Kleppner, Lamb, and Mosher reply: The APS study concluded that a boost-phase defense using airborne interceptors (ABIs) incorporating technology that would be available within the 10-year period considered could be useful in limited circumstances. Dean Wilkening’s analysis agrees with the study’s in most respects, but his assessment of the utility of ABIs is more positive.
One reason the assessments differ is that Wilkening assumes an airborne X-band radar system will be available. No such system yet exists. The study judged that the time required to field such a system would be greater than the 10-year period considered. A second reason is that his interceptor has a higher acceleration than the study’s comparable interceptor I-2 and has a 20-second burn time compared to the 40-s burn time of I-2. Because Wilkening’s interceptor has a higher average velocity, it can reach a greater distance, but there are penalties for such a high acceleration. Wilkening’s analysis appears to neglect one of these: The interceptor’s final stage (kill vehicle) must have more velocity change (“divert”) capability because it is released after only 20 s, when it has less information about the intercontinental ballistic missile’s flight and must maneuver for a longer time. Wilkening’s kill vehicle has a total divert capability of 2 km/s, which is what the study estimated would be required for a kill vehicle released after 40 s. Additional divert capability increases the mass of the kill vehicle and its boosters and reduces the final speed of the interceptor that a given aircraft can carry.
Wilkening says existing US and Russian solid-propellant ICBMs have nominal burn times of 180 s, whereas the computer models constructed by the study have 170-s burn times. But a 10-s increase in burn time would increase the I-2’s ground range by only 40 km, negligible compared to the other uncertainties. Although one must pick specific numbers for any analysis, no countries of concern currently have ICBMs, so the characteristics of missiles they might use are unknown.
Richard Garwin raises three principal issues. In his reference 1, second document, he says “there should be no tactical human decision” and therefore the decision time introduced in the APS report is not needed. Decision time, as defined in the report (page xxiii), 1 refers to the additional time the defense might require for the system to be effective—for example, to communicate between different elements or to estimate better the characteristics of the threatening rocket. No human decision is necessarily implied. The zero-decision-time timeline in the report is a bounding case: Interceptors could not be fired earlier, even if the many optimistic assumptions built into the analysis—including detailed advance knowledge of the attacking ICBM’s performance characteristics, a planar ICBM trajectory, no issues of battle management, communications, command, reliability, and so forth—were satisfied and the system worked perfectly. A practical defense must allow for these factors by basing interceptors closer to the intercept point or using faster ones. Such a margin can be characterized by a time. We chose to illustrate the effects of those margins by showing the effect of a 30-s margin. Building a system with no margin would be unwise.
Garwin argues that radar 400 km from the ICBM launch site could provide the missile warning and tracking information required to fire interceptors and could add tens of seconds to the time available for intercept. The study found that ground- or sea-based radars could provide this information, but would increase the intercept time provided by the advanced space-based detection and tracking system considered in the study only if radars are stationed less than 300 km from all potential ICBM launch sites. However, that would require stationing radars in enemy territory or within 100 km of an enemy coastline, contrary to the study’s guidelines. The study found that if the defense had to rely on the existing Defense Support Program (DSP) system, ground-based radars could improve midcourse guidance of interceptors.
Garwin calls attention to the usefulness of a 14-ton (12.7-tonne), 8.5-km/s, 100-s burn time interceptor for defending against a liquid-propellant ICBM launched from North Korea to the continental US. An 8-km/s interceptor was discussed in section 5.1.2 of the study. If it has a burn time of 40–45 s, its range can be estimated by interpolating between those of the study’s 6.5- and 10-km/s interceptors. Such an interceptor would be more manageable than a 10-km/s one. The study found that the average velocities of 6.5- to 10-km/s interceptors with burn times longer than 40–45 s were generally too low to be effective. Interpolating between the results for 6.5- and 10-km/s interceptors, we estimate that Garwin’s suggested 8.5-km/s interceptor with a burn time of 100 seconds would have an effective range several hundred kilometers less than with a 40-s burn time.
Michael Levi’s proposal to launch enough interceptors to cover all potential ICBM trajectories the moment a possible missile launch is detected raises serious technical, operational, and policy issues that, as far as we know, have not been analyzed. The most serious is the false alarm problem, which would be greatly exacerbated. Space-based IR missile warning and tracking systems must contend with an extremely cluttered environment. It takes time to observe a target long enough to establish that it is an ICBM and not sunglint from clouds and ice, a fighter aircraft using its afterburner, a fire, or a short-range missile. Radars also must wait for an ICBM to accelerate and pitch over to separate its return signal from background clutter and rightly identify it.
Firing a fusillade of interceptors as soon as a possible missile launch is “detected” would cause frequent false-alarm launches, waste dozens of interceptors, and make it possible for the attacker to exhaust the defensive system’s supply of interceptors by, for example, launching shorter-range missiles or other decoys before or while launching ICBMs. In addition to the very large number of interceptors required, the system would have to be able to eliminate the interference with its sensors and the possible fratricide caused by having such a large number of interceptors in flight.
Levi mentions cloud coverage. The study considered separately the available data on the altitude and optical depth distributions of cloud cover over land, sea, and coastal areas as a function of latitude (see section 10.1). To avoid ground clutter, the DSP missile detection and tracking system is designed to detect only IR sources above 10 km. The study also considered a modern, see-to-the-ground missile detection and tracking system. Although cloud tops can extend to 10 km, their average height over land is 4.7 km. An enemy has no obligation to attack only when the weather is clear. At low altitudes, ICBMs would not have sufficient ground speed for the defense to separate them reliably from the background. After considering these and other factors, the study concluded that clouds above 7 km are sufficiently rare that a modern system could reliably detect rockets by the time they reached that altitude, but not at significantly lower altitudes.
Contrary to Levi’s assertion, the 6.5-km/s interceptor would indeed “push the limits of what is possible.” The challenge is not in interceptor speed, but in meeting the total system requirements. Earlier in this piece, we mentioned some of the challenges. Others include solving the “plume-to-hardbody handover” problem, and development and testing of the sensors and the guidance and control algorithms needed for the kill vehicle to hit an unpredictably accelerating target. The 5-km/s interceptor Levi mentions could succeed only under perfect circumstances, and such a scenario is implausible.
Levi attributes to the three of us the quotation concerning the difficulty of defending all 50 states against solid-propellant ICBMs; however, that is the conclusion of the entire study group (page xxii), which also concluded that “when all factors are considered none of the boost-phase defense concepts studied is likely to be viable for the foreseeable future to defend the 50 states against even first-generation solid-propellant ICBMs” (page xxxvii).
The quotation Levi recites on using the 6.5-km/s interceptor to defend against North Korea is incomplete. In full, it is “The lower-left panel of Figure 5.10 reveals that even the 6.5-km/s interceptor could be used to defend Boston only if it were fired very close to the coast with zero decision time.” The restored words, shown here in italic, reverse the meaning. The study concluded that defense of the US against solid-propellant ICBMs, even with a 10-km/s interceptor, is “unlikely to be practical when all factors are considered” (page xxii) because such a defense would be at the limit of what is physically possible and the system would not have enough margin to be robust in the likely event of less than ideal circumstances.
Levi also refers to the study’s discussion of defending against solid-propellant ICBMs from Iran by basing 10-km/s interceptors near Iran’s coast in the Caspian Sea. The study called such basing locations “unconventional” and regarded them as implausible because of operational, security, and policy concerns. We disagree with Levi’s conclusion that an interceptor base in western Afghanistan would be useful.
In response to Truman Hunter’s questions: Analyzing the technical feasibility of using a single interceptor against a single ICBM is the starting point for analyzing the feasibility of a larger system. Analysis of multiple simultaneous threats was beyond the scope of the APS study. The basis for the presumed threats analyzed by the study is described in detail in chapter 3 of the APS report. His other issues lie outside the scope of the study.
