Reducing the cost of tactical BMD

With adversary ballistic threat precision increasing, and the cost of ballistic missile defence rising, many assumptions within our system design requirements for ballistic missile defence need reviewing. This analytical deep dive explores how integrating sensors, adopting ‘high-low’ interceptor mixes, and reconsidering traditional metrics of success could reshape how we think about ballistic missile defence.

Defending forces in the field from ballistic threats is not a new challenge and has been a consideration for Western militaries since the 1980s when the USSR fielded the OTR-23 OKA SRBM.[1] The threat posed by tactical and theatre level ballistic missiles was further underscored during the 1990s, when Iraqi Scud missiles struck US Barracks at Dharan in Saudi Arabia.[2] However, the threat was framed primarily in terms of political risk. Since relatively limited casualties had doomed western expeditions in Lebanon and Somalia among other instances, it was presumed that even limited successes for adversary ballistic missile arsenals could pose unacceptable political risks in an age of wars of choice.[3] Adversary arsenals were not especially large, nor especially capable (the variants of the Scud missile, for example, have a CEP of around 450 m), but they did not need to be to score a lucky hit.[4]

In this context, achieving a defence that was as airtight as possible against a limited threat was the paramount concern.[5] As a consequence, maximising the effectiveness of both sensors and effectors against different ballistic target types such as short-range ballistic missiles (SRBMs), medium-range ballistic missiles (MRBMs), and intermediate-range ballistic missiles (IRBMs) was the key consideration which led to the emergence of specific solutions to individual parts of the threat spectrum. Ballistic missile defence (BMD) systems, such as THAAD, which have specific utility against MRBM and IRBM targets (their IR seekers being of more limited utility at the altitudes at which many SRBMs spend most of their trajectories), were paired with shorter-range systems such as PATRIOT PAC-3 MSE built to provide functionality against SRBMs.

The challenge which the BMD architectures developed over the previous several decades will face is twofold. First, ballistic threats are increasingly coming to combine both mass and precision. For example, Russian ballistic missile salvos against Ukraine have increasingly come to comprise a high-low mix, combining the Iskander-M 9M723 (an SRBM with a circular error probable (CEP) of around 10 m) with North Korean KN-25 SRBMs and Tochka-U tactical ballistic missiles (TBMs) which have considerably lower accuracy but which force the expenditure of interceptors.[6] In the Pacific, the PLA fields a robust and growing arsenal of short-, medium- and intermediate-range ballistic missiles. Between 2015 and 2024, for example, the PLA Strategic Rocket Force produced several hundred DF-26 IRBMs.[7] This poses a challenge to BMD architectures comprised of systems optimised against specific threat and altitude profiles which can, necessarily, only be procured in comparatively limited numbers if each layer of a BMD architecture is to be separately resourced. Even so, however, one cannot ignore specific components of the threat spectrum given that many of the systems involved will be used in convergent strikes on the same targets. A recent example of this is the Russian strike on Dnipro which involved the RS-26/Oreshnik IRBM and the KH-47M2 Kinzhal (which is effectively an air-launched variant of the 9M723 SRBM).[8] Compounding this is the fact that ballistic threats are joined by air breathing threats such as cruise missiles and UAVs. The risk is that the more layers one adds to an air defence system, the more poorly-resourced each layer becomes.

Secondly, the production of interceptors is constrained by a number of factors. Key among them is the fact that the availability of a number of long-lead items is heavily constrained. While industry stockpiling can somewhat mitigate the effects of this challenge, it is only a partial solution given the inherent inefficiencies involved in stockpiling. Moreover, many inputs such as solid fuel and the microchips involved in guidance systems are fungible and thus required by other complex weapons.[9]

As a consequence, the question of how the costs of ballistic missile defence and in particular BMD for fielded forces at tactical and operational ranges might be reduced is a pressing one.

A different operational approach

A major consideration which shapes the conduct of BMD and integrated air and missile defence (IAMD) more broadly is the requirement for a high probability of kill (Pk), which in turn determines system design and requirements. This is, as noted, the legacy of an era in which a high level of security against a moderate-sized threat was required. To be sure, prioritisation does occur in the form of critical asset lists and defended asset lists, but once a target is prioritised, a very high Pk against an incoming target is the norm and this in turn determines the single-shot probability of kill (SSPk; the Pk of any given interceptor) and thus the design requirements of interceptors. The process by which requirements are generated involve the following steps:

Mission requirements are defined in terms of the physical characteristics of the threat, such as speed and altitude, which in turn define system requirements in terms of characteristics such as detection ranges and integration rates (for sensors) or kinematics (for seekers).[10] Notably while trade-offs are discussed, this occurs later in the process after system requirements have been defined and so primarily relate to the choices to be made between, for example, seeker types (where range and resolution must be traded) rather than with respect to more fundamental questions regarding scalability and the capacity to meet a definition of sufficiency that is defined in relation to an operational requirement. Rather, expected Pks are assumed to be high.[11]

This approach is rather problematic because when interceptor design is based on a Pk that is determined at least to some degree in isolation from the operational environment, this necessarily leads to ‘gold-plating’, since the determinant of success is based on engineering qualities rather than desired effects. This represents a challenge in a context where the expectation of being able to intercept every incoming target is both unrealistic and in many cases operationally unnecessary.

Instead, the required Pk might itself be understood in terms of the level of acceptable risk in a specific operational context. This in turn depends on an opponent’s understanding of the criteria for inflicting crippling damage on a target site, and their capacity for salvo coordination. To use an example, Russian military theorists assume that it requires 60 cruise or ballistic missiles striking aimpoints in order to render an airbase non-functional with a high degree of confidence.[12] The likelihood of losing even a marginal part (for example 20%) of the already large missile salvo needed to achieve this thus has ramifications for Russian planners. The reason for this is that the larger a salvo is, the more likely it is to require greater levels of planning and synchronisation, often across services, to coordinate arrival times of missiles launched from different locations. To render a tactical or operational target more difficult to attack, a defensive system need not be hermetic – instead it need only impose a requirement to launch more missiles in any salvo than an opponent can easily coordinate, and to achieve this repeatedly. Consequentially, this involves lowering SSPk and increasing scalability.

Relatedly, it must be remembered that the probability of a missile striking a target is not simply determined by the likelihood of active defences prevailing. First, the missile’s accuracy in CEP and warhead size are determinative. Generally speaking, the number of missiles needed to strike a target with certainty is determined by the equation:

Where rl is the lethal radius of the missile. So, for example, a missile with a CEP of 10 m and a warhead of 450 kg such as the 9M723 has a 50% chance of destroying a target such as a hardened air shelter. By contrast, an IRBM such as the Chinese DF-26 (with a CEP of around 150 m) has a roughly 4% chance of destroying a hardened target. It follows, then, that particularly at battlefield depths in excess of 500 km, the impact of hardening can be considerable. Even assuming more accurate missiles, palliatives such as hardening can increase the number of missiles needed per aimpoint – thereby closing the delta between real and maximally viable salvos further, and reducing the burden on missile defences.

Third, factors such as mobility, dispersion and soft kill can have a considerable impact on ballistic missile effectiveness. Take, for example, the case of the US’ efforts to produce highly-accurate conventionally-armed Manoeuvrable Re-entry Vehicles (MaRVs) under the prompt global strike (PGS) programme, with tests of E2 (Enhanced Effectiveness) in 2002, and LETB (Life Extension Test Bed) modified versions of Trident missile re-entry bodies in 2002 and 2005 respectively.[13] During tests, the loss of access to Global Positioning System (GPS) signals caused the MaRVs to land well away from their targets, though close to the locations where the navigation systems assumed the target to be. Positioning, navigation, and timing (PNT) jamming against ballistic targets is difficult given their trajectories and speed, but not impossible. The key consideration is that jamming need not be a perfect solution, merely a factor that adds to the margin of uncertainty regarding the SSPk of a single missile. The same might be said of other methods of defeat, such as the use of obscurants against missiles with optoelectronic seekers (including variants of the 9M723) and decoys.

None of this would render the ability to hard-kill missiles irrelevant. However, a mission engineering approach which began with a) an assessment of adversary capacity to generate salvo sizes, then b) proceeded to assess how much uncertainty factors such as a missile’s own CEP, soft kill and passive defences added regarding SSPk, and then c) proceeded to the question of how high the likelihood of active defence would need to be, in order to close the delta between currently-assessed salvo size and the salvo size needed to overcome the inherent risks of both active and passive defence, would yield a considerably lower SSPk for interceptors.

For illustrative purposes, let us return to the case of the Russian criterion of 60 missiles on target to render an airbase non-functional. Assuming that the impact of the average CEP inflates this number by at least 50% (in the case of very accurate missiles like the Iskander), if measures such as hardening are taken and that other passive methods (such as the use of obscurants) impose a marginal 10% attrition, a level of attrition inflicted by active defences of around 50% would more than double the size of a required salvo and raise the required number of missiles to a number in excess of almost all of Russia’s missile salvos in the Russo-Ukrainian War (and well in excess of any ballistic missile salvo).

As such a system which combined a highly-effective interceptor such as PAC-3 MSE (which reportedly has a SSPk of 0.8 in test conditions) and a much less-effective interceptor would have a cumulative Pk well in excess of the demand.

If an acceptable SSPk is defined in terms of closing the delta between real and maximum salvo sizes given other factors, this would impact several aspects of systems design. The first factor would be warhead type. Hit-to-kill interceptors geared toward the BMD role, such as PAC-3 MSE, introduce especially stringent demands in terms of missile speed and kinematics, given that the missile must make kinetic impact with its target at a very high speed to achieve target defeat. This introduces a number of design requirements, such as the use of ceramic shrouds (as on PAC-3 CRI) in lieu of more scalable Quartz Duroid shrouds (as on original PAC-3) for seekers, which are used on a number of air defence interceptors.[14] In addition, the use of Titanium warheads rather than tungsten pellet blast fragmentation warheads imposes further costs. These costs are justifiable if one is seeking to maximise SSPk, and bespoke systems still represent an entirely valid first line of defence. However, if one is meeting a sufficiency criterion a shot doctrine for employing interceptors might well employ a ‘high-low’ mix which allowed the employment of bespoke and expensive interceptors in tandem with cheaper systems armed with high-explosive fragmentation (HE-FRAG) warheads, which can also be used in air defence intercept roles.

Integration as a force multiplier

Secondly, it should be noted that there is an inverse relationship between sensor reach and coverage on the one hand, and the cost and complexity of an interceptor on the other. For example, the shorter the homing times of an interceptor are, the more capable of high g manoeuvres it must be. Similarly, the degree to which an interceptor can receive data enabling, for example, target discrimination from other sources impacts the extent to which its own on-board seeker must be able to generate high-fidelity data (which typically requires a Ka-band seeker).[15]

The inverse relationship between sensor range and interceptor complexity was perhaps best illustrated by the interception of an Armenian Iskander missile by Azerbaijan’s Barak-8 in the 2021 Nagorno Karabakh conflict. The Barak-8, which is an air defence missile equipped with an HE-FRAG warhead, and understood to have an average speed of Mach 2, is not an ideal BMD interceptor.[16] The intercept may have been a lucky one, but it is also potentially the case that the early detection provided by the Azeri Green Pine radar meant that the interceptor had a longer homing time and thus a more limited requirement for high-g course corrections.[17] To be sure, this is not a substitute for dedicated hit-to-kill interceptors, but does mean that the longer one’s detection range, the more effective non-dedicated interceptors will be in a BMD role – allowing for a partial erosion of the aforementioned stovepipes, particularly in tactical contexts.

The integration of sensors for air and missile defence is not necessarily new; examples include the pillar programmes of the Naval Integrated Fire Control–Counter Air (NIFC-CA) programme and the more recent cueing of a PATRIOT battery with tracks from an F-35 during US Army tests at the White Sands Missile Range. US aircraft have long been able to conduct cooperative engagements with surface launched missiles as part of NIFC-CA.[18]

However, while the advantages of more sensor coverage and information are well understood, the relationship between sensor and effector integration is often overlooked. The more networked a surface-to-air missile (SAM) system is with sensors within the IAMD system, the less capable its own on board sensors need to be. For example, while a high-performance Ka-band seeker is needed to discriminate a warhead from decoys, such as the six 9B999 radio frequency (RF) decoys carried by the 9M723 Iskander, networking together missiles with Ka-band seekers and lower-cost missiles without them can allow the latter to draw data from the former, meaning that two missiles with costly Ka-band seekers do not need to be ripple fired for a terminal phase intercept. Instead a salvo could combine one high-performance missile and one with a less-capable seeker. Similarly, the ability to cross-reference data feeds from multiple sources (such as in a multi-static radar array) can grant an air defence system both earlier detection and the elimination of false positives, in turn making it possible to lessen the demand for effector kinematics and seeker fidelity.

This does not eliminate the requirement for dedicated BMD effectors, particularly against upper-tier threats such as IRBMs. However, the point is that one consequence of a better integrated system is that the availability of information at the network level reduces the unit-level requirement to gather data, and can enable the hardware of some individual systems (such as interceptors) to be simplified.

Going long or short, and the value of offence-defence integration

One argument which has been made is that a focus on larger numbers of shorter-range interceptors which can be packed in a limited launch space (particularly at sea) might supersede the traditional approach emphasising layered defence. This has, for example, been the basis for recent US Navy tests in which PAC-3 MSE interceptors have been employed in Mk41 VLS cells.[19] However, while there is utility to this, range remains an important means of achieving the functional equivalent of magazine depth.

If one assumes that the SSPk is roughly equivalent across the layers of a system, a multilayered system is generally more-cost effective than a single-layered system because it enables a ‘shoot-look-shoot’ approach. A single-layered BMD system, by contrast, must necessarily ripple fire interceptors in order to achieve a high probability of intercept against targets. This ceases to hold, however, if one of two conditions is met.

Firstly, if the cost of a long-range interceptor must exceed that of shorter-range systems by a factor of two or more, layering ceases to be useful. This is arguably not the case – interceptors such as the MBDA Aster-30 Block 1NT and the Rafael/Raytheon Stunner used with the David’s Sling system even understood to cost less than PAC-3 MSE. The argument might, however, hold with respect to systems such as SM-3 in a counter-IRBM role, although there may be few alternatives to exoatmospheric intercept here. Moreover, costs for exoatmospheric intercept can be reduced considerably as illustrated by the IAI Arrow-3 which costs USD 5 million due to, among other things, its relatively simple terminal phase thrust vectoring mechanism. It would appear, then, that the logic of layering does not necessarily cease to hold.

Moreover, the ability to target launchers can significantly reduce the demand for interceptors. In many instances, the data gathered regarding a missile’s boost phase infrared (IR) signature and/or trajectory can also be employed to triangulate the location of a launcher (with trajectory calculation already a feature of the AN/MPQ-65 radar of the US PATRIOT system).

While air-based and deep strike capabilities may be tasked with functions other than counter-launcher operations, for example suppression/destruction of enemy air defences (SEAD/DEAD), targets such as transporter erector launchers (TELs) do not require heavy payloads to destroy; loitering munitions with payloads of 40 kg (and indeed often considerably less) have been employed against these targets in several conflicts.[20] Over short ranges where accuracy is easier to achieve, dual-use interceptors which combine offensive and defensive functions might thus prove especially useful. The employment of air defence systems in a surface-to-surface role is not new and is a feature of the SM-2, SM-6, and S-300P, among other systems. However design trade-offs have limited the use of SAM systems in this way. Arguably, the counter-battery function at tactical ranges represents the most viable offensive use of these missiles, since the major design trade-off between offensive and defensive missiles (warhead weight) matters less against TEL-type targets, against which a 150 kg warhead, such as that used on many long-range SAMs is entirely sufficient.

Conclusion: More than magazine depth

The challenge of scaling BMD solutions is a real one but the solution to it cannot be to simply make more missiles. Scaling existing capabilities is important, but it will also be important to adopt a less stovepiped approach to BMD as an enterprise, in which operational demands set requirements to a greater degree than engineering characteristics. This, coupled with efforts to use sensor integration to partially erode the silos between BMD and other parts of the air defence spectrum will be of central importance – if the defence of fielded forces against ballistic threats is not to be dependent on bespoke solutions which do not scale.

Dr Sidharth Kaushal

Author Box: Dr Sidharth Kaushal is a Senior Research Fellow at the military sciences team within the Royal United Services Institute (RUSI). His specialisms include Sea Power and Integrated Air and Missile Defence.

 

[1] David Rubinson and James Bonomo. NATO’s Anti Tactical Missile Defense Requirements and Their Relationship to the Strategic Defence Initiative. (Santa Monica:RAND, 1983)

[2] JC Humphrey. Casualty management: scud missile attack, Dhahran, Saudi Arabia. Mil Med. 1999 May;164(5):322-6.

[3] Eric V Larson Glenn A Kent. A New Methodology for Assessing Multilayer Missile Defense Options. (Santa Monica:RAND, 1994)

[4] CSIS. SS-1 Scud. https://missilethreat.csis.org/missile/scud/

[5] Larson Kent A New Methodology

[6] Josh Smith. Explainer: Where Did Russia Get Its North Korean Missiles. Reuters.January 5, 2024. https://www.reuters.com/world/where-did-russia-get-its-north-korean-missiles-2024-01-05/

[7] Defense News. Satellite Images Reveal Rise of Chinese Forces with 59 New DF-26 Missile Launchers. Defense News 2024. https://armyrecognition.com/news/army-news/army-news-2024/satellite-images-reveal-rise-of-chinese-forces-with-59-new-df-26-missile-launchers. Accessed on 20/01/2024

[8] Chris York. Russia reportedly uses new ‘Oreshnik’ ballistic missile against Ukraine — what we know so far. Kyiv Independent. November 2024. https://kyivindependent.com/russia-reportedly-launches-intercontinental-ballistic-missile-against-ukraine-what-we-know-so-far/. Accessed on 31/01/2025

[9] Author interviews with industry SMEs March 2024

[10] Warren Boord and John Hoffman. Air and Missile Defence Systems Engineering.(New York:Taylor and Francis, 2016)p.16

[11] Roy M Smith, ‘Using Kill-Chain Analysis to Develop Surface Ship CONOPs to Defend Against Anti-Ship

Cruise Missiles’, Master’s thesis, US Naval Postgraduate School, June 2010), pp. 84–90, <https://apps.dtic.

mil/sti/citations/ADA524758>, accessed 3 October 2024.

[12] Clint Reach et al. Russia’s Evolution Towards a Unified Strategic Operation. (Santa Monica:RAND, 2023) p.40

[13] Amy Woolf. Conventional Prompt Strike and Long Range Ballistic Missiles. (Washington DC:CRS, 2021) p.21-24

[14] Patrick O’Reilley, Ed Walters. The Patriot PAC-3 Missile Program- An Affordable Integration Approach. https://apps.dtic.mil/sti/pdfs/ADA319957.pdf

[15] Boord and Hoffman Air and Missile Defence Systems Engineering

[16] Naval Barak-8 Missiles Israel. https://www.naval-technology.com/projects/naval-barak-8-surface-air-missiles/

[17] Israeli Defense System Shot Down Russian Missile in Karabakh War – Reports.Moscow Times. https://www.themoscowtimes.com/2021/03/02/israeli-defense-system-shot-down-russian-missile-in-karabakh-war-reports-a73116

[18] See Sidharth Kaushal and Jack Watling. Requirements for the Command and Control of the UK’s Ground-Based Air Defence. (London:RUSI, 2022); NA O’Donaghue. Distributed Kill Chains:Drawing Insights About Mosaic Warfare From the Immune System and From the Navy. (RAND:Santa Monica, 2021)

[19] On this argument see Bryan Clark. Commanding the Seas: The US Navy and the Future of Surface Warfare. Washington DC:CSBA, 2017); Naval News. Lockheed Martin Provides Updates on Naval PAC-3 MSE and G-VLS. https://www.navalnews.com/event-news/sna-2025/2025/01/lockheed-martin-provides-updates-on-naval-pac-3-mse-and-g-vls/

[20] Israel Defense. Armenia and Azerbaijan resume fighting: watch the Israeli Harop destroy Armenian S-300 batteries. https://www.israeldefense.co.il/en/node/55866