GAO and Welch Reports Find NMD Program Still Risky

July/August 2000

Two follow-up reviews of the Clinton administration's proposed limited national missile defense (NMD) system made public in June cautioned that despite Pentagon restructuring of the program, schedule and performance risks persist. Both reports, one conducted by an independent panel of experts commissioned by the secretary of defense and another by the General Accounting Office (GAO), detailed testing delays and the limited number of flight tests against possible threats, though the independent panel concluded that the "technical capability" to develop and field the limited system is available. The Pentagon concurred with the GAO findings and a top defense official said the independent panel report "pleased" the Pentagon.

The Clinton administration is developing a missile defense system for deployment in 2005 to protect all 50 U.S. states from a limited attack by or accidental launch of strategic ballistic missiles. The proposed system, which would violate the 1972 ABM Treaty barring a national defense, would be initially composed of some 20 interceptor missiles based in Alaska with another 80 to be added two years later. President Bill Clinton will decide later this year whether to start the deployment process by officially selecting the missile defense site and authorizing construction contracts for building an advanced NMD radar in Alaska.

Convened for a third time, the independent panel, composed of 14 experts and headed by former Air Force Chief of Staff General Larry Welch, described the program as "on track to achieve the earliest capability to meet the defined limited threat," but warned that meeting the target deployment date with "required performance remains high risk." Appearing before the Senate Armed Services Committee on June 29, Welch, when asked if he thought the system would be ready for deployment by 2005, answered, "Do I believe it's feasible? Yes. Do I believe that's the most likely? No."

Although not predicting whether the deployment goal would be met, GAO, charged with updating its June 1998 NMD report, warned that program "performance and schedule risks remain significant because of the technical challenge, test limitations, and the ambitious schedule." GAO said the program's aim of developing a highly reliable, very effective hit-to-kill capability remains a "difficult one."

Underscoring this challenge, the GAO report, dated May 2000, noted that since 1983 there have been 14 intercept attempts of missiles outside the atmosphere with only four successes. The current system has two misses and one hit, which the Pentagon Office of Operational Test and Evaluation said was aided by a decoy near the target. GAO did not account for the latest intercept failure on July 8. (See p. 30.)

Both reviews highlighted issues surrounding the NMD interceptor's booster, which will carry the exoatmospheric kill vehicle (EKV) to a point in space where the EKV will seek out and collide with the target. At present, a less powerful surrogate booster is being used in flight testing with plans to introduce the actual booster in an early 2001 intercept test. Yet none of the three preintegrated, solo flight tests of the booster, the first of which had been scheduled for April 17, have taken place or been rescheduled because of problems modifying the three commercial boosters that will form the actual booster. The booster program is now eight months behind schedule.

The Welch panel noted in a June 13 executive summary of its classified report (see p. 37) that delaying the integrated booster intercept test would "probably require" an adjustment in the program's decision-making schedule. Top Pentagon officials subsequently said on June 20 that the use of the actual booster in an intercept attempt might be delayed beyond the originally planned test date.

Additionally, GAO pointed out the actual booster will place "much higher acceleration and vibration loads on the kill vehicle" than the surrogate test booster, raising questions as to whether the EKV will be able to withstand the greater forces. Though program officials assert that ground testing indicates the EKV will handle the actual booster's higher acceleration, GAO noted that Welch's November 1999 report characterized this uncertainty as a high program risk. (See ACT, November 1999.) The actual booster will not be mated with the production-representative EKV (as opposed to the prototypes currently being used) until a 2003 test, after which a decision will be made on whether to build and deploy the interceptors.

Flight Testing Limitations

GAO also noted that all NMD intercept-test targets fly from east to west, whereas the proposed defense would have to counter missiles from several directions. Because testing targets are fired from California toward the Marshall Islands in the Pacific Ocean, GAO indicated that the early-warning radar, which is located close to the launch point, immediately has a "large, easy target" to detect, while in the future, the radar would face the more difficult task of initially tracking a target when it is launched from thousands of miles away.

Current flight testing, according to the Welch panel, involves only a "limited part of the required operating envelope." Where a missile is fired from in relation to the location of the missile defense interceptor site will affect the time needed to make the intercept, as well as the angle at which the interceptor will have to fly to hit the target. Due to safety restrictions, however, current NMD testing is restricted to a narrow band over the Pacific Ocean, thereby confining testing to a limited number of intercept variables. The Welch panel recommended the test envelope be "expanded beyond that now permitted."

To compensate for restricted flight tests, the NMD program will rely heavily on computer simulations to assess the system's effectiveness against threat scenarios that cannot be flight tested. But GAO reported that the primary tool for conducting such simulations may not be available this summer for use prior to the Pentagon's deployment readiness review, which will be the basis for Secretary of Defense William Cohen's recommendation on whether the system's technology can support a presidential deployment decision.

GAO said that uncertainty about countermeasures that could be used to penetrate the defense also raises risks to the system's future performance. A September 1999 National Intelligence Estimate warned that Russia and China, two countries that have "developed numerous countermeasures," would likely sell technology related to those countermeasures and that countries seeking long-range missiles— such as North Korea, Iran, and Iraq—could develop countermeasures, including simple decoys, based on readily available technology by "the time they flight test their missiles."

The Welch panel, however, said the system's design capabilities to discriminate between warheads and decoys are "adequate" to meet the few, if any, simple decoys that Pentagon officials contend will be available to potential attackers by 2005. "There is extensive potential in the system design to grow discrimination capabilities," the panel concluded. Nevertheless, the panel recommended starting a program now to address potential future countermeasures.

Originally, the Clinton administration set a 2003 date for NMD deployment, but after the original February 1998 Welch report assessed the program as being on a "rush to failure," Cohen in January 1999 pushed back the deployment goal to 2005. Despite this shift, GAO pointed out that the schedule is still "much shorter" than the timeline for most missile defense acquisitions.

The Welch panel described the NMD schedule as "high risk," a judgment shared by the Pentagon, but the panel saw "no technical reason to change the schedule at present." Yet the panel emphasized the schedule will be "self-adjusting as needed" so long as "meeting performance milestones is the criteria for moving to the next event or next decision."

In his June 29 testimony, Welch stressed the difficulty of the system's task and noted that "there's probably only about a thousand things than can go wrong." He described this estimate as "conservative." Each month of delay, according to the GAO report, will cost $124 million.