BALLISTIC MISSILE DEFENSE - Princeton

Chapter 3

BALLISTIC MISSILE DEFENSE

Chapter 3.— BALLISTIC MISSILE DEFENSE

PageOverview . . . . . . . . . . . . . . . . . . . . .. ....111

Technical Possibilities for BMD . ......111LoADS with MPS Basing. . . . . . . . . . . . .112The Overlay and Layered Defense of

Silo-Based MX. . . . . . . . . . . . .......11 3O t h e r B M D C o n c e p t s . . . . . . . . . . . . . . . 1 1 3T h e A B M T r e a t y . . . . . . . . . . . . . . . . . . . 1 1 3

Endoatmospheric Defense. . . . . . . . . . . . .114Technical Overview of

Endoatmospheric BAD. . . . . . . . . . . .114LoADS With MPS Basing. . . . . ........118O t h e r E n d o C o n c e p t s . . . . . . . . . . . . . . . 1 2 6

Exoatmospheric and Layered Defense. ...129Technical Overview of Exo BMD .. ....129The Overlayand Layered Defense of

S i l o - B a s e d M X . . . . . . . . . . . . . . . . . . . 1 3 1

History of BMD and the ABM Treaty .. ...139The ABM Limitat ion Treaty . . . . . . . . . .140Application of ABM Treaty Provisions

t o M X D e f e n s e . . . . . . . . . . . . . . . . . . . 1 4 1Future ABM Limitation Negotiations ...142

LIST OF TABLES

Table No.20. Army’s LoADSCost Estimate,

October1980 . . . . . . . . . . . . . .

LIST OF FIGURES

Figure No.59.

60.

61.62.

63.

64.

Comparison of Ballistic MissileDefense Systems. . . . . . . . . . . . . .

Page

. . . . 125

Page

. . . . 112LoADS Defense Unit BeforeBreakout. . . . . . . . . . . . . . . . . . . . . . . .119LoADSDefense Unit After Breakout. . .120Overlay/Underlay Layered DefenseSystem . . . . . . . . . . . . . . . . . . . .0. . . . .131Sensitivity of Layered DefensePerformance to Overlay Leakage. . ...134U.S. Defensive Arsenal Needed toAssure l,OOO Surviving U.S. ReentryVehicles ., . . . . . . . . . . . . . . . . . . . . ..135

Chapter 3

BALLISTIC MISSILE DEFENSE

Bal l ist ic m i s s i

also called antterns —would seek

e defense (B MD) systems —ballistic missile (ABM) sys-k to ensure MX survivability

by destroying attacking reentry vehicles (RVS)either in space or after they entered the at-mosphere. Different BMD concepts can havevery different capabiIities and weaknesseswhich suit them for different MX basing rolesThus, it is important to keep clear the contextfor which the defense is intended, i.e , whetherit is desired to defend a large number of mul-tiple protective shelters (MPS) or a relativelysmaII number of siI os This chapter discussesthe technical aspects of the entire range ofendoatmospheric and exoatmospheric defensesystems but will concentrate on the two BMDconcepts most often discussed in the contextof a near-term decision regarding MX basing:the Low-Altltude Defense System (LoADS),

which is suited for the role of enhancing thesurvivability of MX in MPS; and the Overlaycomponent of a Layered Defense, appropriatei n theory for defense of MX based i n conven-tional silos.

There have been many changes in the tech-nical nature of BMD systems in the past dec-ade regarding both systems concept andunderlying technology. Systems contemplatedtoday are quite different f rot-n those discussedin the ABM debate of a decade ago. From atechnical point of view, therefore, the issuesrelevant to that debate have been replaced bya n entire I y new set of issues. Though there aremany paraIIels, intuitions based on previousacquaintance with BMD will not always berelevant — again from a purely technical pointof view — to the systems cent emplated today

OVERVIEW

Technical Possibilities for BMD

It is useful to distinguish BMD systems ac-cording to the altitude regime in which theytrack their targets and make their intercepts,since this Iargely dertermines the effectivenesspossible with such a system. Endoatmos-pheric — or “endo ” — defense systems performtracking and intercept within the sensible at-mosphere, from the Earth’s surface to about300,000-ft altitude, For various technicalreasons, U. S endo BMD efforts have concen-trated lately on the low-altitude regime, belowabout 50,000 ft Low-altitude endo systemssuch as LoADS are Iimited to making a smallnumber of intercepts over a given defendedtarget If the number of targets is relativelysmall, as in the case of silo basing, such defen-sive systems can only exact a small number ofRVS from the attacker Low-altitude systems bythemselves are therefore of limited valueunless the number of targets or aim points islarge, as with MPS basing. The very fact thattheir goal — forcing the offense to target a

small number of RVS at each aim point insteadof one — is modest, means that low-altitudesystems do notachieve this goaI

Exoat mospher

lave to perform very welI to

c — or “exe” – defenses trackand intercept RVS in space I n contrast to low-altitude endo defenses, exo systems can inprinciple intercept many RVS attacking thesame target Systems with an exo componentcan therefore in theory defend a small numberof targets such as silo-based missiIes from alarge attack However, this more demandingtask means that an exo system must be verygood indeed to accomplish it. Thus, an exosystem —even when accompanied by an endosystem in a “Layered Defense” — must have ahigher performance to do its job than a low-altitude system requires to do its more modestjob.

In addition to specifying the capabilities ofa BMD system, the altitude regime determinesthe type of sensor and interceptor required,

111

112 ● MX Missile Basing

which in turn establishes the type of tech-nology required for the system and its poten-tial vulnerabilities (see fig. 59).

Endo systems normally employ ground-based radars and nuclear warheads to trackand destroy targets. Radar blackout caused bynuclear detonations in the atmosphere is not acrippling problem for low-altitude endo sys-tems, as it is for high-altitude endo systems,but it (along with other factors) imposes thelimitation discussed above that only a verysmall number of intercepts can be made withina small area. Operation in the dense air at lowaltitudes means that it is very difficult for anopponent to fool the defense with decoys.

Operation in space would allow exo defenseto make use of nonnuclear kill mechanismsand the tactic of preferential defense. Multiplekill vehicles can also be mounted on a singleinterceptor missile, resulting in some savingsgiven the cost of boosting defensive vehiclesinto space in the first place. Infrared sensorsare preferable to radars for exo defense. With-out the filtering effect of dense air within theatmosphere, exo sensors are vulnerable to of-fensive tactics making use of decoys and otherpenetration aids.

LoADS With MPS Basing

This use of BMD would be an alternative toincreasing the number of shelters in an MPS

system in the face of a growing Soviet threat.In the Air Force baseline horizontal MPS sys-tem, for example, a LoADS defense unit wouldbe hidden in one of the 23 shelters in eachcluster and programed to intercept the first RVapproaching the shelter containing the MXmissile. Since the Soviets would be presumednot tO know which shelter contained the MX,they would have to assume for targeting pur-poses that each of the 23 shelters contained anMX missile defended by LoADS. If the defensewere only able to intercept one RV over eachdefended shelter, the Soviets would have totarget two RVS at each shelter instead of one.Thus, LoADS would increase the attack pricefor an MX missile from 23 to 46 Soviet RVS.

It is possible to have high confidence thatLoAIDS could exact this price of 2 RVS pershelter if the locations of the LoADS defenseunit; and the MX missiles could be concealedand if the defense unit could be hardened tosurvive the effects of nearby nuclear detona-tions. This confidence, conditional on success-ful deception and nuclear hardness, resultsboth from advances in BMD technology in thelast decade and from LoADS’ relat ivelymoclest goal of exacting from the Soviets onemore RV per aim point.

Preservation of location uncertainty (PLU)would be made more difficult with the addi-tion of LoADS to the MPS system, since theLOADS ,defense unit, MX missile, and simu-

Figure 59.—Comparison of Ballistic MissileDefense Systems

“Exo” atmospheric “Endo” atmosphericdefense defense

rSafeguard

Overlayoparatingarea space atomosphere atmosphereType of

of sensorKIHmechanism Non-Nuclear Nuclear N u c l e a r

SOURCE Office of Technology Assessment

Ch. 3—Ballistic Missile Defense ● 113

I a t o r m u s t a l l h a v e i n d i s t i n g u i s h a b l e s i g -natures. The nuclear effects requirements forLoADS are unprecedented. The design goals ofPLU and hardening must furthermore be metsimultaneously. It is not possible to have con-fidence that these goals can be met untildetailed design and testing are done,

In addition to PLU and hardness, there arestylized attacks or “reactive threats” whichcould pose a long-term threat to LoADS. Theserisks are judged moderate,

The Overlay and Layered Defenseof Si lo-Based MX

The Army’s concept of Exo defense, calledthe Overlay, would consist of interceptors,about the size of offensive missiles, launchedinto space from silos. Each interceptor wouldcarry several kill vehicles that would be dis-patched, using infrared sensing, to destroy at-tacking RVS before they entered the atmos-phere. The Overlay could be deployed with anendo “Underlay” to make a Layered Defenseof silo-based MX.

High efficiency would be required of theOverlay if it was to be able to defend a smallnumber of MX silos against a large Soviet at-tack. The Overlay is in the technology explora-tion stage, and there is no detailed systemdesign such as exists for LoADS. There aremany uncertainties about whether the OverlaycouId achieve the high level of performance itwould require to satisfy the needs of MX bas-ing. These uncertainties concern both theunderlying technology and the defense systemas a whole. In addition, there is a potential“Achilles’ heel” in the vulnerability of theOverlay to decoys and other penetration aids.

For the moment it would be quite risky torely on the Overlay or Layered Defense as thebasis for MX survivability.

Other BMD Concep ts

This chapter will also discuss briefly otherBMD concepts which have been studied,

A concept called “Dust,” “Environmental,”or “Ejecta” defense involves burying “clean”nuclear weapons in the vicinity of missile silos.The bombs would be exploded on warning of aSoviet attack, filling the air with dust whichwould destroy Soviet RVs before they reachedthe ground. Though there is little technicaldoubt about the high effectiveness of dustdefense, there is considerable concern aboutpublic reaction to plans for the deliberatedetonation of nuclear weapons on U.S. ter-ritory.

Various low-altitude or “last-ditch” con-cepts based on simple or “novel” principleshave been proposed. Though perhaps relevantfor other BMD roles, these concepts do not ap-pear to have an application in MX defense,given the requirement to preserve a smallnumber of MX missiles against a large numberof Soviet RVS.

The Army’s Site Defense is a derivative ofthe Sprint component of the Safeguard de-fense system of a decade ago. Based on thetechnology of the 1970’s, Site Defense is pre-served as an option in the event of a decisionto field a BMD system based on known tech-nology in a short period of time. Though in-adequate for the role of MX defense, Site De-fense could be appropriate for other limitedBMD roles.

The ABM Treaty

The 1972 ABM Limitation Treaty was nego-tiated as part of the SALT I package of stra-tegic arms limitation agreements. A Protocolspecifying further I imitations was signed in1974. The Treaty is of unlimited duration but issubject to review every 5 years. In addition, theStanding Consultative Commission created bythe Treaty meets about every 6 months to re-view implementation of the provisions of theTreaty and to consider such matters as ‘theparties might wish to raise.

Briefly, the Treaty and Protocol allow de-velopment of some types of ABM systems butlimit their deployment to small numbers at


specified sites. Development of other types ofABM beyond the laboratory is forbiddenaltogether.

No meaningful defense of MX missiles,either in silos or MPS, would be permittedwithin the Treaty, since any such deploymentcould consist of at most 20 radars (18 small, 2large) and 100 interceptors confined to thevicinity of Grand Forks, N. Dak., or Wash-ington, DC.

Limitations on development constrain thetypes of ABM work that can pass beyond thelaboratory stage. Since LoADS consists ofradars and interceptors of the kind permittedby the Treaty, development of this system can

proceed without abrogation or renegotiationof the Treaty except where such developmentconcerns the specific features of mobility,more than one interceptor per launcher, or ahypothetical reload capability. Developmentof the Overlay interceptors can proceed to theextent of testing single kill vehicles on in-terceptors, but development of multiple killvehicles outside of the laboratory is forbidden. ‘Development of space-, sea-, or air-borne ABMsystem components outside of the laboratoryis also forbidden. The Treaty specifies that de-velopment of ABM systems based on “newtechnologies” unforeseen or unspecified at thet ime the t rea ty was d ra f ted canno t bedeployed.

ENDOATMOSPHERIC DEFENSE

Technical Overview ofEndoatmospheric BMD

Endoatmospheric —or “endo” – defense sys-tems perform tracking and intercept within thesensible atmosphere, from the Earth’s surfaceto about 300,000-ft altitude. It is important todistinguish high-altitude systems, which ac-quire and track their targets above about100,000 ft, from low-altitude systems, whichtrack and engage below 50,000 ft. The Sprintcomponent of the Safeguard system is an ex-ample of the former type and the Army’s pres-ent LoADS concept is an example of the latter.

Endoatmospheric defense normally employsground-based radars for tracking. Optical or in-frared sensors would be inappropriate for endooperation because, among other reasons, theycannot supply accurate range information andlow cloud cover or dust could obscure theirview of incoming RVS.

Nonnuclear k i l l is possible in the at-mosphere, but nuclear warheads provide amore certain kill mechanism. A nonnuclear killwouId require that the radar provide very ac-curate trajectory information to the intercep-tor or that the interceptor have its own sensor.Because the kill radius of a nuclear warhead is

much greater, less accurate information suf-f ices to guarantee RV destruction.

Neutrons released from a defensive nuclearwarhead provide the mechanism for disablingthe offensive RV. An RV warhead contains fis-sionable material that absorbs neutrons veryreadily: this is the property that allows thenuclear chain react ion to proceed when the RVis detonated. When the fissionable material inan incoming RV absorbed the neutrons fromthe defensive warhead, it would be renderedunable to detonate. Physical destruction of theRV would therefore not be necessary: thoughblast from the defensive warhead could play arole, it is a less certain kill mechanism. Theneutron kill is sure because the incoming RVmust contain neutron-absorbing material to doits job, and it is very difficult to shield againstneutrons. A relatively low-yield defensive war-head (tens of kilotons) could generate a neu-tron fIuence lethal to RVS at ranges of severalhundred feet from its detonation point. Thedefensive interceptor therefore would nothave to be very accurate to ensure disabling ofthe RV.

Use of nuclear interceptors does involvespecial procedures for their release, however.Release of offensive nuclear weapons must be


authorized by the National Command Au-thorities. The procedures for defensive nuclearrelease have not been worked out since theUnited States has no deployed, working BMDsystem.

Vulnerabilities of High-AltitudeEndo Defense

The radar for the endoatmospheric Sprintcomponent of Safeguard tracked incomingRVS above 100,000-ft altitude, Because of anumber of technical problems associated withsuch high-altitude operation, U.S. BMD effortsin recent times have tended to focus on thelow-altitude regime below 50,000 ft.

Target tracking and discrimination at highaltitudes requires radars which are large andexpensive, These radars, which must for costreasons be few in number, would make tempt-ing targets for a concentrated precursor attackdesigned to overwhelm the defense in the areaof the radars and penetrate to destroy them.The defense system would then be blind.

I n addition to the vulnerability of the radars,high-altitude endo defense suffers from twocrucial technical problems: target discrim-ination and radar blackout. Discriminationrefers to the ability to distinguish RVS from thebus and tank fragments which accompanythem and from light decoys or other pene-tration aids which an attacker could design toconfuse the defense. The defense would wastecostly interceptors if the radar mistook adecoy or other object for an RV, and an RVwould leak through if it were mistaken for anonlethal object. High-altitude systems likethe Sprint component of Safeguard wouldhave high wastage and leakage because of theintrinsic difficulty of radar discrimination inthe upper atmosphere. In the thin air at highaltitudes, objects reentering the atmospherewithout heat shields, such as bus fragments,have not yet started to burn up, and lightdecoys fall at the same rate as heavier RVSThe dense air in the lower atmosphere, on theother hand, acts like a filter: unshielded ob-jects burn up, and light shielded objects slowdown. I n either case the heavy shielded RV can

be distinguished after it has reached lowaltitudes.

Blackout occurs when the heat and radia-tion from a nuclear explosion ionize the sur-rounding volume of air. This ionization causesattenuation and reflection of radar signalspassing through the affected region. At thehigh altitudes where the Safeguard radarstracked their targets, blackout over large areasof the sky could be created by a rather smallnumber of detonations. An attacker was there-fore encouraged to launch a first salvo of war-heads fuzed to detonate at high altitudes,thereby blacking out the defense’s radars. Thenuclear warheads on the defense’s own inter-ceptors could also produce this effect. The at-tacker could then bring in his main attackbehind the protective blackout “shield.”

Advantages and Limitations ofLow-Altitude Endo Defense

Because of the vulnerability and cost of theradars and the severe technical problems ofdiscrimination and blackout for high-altitudeendo systems, U.S. efforts in endo defensehave tended to focus on low-altitude systems,which track targets and perform interceptsbelow 50,000 ft.

Low-altitude systems are relatively imper-vious to decoy attack because it is possible toassess the weight of a body falling throughdense air from its radar return. Weight is astrategically significant discriminant, since of-fensive boosters have limited throwweight. Be-yond a certain point, loading decoys onto amissile requires offloading RVS, a trade thatbecomes unfavorable for the offense if the de-coys must be heavy in order to fool the de-fense, The trade is clearly absurd (leaving asidethe fact that a decoy might be cheaper than anRV) if the decoy must be as heavy as the RVitself, for the RV at least stands a chanceof penetrating the defense and explodingwhereas the decoy does not.

The procedure by which a low-altitude radarobtains a falling object’s weight is difficult foreven the cleverest decoy designer to sidestep


because it is based on fundamental principleswhich is not within the power of the offense toalter: the presence of dense air at low alti-tudes and some basic laws of physics. The rateof fall of an object — RV, decoy, bus fragment,etc. —through the atmosphere is determinedby the ratio of its weight to its area, called itsballistic coefficient, The higher the ballisticcoefficient, the faster the object falls. Of twoobjects of equal area, the heavier will fallfaster because it has more force of gravity toovercome the resistance, or drag, of the air. oftwo objects of equal weight, the smaller ormore streamlined will fall faster because itdoes not have to push as much air out of itsway. Thus, a flat sheet of paper falls slowlywhereas the same sheet, when balled up, dropsrapidly.

By tracking an object, a radar can measureits rate of slowdown and therefore the ratio ofits weight to its area. In the thin air at highaltitudes, however, differences in ballisticcoefficient do not lead to large differences inrate of fall because there is not much drag. Atlow altitudes the differences are quite pro-nounced. Thus, discrimination on the basis ofballistic coefficient is more reliable at lowaltitudes.

Measuring the ballistic coefficient might notbe sufficient for discrimination, however, sincea small light decoy could have the same bal-listic coefficient as a large heavy RV. It wouldin fact be quite difficult to design decoyswhich matched the balIistic coefficient of anRV at low altitudes since the shape of the RV(and hence its ballistic coefficient) changes ina complex way as its heat shield ablates. But asa hedge against a very carefully designeddecoy, the defensive radar can employ anothertechnique, involving the disturbance made inthe air as the body passes through it, to obtainthe area of the falling body. Combining thearea with the ballistic coefficient gives thebody’s weight, a quantity that is not in the in-terest of the offense to match. Thus a low-altitude defense system which made use ofthese radar discrimination techniques wouldbe vir tual ly impossible to sidestep withdecoys, since the fundamental discriminant is

weight and the techniques rely on the basicproperties of gravity and hydrodynamics.

Radar blackout is not a crippling problemfor low-altitude systems as it is for high-altitude systems.

However, fireball effects impose a basiclimitation on the effectiveness of low-altitudedefenses. The ability of low-altitude or “deependo” systems such as LoADS to make multi-ple intercepts within a short time over thesame site — a conventional missile silo or ashelter in an MPS system — is severely con-strained, no matter how many interceptors thedefense deploys. This limitation arises bothfrom blackout in the regions of nuclear fire-balls and from trajectory perturbations suf-fered by follow-on RVS passing through theseregions. The technical nature of this problem,and the extent of the I imitations it imposes, arediscussed further in the Classified Annex. Evenif a hypothetical future technology allowedthe defense to overcome this fundamentalIimitation, there might st i l l be strategiesavailable to the attacker that were more effi-cient than saturation, such as precursor attackon the defense itself or use of various penetra-tion techniques.

How Good is Good Enough?

It is an important feature of low-altitudesystems that only aim to make an attackertarget one more RV at each aimpoint that theydo not have to be very capable to force an at-tacker to pay this price. In fact, if the defenseis only good enough that it succeeds in makingits single intercept more often than it fails —how much more often is irrelevant–the at-tacker will conclude that he makes better useof his RVS by targeting two RVS at a lessernumber of defended aimpoints than by tar-geting one RV each at a larger number. The at-tacker’s conclusion is not a result of con-servative offensive perceptions but of sobercaIcuIation.

To take an explicit, if oversimplified, exam-ple, suppose an attacker has 1,000 RVS totarget at 1,000 aim points, each of which isdefended by a defense system whose goal is a


single intercept per aimpoint. Suppose alsothat the defense performs so poorly that it suc-ceeds in making an intercept only 51 percentof the time and fails 49 percent of the time.The attacker has the choice of targeting all1,000 aimpoints with one RV (Case 1) or 500aimpoints with two RVS (Case 2). In Case 1, theattack destroys 490 aim points because the de-fense fails this many times. In Case 2, all 500aimpoints targeted 2-on-1 are destroyed byassumption. Thus the attacker concludes thathe actually does better by “doubling up” on asmaller number of aimpoints (Case 2). But thisis exactly what the defense seeks to force himto conclude.

Therefore, if the odds that a single-shotsystem actually makes its intercept are greaterthan 50 percent, it achieves its goal of forcingthe attacker to target one more RV at eacha impoint. Whether the odds are 51 or 99 per-cent is immaterial, since the offense does nothave the option of targeting fractions of RVS ateach aimpoint, but only one or two.

Once the limited single-shot goal is ac-cepted, a relatively poor system is as good as aperfect one. Although low-altitude endo in-terception is a very challenging task, defensesystems do not have to perform it very well ifthey accept a goal of only one intercept peraimpoint. This stands i n contrast to exo de-fenses, which aspire to a higher attack pricethan one RV per aimpoint. Such defenses arenot worthwhile unless their performance isvery good.

In the example above, the attacker wasgiven the choice between l-on-l and 2-on-1targeting of ballistic reentry vehicles. Stylizedattacks or “reactive threats” involving non-ballistic RVS, precursor barrages, radar in-terference, etc. pose another set of challengesto single-shot defenses which must be ana-lyzed on a case-by-case basis.

The Need for Leverage

A generic low-altitude defensive system thatcouId only claim a single RV per defended sitewould not be effective unless some additionaldefensive leverage could be found. One U.S.

defense unit (radar plus interceptor) would bea poor cost trade for a single Soviet RV unlessintercept of this single RV resuIted in the sur-vival of a defended target valuable to theUnited States. But this would only be the caseif the number of targets were so large that theSoviets could not afford to target multiple RVSat each one. If the number of targets weresmall, the Soviets could attack each withmultiple RVS, overwhelm the defense, and de-stroy the U.S. value at an extra price, relativeto the undefended case, of a small number ofRVS. For instance, 100 single-shot low-altitudedefense units defending 100 silos containingMX missiles would only be able to claim 100RVS from a Soviet arsenal of thousands.

Additional leverage for the low-altitude de-fense could be provided in three ways.

Deceptive basing, such as for LoADS in asso-ciation with MPS, would allow a small numberof defense units to force the Soviets to expenda large number of RVS because they would notknow which shelters were defended and wouldhave to assume that all 4,600 shelters con-tained MX missi les defended by LoADS.Therefore, 200 LoADS defense units capable ofa single intercept each would be able to exacta price of 4,600 RVS, forcing the Soviets to at-tack each shelter twice for a total of 9,200 RVS.

A so-called “cheap” or “simple” defensesystem such as Swarm jet, to be discussed later,could conceivably improve the cost tradeofffor single-shot defense, but the overall attackprice would still be small if the number ofdefended targets was small, as with silo basing.If the simple system were very inexpensive,one could conceive of deploying one defenseunit with each shelter in a MPS system. Thiswould have the same effect as deceptive bas-ing without the need for PLU. There does notas yet appear to be a simple interceptionsystem cheap enough to allow this possibility.However, dust defense could be cheap enoughto deploy in this way.

Last, a capable “Overlay” defense operatingoutside of the atmosphere would also be apowerful source of leverage for an associated“Underlay” endo defense. The Overlay (if ef-


fective) would thin the attack and break up thestructured Iaydowns of RVS needed topenetrate the Underlay. The Soviets wouldhave to target many RVS at each defended sitein order that a few leaked through in the rightsequence to penetrate the Underlay. Such anattack strategy based upon leakage throughthe Overlay would be costly of RVS and ex-ceedingly risky for the attacker.

Because of the need for extra leverage, pro-posals of low-altitude defense for MX missileshave focused on deceptive low-altitude de-fense for a many-aimpoint MPS basing systemor on Overlay/Underlay (Layered] defense for aforce of MX missiles deployed in a small num-ber of conventional silos.

LoADS With MPS Basing

LoADS Description

THE DEFENSE UNIT (DU)The LoADS defense unit (DU) would consist

of a radar, data-processor, and interceptormissiles. The radar would be of the phased-array type, operating at high frequencies andwith high power and narrow beamwidth for ex-tra anti jam capability. The data processorwould employ distributed processing for rapidthroughput of large amounts of trajectorydata. The interceptor missiles, roughly onequarter the length of an MX missile and half aswide, would be capable of extremely high ac-celerations and rapid change of direction, Theinertially guided interceptor would be directedat launch towards a predicted impact pointwith the RV but its course could be updated inflight as well. The interceptor would be armedwi th a low-y ie ld nuc lear warhead . Thetechnologies embodied in these elements ofthe DU represent significant advances beyondearlier U.S. endo BMD systems.

For the purpose of LoADS/MPS combinationbasing, the elements of the DU would bepackaged into cylinders capable of fitting intothe same spaces in the shelters and trans-porters occupied by the MX missiles and simu-lators (see fig. 60). The DU, MX cannister, andsimulator would be so designed that they pre-sented identical signatures to sensors which

the Soviets might use to distinguish them in theshelters or in transit, It would be essential tothe effectiveness of the LoADS/MPS com-bination that it be impossible to distinguishMX, DU, and simulator.

One DU would be deceptively emplaced ineach cluster of 23 shelters, along with the MXmissile and 21 simulators. The DU would beprogramed to defend the shelter containin g

the MX missile. Upon receiving warning of aSoviet attack, the DU would erect vertically,pushing the radar face and the interceptor can-nister through the roof of the shelter (see fig,61). The DU would then be ready to defend theshelter containing the MX. Breakout would bean irreversible process, since it would destroythe roof of the shelter. Various schemes havebeen studied to avoid breakout. For instance,the DU could roll out the door of the shelterand erect like the MX missile. But the DU inthis exposed position would be too vulnerableto destruct ive ef fects of nearby nucleardetonations. The broken-out DU would stillhave the protective shieldin g and structuralsupport of the remainder of the shelter.

It would be absolutely essential that thedefense received adequate warning that SovietRVS were approaching so that it could awakeelectronic equipment from its dormpnt stateand break out. It appears that this process ofreadying the LoADS DU could be performed ina short period of time. If achievable, thiswouId mean that it would not be necessary tohave warning sensors which detected a Sovietattack at the moment of launch, but only asthe attacking RVS approached the UnitedStates. This late warning would be easier toprovide than the early warning required to sup-port launch under attack or exo BMD. It wouldalso be easier to protect warning sensors ofthis type from a Soviet precursor attack. itmight also be desirable to have some informa-tion ,about the size of the attack before a deci-sion were made to break out. (This is discussedfurther in the context of Shoot-Look-Shoot inthe Classified Annex. ) Finally, the command,control, and communications to support time-ly breakout would require procedures andhard\ware immune to a determined Soviet ef-


Figure 60.— LoADS Defense Unit BeforeBreakout (human figure indicates scale)

Radar Batteries Data processing Interceptors


fort to disrupt them. Several technically feasi-ble approaches to these problems have beenproposed, and their provision would be essen-tial to effective defense.

LoADS OPERATIONThe LoADS DU in each cluster would be pro-

gramed to defend the shelter containing theMX missile, Since the Soviets would not knowwhich shelter contained the MX if PLU weremaintained, they would have to assume fortargeting purposes that each of the 23 shelterscontained an MX missile defended by LoADSLoADS would intercept the first RV attackingthe MX shelter, so the Soviets would have totarget each shelter twice in order to destroy theMX. LoADS would double the price the SovietswouId have to pay for an MX missile from 23 to46 RVS. Thus U.S. deployment of LoADS wouldbe essentially equivalent to doubling the num-ber of shelters in the MPS deployment whilekeeping the number of missiles the same.

deployment while keeping the number ofmissiles the same.

It is desirable for each DU to have morethan one interceptor in order that it could de-fend itself if it came under attack before theMX shelter did.

It would be essential that the location of theMX be unknown to the Soviets. It would alsobe necessary to conceal the location of theDU, since if this were known the Soviets couldattack the defense first, forcing it to use up itsinterceptors in self defense, Subsequent attackon the other shelters would find them un-defended.

LoADS WITH VARIANTS OF MPS

The operation of LoADS would be essen-tially unchanged if the MPS deployment wereorganized into “valley clusters” containingseveral missiles instead of discrete clusters of


Figure 61 .— LoADS Defense Unit AfterBreakout (human figure indicates scale)


23 shelters for each missile. A DU could still beprovided to defend each missile, and the at-tack price per missile would again be doubled.

From the point of view of LoADS defense,there would be significant tradeoffs betweenhorizontal and vertical shelter deployment butno clear reason to prefer one to the other. Forvertical shelters, it would be necessary to putthe radar and missile cannister in differentshelters, since they would be too large to fitside-by-side in a single shelter. There wouldhave to be a data link to connect the twoelements of the defense unit. Since the unitswould be moved from shelter to shelter peri-odically, the communicat ions equipmentwould have to connect all pairs of shelters,potentially a costly addition. The links wouldfurthermore have to be resistant to disruptionfrom nuclear effects. On the other hand,

breakout would not be required, since the de-fense could egress through the blast door ofthe vertical shelter. Matching four objects (MX,simulator, radar module, and interceptor can-nister) would be more difficult than three, butthere would be more design flexibility for theseparate radar module and interceptor can-nister because each would be, so to speak,“half empty. ” The extra room could be usedfor PLU countermeasures. Protecting the DUelements from nuclear effects could con-ceivably be easier for vertical shelter de-ployment.

It is not possible at this time to assess thesetradeoffs in detail, but it is not apparent thateither vertical or horizontal offers clear ad-vantages. More study has been made of thehorizontal alternative.

Ch. 3—Ballistic Missile Defense 121

LoADS Effectiveness

Active defense systems are very complex:the interception process is complicated, withmany distinct sources of leakage and wastage.There are many attack scenarios, offensivecountermeasures, and defensive counter-countermeasures to consider Analysis of theeffectiveness of a BMD system can thereforebe more involved than analysis of basingsystems that ensure survival of MX by passivemeans such as mobiIity, concealment, ordeception It is therefore important in assess-ing how well LoADS would do its job to be veryclear what that job is

Suppose LoADS sought to double the pricethe Soviets would have to pay to destroy anMX missile from 23 RVs to 46 RVs, In this case,LoADS would have the rather modest task ofintercepting the first RV targeted at the MXmissiIe within each cluster I n order to destroythe MX missile within a cluster, the Sovietswould have to target two RVS, or “double up, ”at each shelter, This assumes that PLU wouldbe successful and the Soviets would have noknowledge of the location of the MX or theDU.

In fact, LoADS could exact the price of 2RVS per shelter even if the defense systemwere rather inefficient. Roughly speaking, ifthe Soviets believed that LoADS would suc-cessfully intercept the first RV targeted at theMX shelter more than half of the time— that is,if the efficiency of LoADS were greater thanonly 50 percent — then the Soviets wouldcalculate that they made better use of theirRVS by doubling up on fewer shelters than bytargeting many shelters with one RV,

For example, suppose that the Soviets had6,900 RVS to target at 4,600 MPS shelters,(These numbers are chosen to make the arith-metic easy and for no other reason, ) Supposealso that LoADS were only 51-percent efficientin a 1 -on-1 attack: that is, if one RV weredirected against every shelter, LoADS wouldsuccessfully intercept 51 percent of the RVSdirected at the shelters containing MX. This isthe same as a leakage of 49 percent. Assumealso that all targeted Soviet RVS actually

arrived on target and further that if two RVS ar-rived at the MX shelter within a short space oftime, LoADS would not even attempt to inter-cept the second and the MX missile would bedestroyed,

The Soviets would have the choice of usingtheir 6,900 RVS either to target 100 clusters(2,300 shelters) with one RV and 100 clusters(2,300 shelters) with two RVS (Case 1 ) or to dou-ble up on 150 clusters (3,450 shelters) and leave50 clusters (1,150 shelters) untouched (Case 2).In Case 1, all 100 MX missiles targeted 2-on-1wouId be destroyed, and 49 of the missilestargeted l-on-l would be destroyed becauseLoADS would only be 51-percent efficient byassumption. Thus in Case 1 the Soviets wouIddestroy 149 MX missiles, In Case 2, the 150missiles targeted 2-on-1 would be destroyedand the remaining 50 untouched The SovietswouId therefore actualIy destroy more MX mis-siles by doubling up (Case 2), even thoughLoADS failed to make an intercept almost asmany times as it succeeded.

It therefore appears that LoADS would nothave to be very efficient to exact a price oftwo RVS from the Soviets. At the same time, itwouId be exceedingly difficuIt to exact a priceof several RVS.

So far, the analysis has considered only sim-ple 1 -on-1 or 2-on-1 attacks. The conclusion isthat, as far as these attacks are concerned, andassuming the DU survives nuclear effects to doits job and that PLU is maintained, it is possibleto have confidence that LoADS is capable ofits job. Although low-altitude interception ofRVS is a very challenging technical task, andthere are many uncertainties about LoADS op-eration and potential contributors to ineffi-ciency (radar and interceptor performance, RVradar cross sections, radar traffic handling,kill mechanisms, etc.), there are none whichshould stop LoADS from doing its job as wellas it needs to,

If the defense only sought to make one in-tercept over the MX shelter, then the UnitedStates could assume that the Soviets wouldpay the price of 46 RVS per MX missile if giventhe choice of l-on-l or 2-on-1 targeting. Could

—


t hey do be t te r by us ing spec ia l a t tackstrategies?

There are many such reactive threats toLoADS, For instance, decoys are a hypo-thetical threat: precision decoys seek to foolthe radar into intercepting them, while trafficdecoys simply aim to fool the radar longenough to consume precious data-processingtime. As discussed in the Technical Overview,the ability of radar to weigh falling objects atlow altitudes means that decoys are probablynot a serious threat to LoADS. Jammers de-ployed along with attacking RVS could seek toblind the radars. Maneuverable reentry vehi-cles (Ma RVs) could try to evade the inter-ceptor; and if MaRVs were provided withradar- homing devices, they might destroy theLoADS defense units before they had donetheir job. These reactive threats are discussedin the Classified Annex. Defense barrage,blackout, and exhaustion attacks are discussedunder Hardness to Nuclear Effects and its Clas-sified Annex, and Spoof and Shoot-Look-Shoot, both threats to deception, are discussedunder Preservation of Location Uncertainty[PLU) and its Classified Annex.

One can raise legitimate questions as towhether a prudent Soviet planner would useany of these techniques to sidestep LoADS, butthe defensive planner must fortify the systemdesign against all of them. The attractivenessof these special threats to Soviet plannerswould presumably be weighed against thebenefits they would derive from the simple ex-pedient of deploying two Soviet RVS for everyU.S. shelter. Detailed analysis of these specialthreats, presented elsewhere in this report orits Classified Annexes, indicates that some ofthem are worrisome and represent a long-termrisk to the effectiveness of LoADS/MPS.

Hardness to Nuclear Effects

The close shelter spacing–1 mile–meansthat LoADS must operate in a nuclear en-vironment of a severity unprecedented for socomplex and exposed a piece of equipment.Failure to meet the requirements could lead topronounced degradation in system perform-ance. It is also vital that measures taken to pro-

tect the DU do not betray its location, i.e.,break PLU. Providing for nuclear hardness re-quires detailed understanding of the expectednuclear environment and its effect on criticalmechanical and electrical components. Espe-cially important for LoADS, given the unprec-edented character of the hardness require-ments, is testing of actual equipment. DU de-sign and nuclear effects analysis—and, in thecase of LoADS, PLU analysis — must proceed inconcert. These studies are just beginning. Test-ing is required before it wilI be possible to haveconfidence that LoADS can meet its hardeningneeds, especialIy within the severe design con-straints imposed by PLU.

AS with the analysis of system effectiveness,it is important to have a clear idea of LoADS’hardening needs and of the consequences offailing to meet these needs. The key require-ments concern the survival of the DU, andespecially the radar, after it has broken out ofthe shelter and is waiting to intercept the RVtargeted at the MX shelter. Other concerns,probably less serious, are the hardness of theinterceptor as it flies to make its intercept andthe hardness of the DU before it breaks out.

HARDNESS OF THE DU AFTER BREAKOUT

For LoADS to do a single-shot job, no lessthan 46 Soviet RVS may suffice to destroy anMX missile. The attacker must either be madeto fail to destroy the DU before it has made itsintercept or be made to pay a heavy enoughprice to destroy the DU that nothing is gainedby trying.

The hardness of the broken-out DU defines a“keep-out zone” around the unit: RVS whichdetonate within the keep-out zone are as-sumed to destroy the DU and must be inter-cepted if they arrive before the DU has madeits intercept above the MX shelter. It is for self-defense that each DU should contain morethan one interceptor missile.

Inadequate nuclear hardening would meanthat the keep-out zone was too large. An il-lustrative, if presumably exaggerated, exampleconsists of a DU so soft that a detonationanywhere within its shelter cluster would im-


pair its function. In this case, the Soviets couldtarget a few RVS (perhaps of higher yield thanthose targeted at shelters) to arrive at randomlocations within each cluster a few secondsbefore arrival of the main attack. The main at-tack would consist of one RV on each shelter.The DU would have no choice but to interceptall of the precursors, for otherwise it wouId berendered inoperable, If there were as manyprecursor RVS as interceptors in the DU, thenall the interceptors would be used up in self-defense, The main attack would then find thecluster undefended, as though LoADS did notexist. The attacker would then have paid not46 RVS, but rather the undefended price of 23RVS plus just a few additional RVS to exhaustthe defense,

The defense suppression barrage describedabove is one of several scenarios where LoADShardness plays a crucial role, In all of thesescenarios, the attacker seeks to destroy an MXmissile for an attack price of less than 46 RVS,The results of a more detailed analysis, pre-sented in the Classified Annex, indicates thatthe 1-mile shelter spacing imposes severe re-quirements on the DU. Unlike the MX missile,protected by its steel and concrete shelter andseveral feet of earth, the DU is directly expos-ed to the nuclear effects. Not only must theDU survive, but its complex components mustfunct ion through the attack. Thus, someeffects— prompt radiation, certain effects ofelectromagnetic pulse (EM P), dust, etc. —which are not important for missiIe protectionare severe threats to LoADS. Defense per-formance, measured by vulnerability to thesestylized attacks, might be degraded ap-preciably by shortcomings in hardening.

Work on LoADS hardening so far has con-centrated upon defining quantitatively thenuclear effects which the DU must be able toendure, not providing design fixes for potentialvulnerabilities, Even defining the effects willrequire testing, since in some respects they ex-ceed the predictive power of computer simu-lation codes. Understanding the interaction or“coupl ing” of these effects to the peculiargeometry of the broken-out DU, to electronics,

and to radar performance wiII also requiretesting.

Nothing that is done to ensure its hardnessmust permit the DU to be detected when it is inthe shelter. If the Soviets were able to detectwhich shelter contained the DU, they couldtarget that shelter with a few precursors, forc-ing it to exhaust itself in self-defense. This andother threats to PLU are discussed in the nextsection. The important point is that hardeningthe DU —adding shielding, structural support,etc. — must not provide a signature whichwould allow the Soviets to detect the DU’Slocation. This synergism of hardness and PLUis a matter of testing and detailed design whichhas not yet been done.

Ensuring adequate hardness for the broken-out DU is thus a chalIenging task, and it wiII re-quire some time ‘before uncertainties can bereduced to levels where a final judgment ispossible.

It is important finally to note the constraintsthat would act upon the offense if it were toseek to exploit potential vulnerabilities inLoADS. If the Soviets were to fractionate so asto be able to target as many shelters as pos-sible, they would have to reduce the yields oftheir RVS. The lower yields would significantlyalleviate the nuclear effects on LoADS insome, though not all, circumstances. If on theother hand, the Soviets kept their yields highwith the aim of exploiting potential LoADSvulnerabilities, it would be difficult for themto fractionate their missiIes,

HARDNESS OF THE IN-FLIGHT INTERCEPTORAs the missile flew towards its intercept

point, it would be buffeted by the shock wavesfrom nearby detonations. Though the inter-ceptor has the ability to correct its course, ithas a limited duration of powered flight. Ifintercepting an RV at a relatively distant point,burnout would be complete before the inter-ceptor reached the RV, When coasting in thisway, it would have less ability to correct itscourse than when burning.

Interference with interceptor performancedue to nuclear effects such as shock waves is


one of the many contr ibutors to systemleakage. As described in the previous section,LoADS can tolerate a large leakage withoutimpairing its overall effectiveness. Thus in-flight nuclear effects might not serve to in-crease leakage above an acceptable point.However, interceptors flying out to attemptmultiple intercepts would be forced to fly in aseverely disturbed environment.

HARDNESS OF THE DU IN THE SHELTERIn the context of a Spoof or Shoot-Look-

Shoot attack, to be discussed in the next sec-tion, the DU might be required to survive aIight precursor attack before it broke out of itsshelter. I n this situation the DU would berelatively secure because it would be in theshelter and the scenario cal Is for a Iight attack.

Additional discussion of the problems withmeeting LoADS’ nuclear hardening require-ments can be found in the Classified Annex.

Preservation of Location Uncertainty (PLU)

Successful deception is vital to LoADS’defensive leverage. If the location of the DUwere known to the Soviets, they could exhaustthe defense with a precursor attack. A sub-sequent one-on-one attack would find theshelters completely undefended, What ismore, under certain circumstances, a break-down of PLU for the LoADS DUS could cause abreakdown of PLU for the MX missiles as well.In this case, the United States would be worseoff than if there were no defense at all.

For undefended NIPS, PLU appears to be acomplex and challenging technical enterprise,but no signatures of the MX missile have beenidentified which present clearly insurmount-able problems. PLU for the LoADS/MPS com-bination has not yet progressed this far, andthe problem will have to be reduced to a com-parable “acceptable” level of detail. In par-ticular, the design requirements imposed bynuclear hardening must be taken into account.

Even i f no “Achi l les’ heel , ” or grosssignature of the DU which is fundamentally in-compatible with PLU, is found, a complexengineering task faces the LoADS designer. In

the case of MPS alone, one is presented with200 missiles and the task of creating 4,600simulators which resemble the missiles in allobservable respects. The simulator is createdde novo r with no a priori constraints save tomatch the MX. The LoADS Defense Unit, onthe other hand, is a functional object withunique signatures, related to its operation,which cannot be suppressed. It wouId there-fore be virtually impossible to make the DUmatch a set of missiles and simulators whichwere not designed with the LoADS option inmind. The three objects — MX, simulator, andDU — must all be designed in concert.

PLU is therefore considerably more complexfor MPS defended by LoADS. It is too early totell whether deception can be arranged at all,but it is probable that the 200 missile can-nisters and 4,400 simulators would have to bealtered from ‘time to time as design and testingproceeded to accommodate distinctive fea-tures of the DU. The later that a decision weretaken to give LoADS a place in the design ofthe overall system, the riskier and more costlythe PLU process might become.

in addition to signatures, the operations bywhich the MXS and DUS were shuffled peri-odicalIy among the shelters must not betraythe location of either. It appears that accept-able “movement algorithms” can be devisedto preserve PLU for both MX and DU simul-taneously, whether the system were organizedinto individual clusters of 23 shelters or intolarger “valIey clusters. ” it should be noted thatif rapid reshuffle were required to redress ac-tual or suspected loss of PLU, extra time mightbe required, depending on the availability oftransporters, to move the DU as well as the MXmissiI e.

There is some concern regarding a tactic forattacking LoADS/MPS, called Shoot-Look-Shoot, whereby the Soviets could in principleinduce a breakdown of PLU. I f the Sovietslaunched a first wave of attacking RVS whichcaused the LoADS DUS to break out and ex-pose their locations to remote Soviet sensors, asecond wave could be targeted on the basis ofknown DU locations. They would then be able

Ch. 3—Ballistic Missile Defense 125

to destroy MX missiles at about the unde-fended price or even less. However, to sidestepthe defense in this way, the two waves of at-tacking RVS would have to be well-separatedin time. The Soviets wouId therefore have toreckon with the possibility that the UnitedStates wouId simply launch MX missiIes at theSoviet Union between waves rather than awaitthe outcome of a subtle Soviet strategy. Forthis and other reasons, reliance on Shoot-Look-Shoot would entail high risk to the Soviets, TheSoviets would presumably weigh these risksagainst the simpler expedient of buiIding moreRVS and attacking the shelters directly

There are two scenarios for which a SovietShoot-Look-Shoot capability could be in-tended. In the first (sometimes called “Spoof”)scenario, an initial attack on LoADS/MPS is in-tended to cause the DUS to break out of theirshelters. A second wave of RVS is then targetedon the basis of known DU locations. I n the sec-ond case, appropriate to long war scenarios, asecond attack is not necessarily planned at thet i me of the first, but after an initial exchange inwhich the DUS had broken out to perform theirdefensive job, the Soviets could sense the loca-tions of the exposed DUS. I n a subsequent ex-change, the remaining U.S. force would beleft essentially undefended. Shoot-Look-ShootwouId in this case mean that the LoADS de-fense did not have the endurance that the MXmissiIes themselves wouId have.

Since deception is necessary for defense ef-fectiveness, since the defense unit must breakout and expose itself to defend, and since theSoviets would know that a shelter defended inthe initial attack must have contained an MXmissile, a Shoot-Look-Shoot strategy wouldenable the Soviets to compensate for LoADSwith only a few hundred more RVS than theywouId need to attack an undefended MPS, pro-provided the U n i ted States did not Iaunch out be-tween the first and second waves CaIcuIationsof the outcomes of various Shoot-Look-Shootscenarios, and a discussion of the problemsfaced by the offense in mounting them and thedefense in countering them, are provided inthe Classified Annex.

Cost and Schedule

OTA has not performed independent costand schedule analysis for the LoADS/MPScombination. The data presented in this sec-tion were supplied to OTA by the Army’s Bal-listic Missile Defense Systems Command(BMDSCOM). Comments that accompanythese data are those of OTA and do not nec-essarily reflect opinions of BMDSCOM.

The Army’s most recent (October 1980) costestimate to deploy a LoADS defense for the4,600-shelter Air Force baseline MPS systemand operate it for 10 years is 8.6 bilIion con-stant fiscal year 1980 dollars. The $7.1 bilI ionacquisition cost would include the costs of theDUS, 200 separate transporters to move theDUS, a modest amount of construction ofoperating buildings in the deployment area,and program development and management.Operating costs are estimated at $153 millionper year. A detailed breakdown is presented intable 20. These cost estimates do not includethe costs of potential modifications to the AirForce baseline system in order to accom-modate LoADS nor the cost of additional tac-tical warning and threat assessment systemsand command, control, and communications(C) systems to support LoADS.

The present LoADS Program schedule isfunding- and Treaty-constrained, and preciseschedule information is classified. A schedule

Table 20.—Army’s LoADS Cost Estimate,October 1980’

(constant fiscal year 1980 dollars in billions)

Research, development, testing, andengineering . . . . . . . . . . . . . . . . . . . . . . 1.75

Defense units . . . . . . . . . . . . . . . . . . . . . 3.20Transporters . . . . . . . . . . . . . . . . . . . . 1.13Military construction . . . . . . . . . . . 0.16System engineering and program

management . . . . . . . 0,54Other investment . . . . . . . . . . . . . . . . . . 0.32

Acquisition cost . . . . . 7.10Operations cost (10 years @ 0.15) . . . . . . . . . 1.53

Total . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.63

aFrom figures supplied by Army BMDSCOM

SOURCE Off Ice of Technology Assessment


that assumes that the decision to remove con-straints were not made until late in the decadeprovides for final operating capability (FOC)for the LoADS addition to MPS several yearsafter MPS FOC. This schedule would not re-quire amendment of the ABM Treaty reachedat SALT I until later in the decade.

An accelerated schedule, assuming an earlydecision and release from constraints, couldprovide for LoADS deployment on about thesame schedule as MPS deployment. This wouldrequire early amendment or abrogation of theABM Treaty and funding well above that nowprojected for the LoADS Program.

Other Endo Concepts

Other endo BMD concepts besides LoADShave been proposed and investigated. Dustdefense is technical ly feasible and verycapable but could have very low public appealas welI as a few potential technical drawbacks.Terminal or low-altitude defenses based on“simple” or “novel” concepts could be ade-quate as single-shot last-ditch defenses ofhardened targets against a small attack buthave not been proposed with the demandingMX role in mind. The Army’s Site Defense rep-resents the technology of the 1970’s and is in-adequate for the MX role.

Dust Defense

Dust defense–also called environmentaldefense – provides for burying “clean” nuclearweapons in silo or MPS fields and explodingthem shortly before attacking RVS arrive. Thedust and debris lofted into the air would de-stroy approaching RVS either by direct colli-sion with large earth fragments or by dust ero-sion of the RV’S heat shield. The detonationswould be placed so as not to damage theICBMS in their silos or shelters.

There are two possible ways of employingdust defense. In the first, nuclear weaponswould be buried north of each silo or shelterand exploded seconds before RV arrival. Smallradars placed north of each site would providethe detonation signal. The RV would be de-stroyed in passage through the dense plume of

debris thrown up immediately by the explo-sion. The dust cloud which forms a little laterat higher altitudes would provide additionalprotection for a longer period of time than thedebris stem, which falls back to Earth in a shorttime.

In the second scheme, a smaller number ofweapons of higher yield would be explodedthroughout the fields several minutes beforeRv arrival, The heavy debris would thus havefallen by the time the RV arrived, but by thattime the dust cloud would have formed. Sincethe dust cloud from a high-yield weapon canbe tens of miles in width and breadth, manysiIos couId be protected by a single dust cloud.Protection would last for approximately 20minutes after which another set of weaponswould have to be detonated to provide contin-uing protection.

The weapons detonated would destroy farmore megatonnage than they const i tutedthemselves, a fact which makes the deliberatedetonation of nuclear weapons on U.S. ter-ritory somewhat more palatable from thestandpoint of fallout. But a more importantfactor in reducing fallout is the possibility,much discussed in the 1960’s at the time of thePLOWSHARE Program studies of the peacefuluse of nuclear explosions, of constructingnuclear weapons which produce very Iittleresidual radioactivity.

Conventional nuclear weapons give rise toradioactive fallout in two ways. First, a certainfraction of the weapon yield is produced byfission. The fission products are unstable iso-topes which give off harmful radiation whenthey decay into more stable species. The restof the weapon yield is provided by fusion.Large numbers of neutrons are formed in thefusion process, and when these neutrons en-couunter ordinary material in the vicinity of thedetonation, they transform it into radioactivematerial.

Clean weapons reduce both sources of fall-out. First, the clean weapon is constructed insuch a way that very Iitle of the yield is due tofission, Second, one can surround the weapon

with material, such as berated water, which ab-


sorbs the fusion neutrons readily without be-coming radioactive Such a clean bomb is notas compact as an ordinary nuclear weapon. Itmight occupy the volume of a room. For thatreason, underground vauIts wouId be dug inthe silo fields to house the clean weapons. Inthis way the radioactivity from the clean explo-sions could be reduced to about one o n e -hundredth of the radioactivity from conven-tional nuclear weapons of equal yield

Though there is some uncertainty in thecomposition of the stems and dust cloudsformed in nuclear explosions (which cannot beentirely resolved within the Test Ban Treaty),there is general agreement that dust defense isan effective way to destroy attacking RVSThere appear to be no effective measures thatthe Soviets could take to protect their RVS.Moreover, large numbers of RVS could be de-stroyed within a short space of time i n thisway, a feat that is impossible for more conven-tional endo defenses

Potential drawbacks to dust defense, be-sides its perceived unpalatability, are the needfor warning, the need to provide multiple ex-plosions if the attack occurs in waves well-spaced in time, and the fear of error

The cloud variety of dust defense requireswarning because the weapons must be deto-nated sever-a I minutes before attacking RVS ar-rive. I n principle, the stem variety does not re-quire warning beyond that provided by itsradar, but it might be considered inadvisableto keep the system activated at all times, sincea radar malfunction in peacetime might causeinadvertent detonation Since warning is notneeded until late in the flight of the attackingRVS in either case, it is easier from the tech-nical point of view to provide an adequate sys-tem of this type than one which provides warn-ing at the time of missile launch. This type ofwarning is needed by alI endo defense systems.

An attack that came in waves could requiremultiple detonations. If backup weapons wereburied to provide the capability for multipledetonations, the weapons would have to bespaced far enough apart that the first detona-tion did not destroy the remainder One U.S.

response to a muItiple-wave attack would sim-ply be to launch in retaliation rather thanawait the next wave, Offensive missiles can bemade that can launch through dust cloudswithout damage. Dust defense could thereforeextend the timeline for launch under attack byforcing the Soviets to attack in two waves. Thefirst wave would be destroyed by the dust, andthe second wave could not be launched untilthe dust cloud had dispersed. The UnitedStates would have this extra time to decide ona reponse.

Error in the form of inadvertent or unau-thorized detonation of the buried weaponscouId be avoided by the same set of proce-dures which prohibit launch of offensive mis-siles. The real possibility of error Iies in a falsewarning message causing authorized detona-tion Fear of this type of error and proceduresto avoid it could lead to another type of error:failure to authorize detonation when the warn-ing information was correct, The problemshere are similar to those of a launch underattack system.

Dust defense could therefore be by far themost potent endo defense system. However, itis seldom taken seriously because of concernfor public reaction.

Simple/Novel Systems

Simple/novel systems is a catch-all for awide var iety of low-al t i tude or last-di tchdefenses of hardened targets. Examples go bysuch names as Swarm jet, Porcupine, GatlingGuns, SID CEP, Quickshot, SSICM, Bed ofNails, and Agile, The interceptors consist ofrockets, shells, or inert projectiles with orwithout nuclear warheads and guided by land-based radars or homing sensors. Not all aresimple, though many are novel indeed. LoADSitself could be classed with these systems,since it has a similar goal.

Because low-altitude defenses cannot guar-antee muItiple intercepts over a single target inrapid succession, they are inadequate to de-fend a small number of targets against a largeSoviet threat Indeed, most simple/novel sys-tems were conceived as cheap and quickly de-


ployable ways to increase the Minuteman at-tack price and create uncertainty for Soviettargeters.

Some simple/novel systems might thereforehave capabilities similar to LoADS, thoughnone has been studied in the depth that LoADShas A simple/novel system might therefore inprinciple be capable of replacing LoADS in theMPS basing role. This could come about eitherby providing a last-ditch system simple andcheap enough to deploy in association witheach of the 4,600 MPS shelters, or compactenough to fit into a shelter deceptively I ike theLoADS Defense Unit, However, none of theconcepts yet proposed combine confidence intechnical feasibiIty with low cost or smal I sizesuch that they would be attractive replace-ments for LoADS in the MPS role.

Because of the interest in simple/novel sys-tems, two of the most promising examples arediscussed briefly below,

SWARM JET

The Swarm jet concept consists of radars de-ployed north of each defended site and alauncher located near the site and containingthousands of spin-stabilized, rocket-propelledprojectiles, When the radar detects an attack-ing RV, the launcher pivots in the direction ofthe predicted intercept point and the projec-tiles launch into the threat tube in a swarm.Each projectile weighs a few pounds and is de-signed to destroy an RV completely in a hyper-velocity COII ision seconds before arrival at thesilo Swarm jet is designed to be constructedfrom already-available or easily manufacturedcomponents.

The object of the defense is to fill the sky inthe path of the attacking RV with enough pro-jectiles to assure high probability of collision.Though there is agreement among those whohave studied Swarm jet that collision with aprojectile will indeed destroy an RV, there isdisagreement about how may projectiles areneeded to assure a high collision probability.This disagreement translates into uncertaintyin the size and cost of an effective Swarm jetdeployment. Factors that enter into the uncer-

tainties are radar performance, the pointingand aerodynamic properties of the projectiles,and the effects of blast waves from precursoror nearby nuclear detonations.

Like other low-altitude defenses, Swarm jet isessentially a single-shot system and couldtherefore claim with confidence only one RVper si lo from an attacker. The Swarm jetlauncher might be too large to fit into an MPSshelter; if this were the case, the only way todeploy it with MPS basing would be to provideone Swarm jet unit for each shelter. This wouldbe costly but might deserve consideration ifdeception proves too cliff icult for LoADS/MPS.

AGILE INTERCEPTOR

The idea of an Agile interceptor is to getbeyond the single-shot limitation of low-alti-tude systems by providing an interceptor somaneuverable that it can intercept follow-onRVS after detonation of a first despite poorradar impact-point prediction due to firebal isand despite being thrown off-course by blastwaves and winds, This program is i n the re-search stage.

The goal of the Agile interceptor is to in-tercept a few, but not many, RVS over a singlesilo. Because its goal remains modest and be-cause the technology is yet unproved, this con-cept is considered unsuitable for MX defense.

Site Defense

The Army’s Site Defense is a derivative ofthe Sprint component of the Safeguard de-fense system of a decade ago. As a high-altitude endo system, it is susceptible toblackout, penetration aids, and direct attackon its few, large radars, as described in the7-echn;ca/ Overview. Based on the technologyof the 1970’s, Site Defense is preserved as anoption in the event of a decision to field aBMD system based on known technology in ashort period of time.

Though inadequate for the role of MX de-fense, Site Defense could be appropriate forother BMD roles. For instance, it could be usedas a “threshold defense” for some importantU.S. assets such as warning sensors. In the con-


cept of threshold defense, no pretense is made tempt to destroy the defended targets wouldthat the defense can ensure survival of the de- require such a Iarge attack as to constitute afended asset; its role is rather to increase the major provocation deserving of major U.S. re-attack price to the point where a Soviet at- sponse.

EXOATMOSPH ERIC AND LAYERED DEFENSE

Technical Overview of Exo BMD

Exoatmospheric — “exe” — defense holdshigh promise in theory because the long flighttimes of RVS outside the atmosphere and thelarge battlespace mean that many RVS tar-geted at the same site can be destroyed. Incontrast to low-alt i tude systems, systems withan exo component could in theory defend asmall number of targets such as MX silosagainst a large number of Soviet RVS, Addi-tional strengths of exo BMD are the feasibilityof nonnuclear kill, the posslbility of mountingmultiple interceptors on a single boosterrocket, and the concept of adaptive preferen-tial defense

Theoretical Advantages to Exo Operation

Nonnuclear kill is possible in space for sev-eral reasons First, the defensive sensors wouldhave a relatively long time–minutes, as op-posed to seconds for an endo defense— totrack their targets, and the trajectories ofattacking RVS would be predictable becausethey would be passing through empty space. Itwould therefore be possible for the interceptorto aim close enough to the RV that the largedestructive radius of a nuclear warhead wouldnot be necessary Deployment of barriers ofmaterial in the paths of approaching RVSwouId aIso be easier in the vacuum of space.Nonnuclear k i l l is preferable to nuclearmethods because a nuclear defense’s ownwarheads couId interfere with its sensors(assuming the offense did not employ its ownnuclear precursors), nuclear warheads arerelatively expensive and heavy, and activationof a nuclear defense would require proceduresfor authorized nuclear release.

Interceptors boosted into space for exo de-fense could also carry many individual killvehicles — much as a MIRV’d missiIe carriesmany RVS — resuIting i n some savings consider-ing the cost of putting defensive vehicles intospace in the first place. Multiple warheads onthe same interceptor are impractical for usewith in the atmosphere, where the engagementtimelines are too short to make multiple de-ployments feasible.

Preferential defense is a tactic for multiply-ing the effectiveness of a defensive system if itis only required to defend a subset of the tar-gets under attack, For instance, suppose MXmissiles were deployed amongst the six Min-uteman wings and that survival of the missilesin two of these wings against a Soviet attackwas considered a sufficient goal for the de-fense. The defense could then concentrate itsexo interceptors upon destroying RVS targetedat the two defended wings and abandon theother four wings to the attacker. Which twowings were chosen for heavy defense could bekept secret from the Soviets or decided by thedefense at the last minute, In their targetingplanning, the Soviets would be unable to con-centrate their RVS on the defended wings: theywould either have to do their targeting asthough all the wings were heavily defended orgrant the defense its goal of two survivingwings. Adaptive preferential defense thereforeeffectively multiplies the number of defensiveinterceptors, In this example the Soviets wouldbehave as though all six wings were defendedas heavily as the two singled out by the UnitedStates. Adaptive preferential defense is not aneffective tactic for endo systems because endointerceptors must be located near to the tar-gets they are defending. The presence of the


defense near the defended sites therefore givesthe game away.

Infrared Sensing

An exo system that used large ground-basedradars to acquire targets outside of the at-mosphere could be blinded by direct attack orhigh-altitude blackout, and the view of ground-based optical sensors, which would be inade-quate in any event, could be obscured byclouds and dust. It is therefore desirable to putthe defensive sensors into space, either on theinterceptors themselves or on other space vehi-cles. Space-based radars would be heavy,costly, and susceptible to jamming. For thesereasons, many exo concepts employ space-borne passive infrared sensors, which are rela-tively light and compact, However, infraredsensors are susceptible to offensive counter-measures such as decoys and other penetra-tion aids.

Layered Defense

Like endo systems, exo defense alone wouldbe indequate to defend a small number of tar-gets against a large number of attacking RVS.In this case, the reason is not saturation of thedefense, but the cumulative effect of leakage.It only takes one leaker to destroy a silo. Ifmany RVS were aimed at each silo, the oddsthat one would get through could be high evenif the probability that each individual RV wasintercepted were high. The defense could at-tempt to stanch this hemorrhage of leakers byattempting to intercept each RV more thanonce (assuming that the multiple interceptorvehicles targeted at the same RV would not in-terfere with one another). This tactic could beeffective but would drive the defense to enor-mous arsenals of interceptors.

If an endo defense were associated witheach silo, the combined exo/endo LayeredDefense would be more effective. The endosystem could catch leakers from the exo sys-tem and, moreover, the exo system would im-prove the performance of the endo systemsince it would break up the concentrated,

structured attack patterns which saturate endosystems. Thus, exo (Overlay) and endo (Under-lay) defenses in a Layered combination have asynergistic effect wherein the principal I imita-tion of each is alleviated by the presence ofthe other. An endo defense could also help toprotect the launch sites for the Overlay inter-ceptors from a disabling precursor attack. Itwould also be difficult for an attacker to de-signl decoys to confuse both the Overlay sen-sors and the Underlay sensors. However, sincedecoys are ineffective against low-altituderadar sensing anyway, an attacker would prob-ably concentrate his penetration aids againstthe Overlay and not try to fool both layers.Last, the tactic of adaptive preferential de-fense for the Overlay loses some of its attrac-tiveness when there is an Underlay because theUnderlay cannot adapt: it can only defend thearea (or individual silo) near which it is de-ployed. If the defense concentrates its Overlayresources on a subset of the silos and aban-dons the others to the attacker, then it leavesthe endo defenses associated with the aban-doned silos open to easy saturation and pene-tration. These endo resources—all bought andpaid for— are wasted, whereas the whole pur-pose of adaptive preferential defense was tomake optimum use of defensive resources.

The Importance of overlay Ieakage

In contrast to low-altitude systems, whichcan accept relatively high leakage and still doa single-shot job, high performance is requiredof the Overlay component of a Layered De-fense. Thus one must take seriously the manysources of leakage which can be present in thecomplex process of exo interception and alsothe possibility of having to face attacks involv-ing decoys and other penetration aids, In prac-tice, poor Overlay performance drives the de-fense to large inventories of interceptors inorder to maintain a given level of silo survival.The effect of this sensitivity to Overlay Ieak-age is best illustrated in the context of specificcalculations, Such calculations are presentedin the next section,

Ch. 3—Ballistic Missile Defense . 131

The Overlay and Layered Defenseof Silo-Based MX

Overlay Description

The Army’s concept of the exo componentof a Layered Defense — called simply the Over-lay– is in the technology exploration stage. Nodetailed system design is available as exists forLoADS.

In outline (see fig, 62), the concept consistsof interceptors roughly the size of offensivemissiles equipped with infrared sensors andcarrying several kil l vehicles (KVS), alsoequipped with infrared sensors. The multipleKVS would be mounted on the upper stage ofthe interceptor. The interceptors, of whichthere might be several hundred, would bebased in silos in the Central United States inthe same manner as offensive missiIes.

When attacking Soviet RVS were about two-thirds of the way through their flight to U.S.targets — about 10 minutes before impact — theinterceptors would launch into space, Whenan interceptor reached space, its infrared sen-sor would scan its field of view and attempt todiscriminate approaching RVS from tank andbus fragments and from decoys or other pene-tration aids. The infrared sensors would detectthese objects as warm spots —warm since theywere launched from the Earth — against thecold background of space.

Each KV would be assigned a target deter-mined to be a true RV and dispatched to in-tercept it. Using its own rocket power and in-frared sensor, the KV would home in on the ob-ject and destroy it either by colliding with itdirectly or by deploying a barrier of material inits path, Since the closing velocity of RV and

Figure 62.—Overlay/Underlay LayeredDefense System

Warning sensor

Reentry vehicles,

Underlay/

/

Intticeptor/ ~“ “ ; ” A P

interceptor z z . “ 9#/- / 1 1 f -i 3 /

\ / - c

Interceptor 1silofieids 8attie

manager

SOURCE Off Ice of Technology Assessment


KV would be about 25,000 miles per hour, asmall fragment of material from the KV couldcompletely destroy the RV. Even a glancingblow could damage the RV’S heat shield,mean ing that it would either burn up or be car-ried off-course as it reentered the atmosphere.

System studies of the Overlay concept haveshown that performance would be improveddramatically by providing an additional sensorwhich would make an early assessment of thesize and nature of the attack, allowing in-terceptors to be assigned more efficiently toregions of space. One idea for such a forwardacquisition system (FAS) would be rocket-Iaunched infrared sensing probes lofted intospace as soon as warning sensors indicatedSoviet launches. The probes would arrive onstation within a few minutes and remain therefor a short time before falling back to Earth.During this time, they would relay informationon the trajectories of attacking RVS back tothe interceptor silo fields. Since at any onetime in the attack a number of probes wouldbe required to cover all the attack corridorsand their time on station would be limited, alonger lived FAS system might be required aswell. This might consist of infrared sensorsmounted on high-flying aircraft maintained oncontinuous alert and capable of several hoursof time on station. Alternatively, satellitescould perform an FAS function. However,neither of these longer liived FAS systemswould be as capable as the probe.

The data acquired by the FAS would be inte-grated and interceptors assigned by a CentralBattle Manager or by Wing Battle Managersassociated with each set of defended silos. Thebattle managers would decide on a defensestrategy and make interceptor assignments ac-cordingly.

The battle managers and their data linkswouId have to be immune to disruption by pre-cursor SLBM attack.

Last, an exo defensive system would requireearly, secure, and reliable warning of Soviet at-tack. Systems to provide this warning must beconsidered part of the Overlay architecture.

Risks to Overlay Effectiveness

The interceptors and kill vehicles, intercep-tor silos, FAS (probes, aircraft, or satellites),battIe managers, and communications systemsdescribed above wouId comprise an extremelycomplex defensive system. The system ar-chitecture remains to be worked out in detail,a n d m a n y technoIog y issues are yet u n re-solved, 1 n the absence of a detailed systemdesign, it is not possible to analyze in quan-t i tat ive detai l the effect iveness and vul-nerabilities of a Layered Defense system basedon the Overlay in the way that such analysis ispossible for LoADS. Analysis of the Overlay inthe context of MX basing must instead rely ona qualitative estimation of technical risk andthe sensitivity of Overlay performance to fac-tors which are yet unknown.

As in the discussion of LoADS effectiveness,this section begins by asking how well theOverlay must perform in order to guaranteeacceptable protection for silo-based MX mis-siles. Unlike a LoADS deployment with a sin-gle-shot goal, the effectiveness of a LayeredDefense based on the Overlay is very sensitiveto the details of system performance. Onemust therefore take seriously the uncertaintiesin overlay performance which exist at present.These uncertainties concern both the fun-damental technologies in the Overlay conceptand potential vulnerabilities in the workingsystem as a whole. For the moment, facing therelative immaturity of the Overlay conceptand a near-term decision regarding MX basing,it would be quite risky to rely on Layered De-fense as the basis for ensuring MX surviv-able it y.

THE IMPORTANCE OF LEAKAGE

“The ability of a Layered Defense to protectsilo-based MX missiIes against a massive Sovietattack depends sensitively on the performanceof the Overlay component. For LoADS/MPS,by contrast, the defense would achieve a sin-gle-shot goal even if interception failed almostas often as it succeeded.

The effectiveness of a Layered Defense is amatter of probabilities, and the confidence of


attacker and defender to acheive their goalscould depend not only upon the odds them-selves but upon how willing either side wouldbe to “play the odds” in a nuclear war, Thediscussion which follows is intended t. beiIIustrative only: the precise numbers com-puted depend upon the assumptions and amyriad of details, but the overall trends, in-d ica t ing the sensitivity to Overlay perform-ance, do not. The assumptions made here tendto be rather favorable to the defense.

An illustrative silo basing arrangement forMX might consist of 200 MX missiles deployedin Minuteman I I I silos, The total U.S. ICBMforce would then consist of 450 Minuteman I Is(one RV each), 350 Minuteman I I Is (three RVSeach), and 200 MX (10 RVS each, say), for atotal of 3,500 RVS in 1,000 silos. In what fol-lows it is convenient to take each silo to be atarget of equal value, as though all missileswere identical and carried between three andfour RVS. In actual fact, of course, the Sovietswould be Iikely to target, and the United Statesto defend, MXS more heavily than MinutemanIIIs and Minuteman 11 Is more heavily thanMinuteman I IS

Assume further that the Soviets deploy n o

penetration aids or alternatively that discrim-nation is pertect. This assumpt i on concedesquite a bit to the defense, as the section onDecoys and Other Penetration Aids shows.

There is a certain probability that the Over-lay will succeed in destroying an RV if it de-tects, tracks, and assigns a KV to it. Call thisprobability the “efficiency” of the Overlay;the leakage is one minus the efficiency. (It willbe necessary to assume that the probabilitiesthat individual intercepts succeed are statis-tically independent; this would not be true if,e.g , the KVS concerned originated on the sameinterceptor, ) A quite respectable value for theefficiency in the absence of Soviet penetrationaids would be 0.85 (85 percent); this valuewould assume achievement of al I of the Over-lay “specifications “ A more modest valuewould be 70 percent, and so percent would bedisappointing indeed. The point of this anal-ysis is to show how strongly the number of U.S.

RVS surviving Soviet attack depends on Over-lay efficiency.

The Underlay must also be specified. Heremany choices are possible, ranging from ahigh-altitude Site Defense to simple single-shot“terminal” defenses. Assume that associatedwith each silo is a low-a It itude defense with thecapability to make a single intercept 70 per-cent of the time, a second intercept 50 percentof the time, and no capabiIity to make three ormore intercepts above the same silo. This con-stitutes a rather large and costly deploymentand assumes effectiveness typical of low-altitude systems. A second endo intercept isallowed on the assumption that the secondleaker could follow the first with a time delay.

For the Soviet arsenal, targeted against bothMX and Minuteman, a range from 5,000 to12,000 (reliable arriving) RVS could be con-sidered; calculations wilI be done for a repre-sentative value of 8,000, Each arriving RV isassumed to have a single-shot kilI probabilityof one.

If the Soviets had 8,000 RVS and no reasonto target particular silos preferentially, theycould direct 8 RVS at each silo, timed to arrive(if they penetrate the defense) within a shorttime of one another. If the Overlay has an effi-ciency of 85 percent, then the probability thatall eight RVS aimed at a defended silo aredestroyed by the Overlay is 0.85 to the eighthpower or 0.27 (27 percent). The probability thatone RV penetrates is the probability that sevenRVS are intercepted (0.85 to the seventh power)times the probability that one penetrates (0,1 5)times the number of RVS (8) which have achance to penetrate, for a total probability of0.38 (38 percent). (The apparently paradoxicalresuIt that one RV penetrates more often thannone refIects the fact that zero penetrationcan only occur one way, whereas single pene-tration can occur eight different ways, ) Theprobability that two RVS penetrate is 24 per-cent.

Thus, 27 percent of the time the silo is safebecause all RVS are destroyed in space. Onegets through 38 percent of the time, but the


endo Underlay destroys this RV 70 percent ofthe time. Two RVS get through 24 percent ofthe time, but the first of these is destroyed T O

percent of the time and the second 50 percentof the time. Thus the overall probability thatthe silo survives is (0.27) + (0.7) (0.38) + (0.24)(0.7)(0.5) = 0.62, or 62 percent.

How many silos would actually be defendedat all depends on the number of interceptorsthe United States deployed. For instance, if theUnited States deployed 400 interceptors with10 KVS each, 500 silos could be defendedagainst the 8,000-RV Soviet attack. Since 62percent of the defended silos would survive, atotal of 310 silos or 1,085 RVS would survivethe attack. The Soviets would have expended8,000 RVS to claim 2,415 U.S. RVS, and over4,000 Soviet RVS would have arrived on theUnited States.

Suppose now that the Overlay efficiencywere not 85 percent, but only 65 percent. Theprobabilities of none, one, or two RVS pen-etrating to the Underlay are then 3, 14, and 26percent, respectively if the defense persists indirecting one KV at each RV. The result thattwo RVS can penetrate more often than onereflects the fact that there are more pairs o fRVS (28) than individual RVS (8). The higherprobabilities for multiple leakers means thatthe endo defense has a harder job. In this case,only 22 percent of defended U.S. silos survive.If 4,000 KVS were used to defend 500 silos, thetotal U.S. survivors would be 110 silos or 385RVS.

But this would not in fact be the best U.S.defense strategy if the Overlay efficiency werelow. A better result would be obtained by de-fending half as many silos with twice as manyKVS per silo, i.e., defending 250 silos anddirecting two KVS against each Soviet RV. Thisstrategy would result in 70 percent survival ofthe 250 defended silos, for a total of 175 silosor 613 RVS surviving. The Soviets would againhave paid 8,000 RVS, and over 6,000 RVS wouldhave arrived on the United States. The Under-lay defense at the 750 undefended silos wouldhave been saturated.

In this example, a 20 percent degradation inOverlay efficiency (from 85 percent to 65 per-cent) results in the number of U.S. survivorsbeing reduced by almost one-half. This effectdemonstrates that the effectiveness of LayeredDefense to protect silo-based missiles dependssensitively on the Overlay efficiency.

Figures 63 and 64 further demonstrate theimportance of Overlay efficiency. Figure 63shows that the number of U.S. survivors de-creases dramaticaIIy as the Overlay efficiencydegrades. Figure 64 shows that the size of thedefensive arsenal needed to assure survival of1,000 RVS (286 silos) quickly gets out of hand ifthe Overlay efficiency begins to slip. For afixed number of U.S. silos the sensitivity toOverlay performance is more pronounced thelarger the Soviet threat. For a fixed threat, thesensitivity is less pronounced the larger thenumber of U.S. aim points.

TECHNOLOGY ISSUES

Because of the sensitivity to performancedescribed above, one must take seriously theuncertainties which exist at this stage of the

Figure 63.—Sensitivity of Layered DefensePerformance to Overlay Leakage

(Soviet attack consists of 8,000 reentry vehicles)(U.S. deploys 400 overlay interceptors)

900-800

7 0 0 “

6 0 0 –

5 0 0 -

4 0 0 —

3 0 0 -

200 –

100 —

o I 1 I I 1O 10 20 30 40 50 60 70 80 90 100

Overlay efficiency

SOURCE: Off Ice of Technology Assessment

Ch. 3—Ballistlc Missile Defense ● 135

Figure 64.— U.S. Defensive Arsenal Needed toAssure 1,000 Surviving U.S. Reentry Vehicles

(Soviet attack consists of 8,000 reentry vehicles)

10,000“

9,000 - 900 —u8 , 0 0 0 - 8 0 0 —

7,000 - 700 —

6,000 - 6 0 0 —al

5,000 “ 5 0 0 —

4,000 - 4 0 0 —

3,000 “ 3 0 0 —o 2,000 - 2 0 0 —

1,000 “ = 100 —

o - 0O 10 20 30 40 50 60 70 80 90 100

Overlay efficiency


Overlay’s development. Many aspects of theOverlay interception function require the fron-tier technologies of infrared sensor design,compact computer design, rapid computerthroughput, software architecture, small hom-ing and KV technology, and so on. The systemrequirements are demanding, and at al I stagesof the interception process advances must bemade to meet them. Furthermore the stagesare interlinked in such a way that failure at onestage could cause failures at others and de-grade overalI system performance signif-icantly, There is no reason to believe that,given time and money, the uncertainties inOverlay technology could not be reduced to apoint where a clearer estimation of the valueof an Overlay system for the protection of silo-based missiles could be made Nor is there anyparticular reason —with one exception, dis-cussed in the section Decoys and Othe rPenetration Aids – to believe that fundamentaltechnical problems will be encountered whichby themselves would constitute “Achilles’heels” for an Overlay defense. Indeed, there issome reason for optimism, since the neededtechnology elements fall into categories—compact data processing, infrared sensing,miniature guidance systems, homing, etc. — inwhich rapid progress is occurring now and

more is expected in the future. Rather, the risklies in the cumulative effects of shortcomingsin the performance of technology elements atmany stages of the interception process. Inmany cases the capabilities of today’s tech-nology falls short of the requirements of theOverlay by wide margins. Though progress canand should be expected, even small shortcom-ings couId uItimately prove significant, sincepoor performance in one area can inducefailure elsewhere, and the cumulative effectcould be to reduce total system performancebelow the high standard required for MX de-fense. Technology forecasting is always risky,and in the case of the Overlay one is simplyfaced today wi th an unknown quant i ty.Though the level of confidence in Overlaytechnology may in time increase to the pointwhere it could support a decision regardingMX basing, at the moment it appears quiterisky to depend on successful resolution of alloutstanding issues. The following discussionseeks to highlight unresolved technologyissues and convey a sense of the complexity ofexo defense.

The interception process begins when theprobes or other FAS vehicles survey the attack-ing “threat complexes, ” the clusters of RVS,bus and tank fragments, and possibly decoysor other penetration aids, which are boostedinto space by Soviet offensive missi les. I n alarge Sov iet a t tack there would be tens o fthousands o f such objects in the sky. Eachprobe sensor should have as wide a f ield ofv iew as poss ib le so that a smal l number o fthem can survey the whole sky. But a widefield of view means either a large sensor, whichis hard to protect from interfering backgroundfrom the Sun, Earth, atmosphere, and other ef-fects; or a slow scanning rate, which makes itdifficuIt for the sensor to correlate what it seeson one scan with what it saw on previousscans, essential for discriminating RVS fromother objects and for compiling accurate tra-jectory information. Unlike radars, infraredsensors cannot determine the ranges of distantobjects but only their angular positions in thesky. Range must be inferred from changes inthe angular positions of objects with time.Angular position must therefore be measured


very accurately, requiring precise orientationof the sensors in space. The threat complexesmight also be rather dense: a single spot in thesky could later resolve itself into several ob-jects. The probes must resolve all objects andobserve them long enough to attempt to dis-criminate RVS from nonlethal objects andcompute accurate trajectories. If all this isdone poorly or too late, too many interceptorswill be allocated to some regions of space andnot enough to others.

The probes must then “handover” their filesof objects to the interceptors, telling the in-terceptor sensors where to find each objectand what it looks like. The interceptors mustreacquire each object in their respective sec-tors, attempt discrimination again, and deter-mine how best to release their KVS. Again,poor performance at this stage degrades per-formance at the next,

Last, the interceptor sensors must hand overtargets to individual KVS. The KVS must thenmaneuver in such a way that they come withinlethal range of an RV approaching at 25,000mph The KVS must also be able to distinguishthe true RV from objects placed nearby. If thefirst intercept failed, a second wave of in-terceptors wouId have a very short time to per-form all the required functions. It is not clearin any event that it would be possible to tellwhich intercepts had failed on the first at-tempt.

The infrared sensors are the most delicateelement of the Overlay system. Infrared sen-sors can be disrupted by heat, nuclear radia-tion, and rocket exhaust gases. They wouldhave to be mounted on very sensitive gimbalsystems with accurate inertial guidance, ln-frared sensors measure the temperature char-acteristics of approaching objects. These char-acteristics depend sensitively upon the posi-tion of the object relative to the Sun and Earthand on the time of day, season, and weatherconditions on the Earth below. All of this datawould have to be made available to the sen-sors before they could interpret what they saw.Infrared sensors of great sensitivity are alsorather temperamental in that each behaves dif -

ferently and must be calibrated separately,and a given i I I urn i nation of the same sensorcan sometimes result in different output volt-ages. These latter factors introduce some fun-damental limitations in temperature resolu-tion, an important factor in discrimination.

SYSTEM ISSUES

In addition to the risk introduced by the hightechnology required, the Overlay as a systemcould have vulnerabilities much Iike thevulnerabilities of other basing modes. For in-stance, the Overlay would depend on tacticalwarning, since the system must begin to func-tion early in the flight of Soviet ICBMS. TheOverlay could thus share some of the potentialvulnerabilities of other basing systems whichdepend on warning such as launch underattack and air mobiIe MX. The battle manage-ment function requires survival of the com-mand centers and of secure, high-data-ratecommunications among the FAS, battle man-agers, and interceptor silo fields. The concernshere are similar to those regarding wartime sur-vival of our national military C systems. Also,as with LoADS, attention must be given to of-fensive tactics designed to confuse or bypassthe defense.

Limitations of the Overlay defense againstsubmarine-launched balIistic missiles (SLBMS),which couId attack the Overlay’s own com-ponents, are discussed in the Classified Annex.Leakers from a precursor ICBM attack mightalso threaten critical system elements.

In the absence of a specific system design, itis again not possible to estimate vuInerabilitiesprecisely. The interceptor silos and especiallythe probe silos would be few in number. Pres-ent Soviet SLBMS have limited silo-kill ingcapability, and they are not normally deployedin large numbers near U.S. shores. Still, theOverlay defense assets would be high-valuetargets for an SLBM precursor attack, and de-pending on their hardness they could be vul-nerable. Some thought has been given to pro-viding an endo defense for the Overlay mis-siles themselves. Softer targets such aspossible ground-based battle management


bunkers and their communications links couldbe vulnerable to less accurate SLBMS.

If strip-alert aircraft were used to supple-ment the probes or to serve as battle manage-ment centers, they would have the same vul-nerabilities as the bomber force and AirMobile MX. One must also consider an antisat-elIite threat if satellites were used to aid anFAS,

Care must also be taken to provide for sur-vival of the high-data-rate communicationslinking FAS, battle managers, and interceptorfields, The generic technical problems of pro-viding survivabiIity for communications aresimilar to the case of launch under attack andto the C systems of other basing modes.

A complex defensive system like the Over-lay must also reckon with offensive counter-measures. The most important of these is theuse of penetration aids, discussed in the nextsection. Other tactics are discussed i n the Clas-sified Annex,

Detailed study of vulnerabilities at the totalsystem level must wait until the Overlay con-cept is translated into a working design. Ex-perience with the national military C’ systemand studies of launch under attack, air mobileMX, LoADS/MPS, and other basing systemsgive an idea of the scope of problems whichcan be encountered when a complex MXbasing system faces a future Soviet threat. Forthe moment, the uncertainties in whether arobust wartime system can be fashioned fromthe Overlay concept are another source of riskto a decision to make Layered Defense thebasis for MX survivability.

Decoys and Other Penetration Aids

The Overlay concept is based on the prac-ticality of infrared sensing in space. However,infrared sensing is potentially critically vul-nerable to offensive countermeasures in theform of decoys and other penetration aids.Unlike the LoADS radar operating at low alti-tudes, which could measure the weight of ap-proaching objects, the Overlay infrared sen-sors would measure their temperature charac-teristics. Decoys able to fool the LoADS radar

83-477 0- - 8 1 - 1 0

must be heavy, and adding heavy decoys to anoffensive missile requires removing RVS, sincethe missile has Iimited throwweight. Measuringweight is therefore a strategically significantmethod of discriminating true RVS from de-coys, since the offense would presumably notchoose to offload RVS and replace them withequalIy heavy decoys. On the contrary, there isno impediment in principle to deploying ligh-weight decoys which have temperature char-acteristics indistinguishable from those of trueRVS. Temperature, which is fundamental tothe Overlay sensing method, is not a stra-tegically significant discriminant. The offensemight therefore be able to deploy a Iarge num-ber of excellent lightweight decoys on itsoffensive missiles without having to offloadmany RVS. This would call into question theeffectiveness of an exoatmospheric defense ofMX.

This section will indicate some of the ele-mentary physical principles that permit the de-sign of Iightweight penetration aids. It wilI alsoindicate the practical difficulties which the of-fense would face in mounting decoyed attacksas welI as those the defense wouId face incountering them. A more complete discussionis relegated to the Classified Annex.

To get a feeling for the importance of dis-crimination, consider the case if the offenseprovided along with each RV a sing/e perfectdecoy. A KV approaching the two objectswould then intercept the true RV 50 percent ofthe time. If the efficiency of the defense, asdefined in the last section, were 85 percent inthe absence of decoys, and if one KV were dis-patched against each RV/decoy pair, then thetrue RV would be intercepted only (0.85) (0.5)= 43 percent of the time. This low efficiency

(high leakage) would be catastrophic to thedefense, as figures 63 and 64 in the previoussection show. If on the other hand the defensedirected a KV at both the RV and the decoy,then the same number of U.S. silos would sur-vive as in the no-decoy case, but an arsenal ofdefensive missiles twice as numerous would berequired to produce this resuIt

In practice, no decoy is perfect, and on theother hand the offense could deploy many


more than one decoy with each RV. In practicealso, a tradeoff is made between leakage (in-tercepting the object judged most likely to bean RV and allowing an RV to penetrate if theguess is wrong) and wastage (interceptingeverything). In the case of the Overlay, theneed to keep leakage low means that the bestsolution for the defense is usually to accepthigh wastage. Thus, an offensive decoy deploy-ment would tend to drive the defense to largernumbers of defensive interceptors — in the ex-ample above, applied to the model in the lastsection, twice as many, or almost as many de-fensive missiles as offensive missiles.

INFRARED SENSING AND TEMPERATURE

A metal bar heated to very high tempera-tures glows white-hot. If its temperature islowered somewhat, it glows red-hot. if cooledfurther—to room temperature— it no longerglows in the visible part of the spectrum but atlonger wavelengths, in the infrared part of thespectrum. A room-temperature object thus“glows infrared” and can be “seen” by a detec-tor sensitive to infrared light.

An RV launched into space from the approx-imately room temperature condition of its siloforms a glowing spot against the dark (i.e.,cold) background of space. Infrared sensorscan measure both the color (i. e., the preciseshade of infrared light) and the brightness ofthe RV and of any other objects, such as de-coys, which accompany it. The color of the ob-ject reveals its temperature and its brightnessreveals its size and the type of material it ismade of. Decoy/RV combinations that appearto the infrared sensors to have identical colorand brightness cannot be discriminated.

In addition to the warmth it brings with itfrom the Earth, the object absorbs energy fromthe Sun above and the Earth below and losesenergy to cold space. Its color could thereforechange as it were warmed or cooled. In addi-tion to emitting light because of its tem-perature, the object also reflects infraredradiation from the Earth. An infrared sensortherefore senses the combination or sum of theemitted and refIected energy from the object.

T here is a relationship between the tendencyof a body to emit thermal radiation because ofits temperature and its tendency to reflectradiation which shines on it. For a body of agiven size (more precisely, surface area) at agiven temperature, the sum of its effectivenessin emitting radiation of a given wavelengthand in reflecting it is the same no matter whatthe body is made of. Therefore, the less in-frared radiation a body in space emits, themore it reflects from the warm Earth and viceversa. The relative emittance and refIectancedoes depend on the nature of the body, butonly on what the surface of the body is madeof and not what is inside it.

Using only these elementary principles ofphysics, it is a straightforward matter to designRV/ light-decoy pairs which appear identical toinfrared sensors. A wide variety of other ex-amples of penetration aids based on simplethermal properties of materials can also bedesigned. These are discussed at some lengthin the Classified Annex.

MEASURE AND COUNTERMEASURE

In practice, the situation is more complexthan simple physical principles alone would in-dicate, with many constraints and opportu-nities both for the offense and the defense.Despite these complexities, the fact remainsthat there is no principle and no detector thatcouId guarantee perfect Overlay discrimina-tion. The burden would thus rest with the de-fense to maintain i ts conf idence that i tsmethods of discrimination were adequate tomeet a decoy threat.

l-o begin with, there would be practical con-straints on the offense. Foremost among theseis the fact that the best results would be ob-tained by altering the RV to make it easier fora Iight decoy to match. Though these changesare minor, inexpensive, and need not affect RVperformance in the least, there could be somepsychological reluctance to tamper with thelethal RV for the sake of a nonlethal decoy. Itis also one thing to design the perfect decoyancl quite another to package it, mount it onan ICBM, and deploy it in space so that its

Ch, 3—Balllstic Missile Defense ● 139

deployment process and in-flight motions (as-suming these could be observed by the defensein an attack) resemble those of a true RV.

Constraints on the defense include the factthat infrared sensors are not perfect (i. e.,cannot determine brightness and temperatureprecisely, especially in the presence of back-ground) and they would not have a very longtime to observe objects in an engagement. Thetemperature characteristics of objects in spacefurthermore depend on the position of the Sunand Earth relative to the sensor and the time ofday and weather conditions on the Earth be-low. Overall, the interception process is dif-ficult enough even in the absence of decoys, asdiscussed in the last section.

Since there is no fundamental principle onwhich infrared sensors can rely to guaranteediscrimination, there could be value in someadvance knowledge of the type of penetrationaids the offense deployed. It is possible (likelyis too strong a word) that by observing flighttests, the defense could Iearn enough aboutthe offense’s penetration aids to devise a dis-

crimination scheme based on some distinctivefeature or detail of the offense’s design ordeployment procedure. However, since thereappears to be a wide variety of effective pene-tration aids which the offense could use, thisapproach based upon particulars rather thanprinciples could succeed for one penetrationaid but faiI for another.

Details and further discussion of penetra-tion aids, constraints, and tactics can be foundin the Classified Annex.

Overlay discrimination is a difficult prob-lem, the practical details of which are not un-derstood, though the principles are. Testing ofpenetration aids designed expressly and ex-clusively for the purpose of Overlay penetra-tion is required before it can be known wheth-er the perfect decoy of principle can be real-ized in practice and whether less-than-perfectdecoys still make defense based on infraredsensing too difficult and costly to undertake.For the moment, the very fact that effectivedecoys are possible counsels caution.

HISTORY OF BMD AND THE ABM TREATY

The development of ABM systems by theUnited States in the early 1950’s followed thedecision to begin development of ICBMS.During the mid-1 960’s, the johnson administra-tion proposed the deployment of the so-calledSentinel ABM system to provide both area andpoint defense against a limited nuclear attack.This proposed ABM system was reviewed in1969 by the Nixon administration which optedinstead for an ABM system to defend Minute-man siIos. Deployment of the Nixon adminis-tration’s Safeguard ABM system was begun inthe early 1970’s but was brought to a haltfollowing negotiation and ratification of theABM Limitation Treaty of 1972. The Treatywas subsequently amended by the Protocol of1974.

With the development of ballistic missiledefense, doubts about the long-term viabilityof international security based on a “balance

of terror” began to mount WhiIe alternativesto maintaining a balance of terror were ex-plored through a variety of formal and in-formal channels, by the mid-1960’s it seemedto many senior U.S. policy makers that somesort of arms Iimitation on balIistic missiledefenses would be preferable to either an ABMarms race or a major revision in the post-WorldWar I I “balance of terror”.

For example, Defense Secretary Robert S.McNamara noted:

Should they elect to do so, we have both theIeadtime and technology available to so in-crease both the quality and quantity of our of-fensive strategic forces –with particular atten-tion to highly reliable penetration aids—thattheir expensive defensive efforts wilI givethem no edge in the nuclear balance whatever.

But we would prefer not to have to do that.For it is a profit less waste of resources, pro-


vialed we and the Soviet can come to areal istic strategic arms-limitation agreement.

Even though Secretary McNamara had seri-ous reservations about ABM systems in gen-eral, he nevertheless proposed to deploy anABM system to defend the United Statesagainst some nuclear attacks. The SentinelABM system proposed by the johnson admin-istration included a long-range, high-altitudeexoatmospheric interceptor missile, the Spar-tan, guided to targets by a very large radar anda smaller, shorter range interceptor missile,called Sprint, also guided to its targets byradar. Both missiles were armed with nuclearwarheads which destroyed incoming reentryvehicles. The original johnson administrationproposal envisioned deployment of the Sen-tinel ABM System at some 14 sites includingICBM silo fields in Montana and North Dakotaas well as several major cities.

The Nixon administration reviewed the pro-posed Sentinel ABM system in light of bothU.S. strategic requirements and the intensepolitical opposition that arose over the poten-tial deployment of nuclear weapons adjacentto American cities and concluded that the useof Sentinel radar and interceptor componentsto defend U.S. Minuteman fields would be anappropriate and strategicalIy significant re-sponse to the Soviet deployment of an ABMsystem around Moscow. On March 14, 1969,President Nixon announced his plan to deployan ABM system to defend ICBM silos in Mon-tana and North Dakota:

This measured deployment is designed tofulfilI three objectives:

1. Protection of our land-based retaliatoryforces against a direct attack by theSoviet Union

2. Defense of the American people againstthe kind of nuclear attack which Com-munist China is likely to be able to mountwithin the decade,

Ch. 3—Balistic Missile Defense ● 141

In general, the ABM Limitation Treaty of1972, as amended by a Protocol negotiated in1974, prohibits ABM systems based on thetechnology deployed by the United States andthe Soviet Union in the late 1960’s and early1970’s, The Treaty as amended does permiteach side to deploy one ABM system for de-fense of either its national capital or an ICBMsilo field. The Treaty also permits continuedresearch and development on allowed ABMsystems, bans other types of ABM develop-ment, test, and deployment, and provides forfurther negotiations on specific limitations ofnew ABM systems based on technologies notdeployed in the 1970’s

Article I of the Treaty) prohibits ABM sys-tem deployments other than those specificallypermitted by subsequent articles of the Treaty.Article I I defines ABM system components. Ar-ticle I I 1, paragraph (b), permits the UnitedStates to deploy one ABM system with the fol-lowing characteristics:

● not more than two Iarge phased-array ABMradars,

● not more than 18 smalI phased-array ABMradars,

● not more than 100 ABM interceptor Iaunch-ers and not more than 100 ABM interceptormisslIes i n a deployment area having aradius of less than 150 kilometers centeredon the middle of an ICBM silo field.

Article IV of the ABM Limitation Treaty per-m its development and testing of ABM com-ponents at designated test ranges withoutcounting such components in the quantitativelimits established in article I I 1,

Article V of the Treaty bans the develop-ment, test, or deployment of sea-based, air-based, space-based, or land-mobile ABM sys-tems or components. Article V also bans thedevelopment, test, or deployment of ABMinterceptor launchers which contain more thanone interceptor or which are capable of auto-matic or semiautomatic interceptor reload.

Other official statements incorporated intothe legal restrictions of the ABM LimitationTreaty also affect future ABM system develop-ment, test, and deployment. Agreed statement(D) contains the following provision:

In order to insure fulf i l lment of the obliga-tion not to deploy ABM systems and theircomponents except as provided in Article I I Iof the Treaty, the Parties agree that in theevent ABM systems based on other physicalprinciples and including components capableof substituting for ABM interceptor missiles,ABM launchers, or ABM radars are created inthe future, specific limitations on suchsystems and their components would be sub-ject to discussion in accordance with ArticleXl I I and agreement in accordance with ArticleXIV of the Treaty. ’

Agreed statement (E) prohibits deployment ofABM interceptor missiles with more than oneindependently guided warhead.

The ABM Limitation Treaty Protocol nego-tiated in 1974 and ratified in 1976 furtheramended the ABM Limitation Treaty in the fol-lowing respects. Article III of the Treatyoriginal Iy permitted both the United Statesand the Soviet Union to deploy two ABM sys-tems in two deployment areas. One permittedsystem could defend an ICBM silo field;another could defend the national capital. Ar-ticle I of the 1974 Protocol limits each side toonly one ABM system deployment.

Article I I of the Protocol permits each sideto shift deployment of its permitted ABM sys-tem once. In the case of the United States, theprotocol would permit the dismantling anddestruction of the ICBM silo ABM defense sys-tem at Grand Forks and the relocation of theABM system to the Washington, D.C. area.

Application of ABM Treaty Provisionsto MX Defense

ABM systems deployed to defend MX arelimited by the ABM l-imitation Treaty of 1972(as amended by the 1974 Protocol) in two dis-tinct ways. First there are limitations on thedeployment of ABM systems. Second, there

‘I bid, p 143


are limitations on the development of newABM systems.

ABM deployments are limited by the ABMTreaty as amended in the following respects:

1

2

3.

4

5.

6,

7.

any U.S. ABM system may only be de-ployed in a c irc le of 150-km radiuscentered on the Grand Forks, N. Dak.,ICBM field;LoADS defense units could not be de-ployed if any system component were ex-plicitly not of a fixed type;since each LoADS unit would contain onesmall radar, no more than 18 LoADS DUScould be deployed under terms of theradar limitations of the ABM Treaty;the total number of LoADS or OverlayABM interceptor launchers and ABMinterceptor missiIes couId not exceed 100;LoADS defense units could not be de-ployed if each unit contained more thanone ABM interceptor launcher. Alterna-tively, even if each LoADS defense unitcontained only one interceptor launcher,it would stilI possess in principle auto-matic, semiautomatic, or rapid reloadcapability, barred by the Treaty;since each Overlay KV is an independ-ently guided warhead within the meaningof the Treaty, deployment of more thanone such warhead on each Overlay in-terceptor missile would be prohibitedunder provisions of agreed statement (E).deployment of space-based laser ABMsystems is explicitly prohibited by articleV of the ABM Treaty,

The development of future ABM systems isalso limited under terms of the ABM Limita-tion Treaty. The United States and the SovietUnion have defined development of ABMs forpurposes of the Treaty as follows:

The obligation not to develop such systems,devices or warheads only to that stage of de-velopment which follows laboratory develop-ment and testing, The prohibitions on develop-ment contained in the ABM Treaty would startat that part of the development process where

field testing is initiated on either a prototypeor bread-board model. 5

Thus, the following limitations would apply tothe development of specific ABM systemssuch as LoADS, Overlay, or even space-basedABM systems:

1.

2.

3.

mobi le components of ABM systemsdeveloped beyond the laboratory such asLoADS defense units would be banned;multiple independently guided KVS forthe Overlay could not be tested beyondthe confines a laboratory;development of unique components forspaced-based laser ABM systems wouldbe banned.

Future ABM Limitation Negotiations

The ABM Treaty provides that either sidemay propose amendments during semiannualmeetings or special meetings of the Standingconsultative Commission which was estab-lished to resolve questions of interpretation inthe Treaty as well as to supervise and resolvequestions of verification. The 1974 Protocolamending the Treaty arose out of just suchStanding Consultative Commission discus-sions. The Treaty, which is of unIimited dura-t ion, also provides for a formal review con-ference every 5 years at which time either sidemay propose changes or amendments. Thenext ABM Limitation Treaty Review Con-ference is scheduled for October 1982.

Present ABM options for the defense of MXdeployments are significantly constrained bythe Treaty from the standpoint of final en-gineering. Substantial research on new ABMsystems can be undertaken, and developmentand testing of ABM components whose pur-pose is to modernize the mothballed Safe-guard system can also be undertaken, Newradars, new interceptors, and new warheads

“ Ball Istlc MI$SIIP Defense, ” In U S Congress, House Commit-tee on Foreign Affalr$ and Senate Commi t tee on Foreign Rela-

t Ion 5, 4 rrrs C-on tro/ /mpa c t Statement J tor f I >ca / Year / 982( W a s h i n g t o n , D C U S ( g o v e r n m e n t Printing Office, 1981, p

195


for Safeguard are all testable and deployableunder terms of article VI I of the Treaty permit-ting modernization of existing ABM systems,Even development of directed energy weaponsfor possible use as ground-based ABM systemswould be permitted under terms of the Treaty,so long as deployment as modernization forthe Safeguard system was envisioned.

The United States might wish to explore thepossibility of further amending the ABM Lim-itations Treaty in a manner that would permitengineering development and possible de-ployment of the LoADS or Overlay ABM sys-tem as they are presently envisioned during thecourse of the 1982 ABM Limitation Treaty Re-view Conference. Reopening discussions of thesubstantive provisions of the ABM LimitationTreaty does, however, raise serious questionsin need of further analysis beyond the scope ofthe study.

The process of renegotiating the ABM Limi-tation Treaty is subject to uncertainty. TheSoviets, too, have an active ABM research anddevelopment program which is also con-strained by the ABM Limitation Treaty. Modifi-cations in the terms of the ABM LimitationTreaty which would permit the United Statesto proceed with development and testing nec-essary to advance the LoADS ABM technologyinto engineering and full-scale engineering de-velopment, or permit development of Overlaytechnology, would also permit comparabledevelopments in the Soviet ABM program.

Hence judgments of the technical, political,and military benefits to be gained by reopen-ing negotiations on ABM I imitations will haveto be made should some basing mode for theMX missile requiring ABM systems be con-templated.

Documents

BALLISTIC MISSILE DEFENSE - Princeton