National Missile Defense: Test Failure and Technology Development

September 8, 2000—Mr. Chairman, members of the Subcommittee, I appreciate the opportunity to appear before you today to discuss the critical issue of National Missile Defense (NMD) technology development. I am honored to be in the company of those who have the challenging task of managing and overseeing this critical program. As you requested, I will focus my testimony only on the impact of test results to date and what they mean for the deployment of an NMD system. But as we all know, there are many other critical factors that will impact a deployment decision, and I would urge you and your colleagues to consider these factors at some future time.

I commend the Subcommittee for holding this hearing because it is difficult for an executive agency to evaluate objectively a program which it is developing. Indeed, that is why Congress in the early 1980s established a separate Director of Operational Test & Evaluation within the Department of Defense (DOD).

Before discussing the results of the tests, which have been conducted, it is important to be clear, first of all, of exactly what specific NMD system we are assessing. Too often, people engaged in the debate talk past one another because they are discussing different NMD systems, e.g., land or sea-based, or space-based, or some combination thereof, or different components of a particular system.

The proposed NMD system, which we are assessing today, would use ground-based interceptors (GBI) deployed initially at one site and eventually at two sites, supported by an extensive network of ground-based sensors and space-based infrared sensors. This system will have five (5) key components.

1. Space-based early warning satellites deployed in a geosynchronous orbit some 36,000 km above the earth. Their role is to detect an ICBM within a minute of launch and provide missile launch location and targeting information.
2. Ground and space-based sensors which are the key to the effectiveness of the NMD because they track the warheads, as well as discriminate the warhead from any decoys. These sensors include band ground-based phased array radars (a prototype is in place in Kwajelein and is being used in the current testing of the NMD system); five large phased array radars (located in Alaska, California, Massachusetts, Greenland and Britain) which will be upgraded to track incoming warheads accurately enough to guide interceptors; and, a space-based missile tracking system known as SBIRS-LOW which will not be ready before the scheduled 2005 initial deployment.
3. Command, Control and Communication Centers in Cheyenne Mountain in Colorado and at the interceptor basing sites. The Centers are supposed to determine when the interceptors will be launched and update the information to them during their flight.
4. The Ground-Based (GBI) Interceptor, a three-stage rocket, which must be capable of accelerating the payload or weapon to speeds of 16 to 18 thousand miles per second.
5. Exoatmospheric Kill Vehicle (EKV) is a 130-pound package of miniature rocket motors, computers and sensors which is to be released by the GBI and coast through space and search for its target and kill it by colliding with it at speeds of up to 34,000 miles per hour.

All five components, to various degrees, are pushing technological frontiers and to state the obvious, in order to have an effective NMD, all five components of the NMD system must work all the time. For example, the July 7, 2000 test failed because the GBI did not function properly.

The current plan calls for deploying by 2005, twenty interceptors in central Alaska, a new X band radar in Shemya in the western Aleutians, and an upgrade of the existing early warning radars in Alaska, California, Massachusetts and Britain. By 2007, the number of interceptors in central Alaska would grow to 100. The complete NMD system with some 200 missiles and the other components would be in place by 2010, if construction in Alaska begins in the spring of 2001.

If the proposed system is to enhance United States security and be worth the tremendous monetary and diplomatic cost, it must work and work well. There will be no second chances. Unlike the Safeguard ABM system, which we deployed in 1974, the proposed system is single shot as opposed to a layered defense. Thus, if this NMD system fails to work the first time it is used against a deliberate attack by a rogue state or an accidental launch by another nuclear power, the nation will have wasted some $100 billion, and caused unspeakable damage to its populace.

Many people argue that the United States has overcome difficult technological challenges in the past 50 years: the Manhattan Project to develop the atomic bomb in the 1940s; the development of the ICBM in the 1950s; and, the Apollo program of the 1960s. There is no doubt that this nation has the technological capacity needed to build a homing kill vehicle that can hit a high-speed warhead traveling through outer space. And, given enough time and money, the Department of Defense could do this with a high degree of reliability on the test range, i.e., 85 percent effectiveness at a 95 percent confidence level against a small-scale missile attack.

But, this is not what we are doing. In order to meet the initial deployment date of 2005, the Ballistic Missile Defense Office (BMDO) is doing concurrent weapons development—researching, testing and building all at the same time. It is compressing the work of 12 - 16 years into eight. This has lead to what General Larry Welch’s panel aptly describes as a rush to failure. It is no wonder that two of the last three have been complete failures and one only a marginal success. The last major concurrent weapons development program undertaken by the Pentagon was the B-1 Bomber. In the early 1980s, the Department of Defense began production of the B-1 three years before its development testing was completed. This rushed development led to chronic problems with the aircraft’s electric system which persist 20 years later.

Moreover, the current testing program does not simulate real world conditions, i.e., the type of countermeasures likely to be employed by an enemy with the capability to launch an ICBM thousands of miles. Rather, the three tests of the proposed system have been against cooperative targets and even under these less demanding requirements, two of three tests have failed. Remember that the Patriot anti-missile system had a perfect record (17 out of 17) in tests, but failed almost completely against Iraqui Scuds during the Persian Gulf War because of inadvertent Iraqui countermeasures (the missiles broke apart during reentry). Since the proposed system will have to hit its target in space, and since in the vacuum of space, both heavy and light objects travel on nearly identical trajectories, large numbers of cheap decoys most likely would be deployed along with the warhead by the attacking nation.

To obtain a 95 confidence level that the kill probability will be 85 percent, the proposed system should be tested successfully in a development mode a minimum of 20 times against the type of countermeasures that could be used by an attacker. (By way of contrast, the Sentinel System was tested over 100 times and Safeguard 42 times.) If there are three failures in a test series, 47 tests would need to be successful in order to provide 95 percent confidence that there would be an 85 percent probability of a single shot kill probability.

What is the record to date? Overall, since 1976, when research on hit-to-kill weapons began, attempts to destroy mock warheads have failed more than 70 percent of the time. Since last October, the proposed system has been tested on three occasions against limited countermeasures (one decoy rather than nine). Two of the tests have been outright failures: in January 2000, a leak of sensor coolant made the EKV miss the target altogether and in July 2000, the EKV failed to separate from the booster rocket. Even the one successful test in October 1999 raises doubts about the effectiveness of the system. The EKV drifted off course and actually was homing in on a large balloon (decoy) when the warhead drifted into its path.

By any reasonable criteria, we have not come close to meeting the 95 percent confidence level that there is an 85 percent probability that this proposed system will be effective, and even the strongest proponents of NMD do not favor deploying this system. Some of these advocates are now recommending that we switch to a Boost-Phase system which would attempt to destroy an enemy missile soon after its launch, before its warhead deploys. But, this is easier said than done and involves developing a new, more advanced interceptor, as well as more sophisticated sensor, radar and command systems than now exist. In my view, this will take a minimum of seven years of vigorous research and development before we can make an informed deployment decision.

Lawrence J. Korb

Vice President and Director of Studies

Council On Foreign Relations

58 East 68th Street

New York, NY 10065