/742.
[Wright 2002] Wright, David, and Laura Grego. 2002.
[Van Diepen 2020] Van Diepen, Vann H., and Michael “Anti-Satellite Capabilities of Planned US Missile
Elleman. 2020. “North Korea Unveils Two New Strate- Defence Systems.” Disarmament Diplomacy, no. 68,
gic Missiles in October 10 Parade.” 38 North, October December 2002 - January 2003. http://www.acronym.
10. https://www.38north.org/2020/10/vdiepenmelle- org.uk/old/archive/dd/dd68/68op02.htm.
man101020/.
[Wright 2017] Wright, David. 2017. “Reentry of North
[Van Diepen 2021a] Van Diepen, Vann H. 2021a. “It’s Korea’s Hwasong-15 Missile.” Union of Concerned
the Launcher, Not the Missile: Initial Evaluation of Scientists: All Things Nuclear, December 7. https://allth-
North Korea’s Rail-Mobile Missile Launches.” 38 North, ingsnuclear.org/dwright/reentry-of-hwasong-15/.
September 17. https://www.38north.org/2021/09/its-
the-launcher-not-the-missile-initial-evaluation-of-north- [Wright 2019] Wright, David. 2019. Decoys Used in
koreas-rail-mobile-missile-launches/. Missile Defense Intercept Tests, 1999-2018. Union of
Concerned Scientists, January. https://www.ucsusa.org/
[Van Diepen 2021b] Van Diepen, Vann H. 2021b. “Six sites/default/files/attach/2019/01/Missile-Defense-Inter-
Takeaways from North Korea’s ‘Hypersonic Missile’ cept-Test-Decoys-white-paper.pdf.
Announcement.” 38 North, October 13. https://ww-
w.38north.org/2021/10/six-takeaways-from-north-ko- [Yeo 2017] Yeo, Mike. 2017. “Can North Korea Shoot
reas-hypersonic-missile-announcement/. Down a US Military Aircraft?” Defense News, Sep-
tember 28. https://www.defensenews.com/flash-
points/2017/09/28/can-north-korea-shoot-down-a-us-
military-aircraft/.
American Physical Society  •  1 Physics Ellipse  •  College Park, MD 20740  • U.S.A.  • aps.org