Weak Modeling Hampers Plans for Operating U.S. Missile Defenses, Pentagon Study Finds

By David Ruppe
Global Security Newswire

WASHINGTON — U.S. military officials lack adequate tools for judging the likely effectiveness of the long-range national missile defense system President George W. Bush has ordered activated later this year, according to a recent Defense Department report and nongovernmental experts.

As a result, the U.S. Strategic Command, which has overall responsibility for operating the system, has not received critical information needed to develop operational plans, according to the experts and the Defense Science Board report Missile Defense Phase III, Modeling and Simulation, which was released by the Pentagon in March.

Furthermore, Missile Defense Agency development and procurement decision-makers also lack sufficient assessments of capabilities needed to perform cost-benefit assessments, they said.

The problems are caused by the agency conducting insufficient modeling and simulation of the Ballistic Missile Defense System and its components, tools that are key for understanding the operational capability of the system. Modeling and simulation is a discipline that enables weapons developers to examine potential capabilities of weapons systems by approximating operations in the field to provide data for making managerial or technical decisions.

The report concludes, “The GMD model treats some critical parameters such as discrimination [warhead identification] and overall probability of kill too simplistically for adequate credibility,” referring to the Ground-based Midcourse Defense system, pieces of which were ordered by Bush to be activated this year.

The ultimate objective of the ballistic missile defense system’s development is to produce a high “kill probability,” which is Missile Defense Agency jargon for the estimated probability of successfully destroying an enemy warhead.

For modeling and simulation to be useful, the report says, agency models must include “credible, higher-fidelity” assessments of the GMD system’s discrimination capabilities and kill probability.

Modeling of the overarching Ballistic Missile Defense System, of which the GMD project is just one component, also is deficient, it says.

The Missile Defense Agency lacks useful models for boost-phase intercept systems, it says, and lacks a single model for assessing the potential capabilities of the entire system that could be useful in deciding which systems to develop for fielding in two-year “block” increments, it says.

The agency should “improve the general level of fidelity for ‘hardware-in-the-loop’ testing [simulation testing of hardware] of all BMDS elements,” the report states.

Critics reviewing the report said its findings undercut Pentagon assertions about the system’s capability used to justify the deployment plans.

“They’re making detailed claims about how well it will work and they have no idea about whether it will work at all. … They have no idea of what the capability of this system is,” said MIT professor Ted Postol, a prominent critic of the program.

Last year, a senior Pentagon official estimated to Congress that the system would have a 90-percent kill probability using multiple interceptors against a single North Korean ICBM warhead (see GSN, March 21, 2003).

The Missile Defense Agency this year invited reporters to participate in a computerized war game that reportedly assumed a 91-percent kill probability for the entire BMDS system as it is projected to be configured years from now (see GSN, March 17).

Missile Defense Agency Director Lt. Gen. Ronald Kadish has been less specific about the system’s kill probability. In the June issue of the National Defense Industrial Association’s National Defense magazine, he cited agency modeling and simulation efforts to argue that the fielded system would be effective regardless of the outcome of two flight-intercept tests scheduled for this year.

“We will have the capability in September. Modeling and simulation predicts with great precision what would happen if it works as designed,” he said.

Philip Coyle of the Center for Defense Information, who was the Pentagon’s top testing official during the Clinton administration, challenged that level of certainty.

“Modeling and simulation of missile defense has never predicted actual system performance with great precision,” he said.  

“Current models and simulations simply do not capture the most significant variables in a missile defense engagement, and aren’t likely to any time soon. The battle environment is highly complex, the software programming requirements are daunting, and the models don’t begin to capture the real physics,” he said.

Coyle, for example, noted the report’s conclusion that models of potential targets were based on “uncertain or unavailable data.”

Bill Sowder, a missile defense expert at Teledyne Brown Engineering, agreed that Pentagon testing has been simplistic so far.

“We’re sort of in kindergarten right now entering first grade and we’re trying to move up,” he said.

“This is a monumental program, much beyond the Manhattan project or the development of Atlas, the first ICBMs. This is much bigger than that.  We have not had many chances in this country to look at, and develop and actually test and actually deploy a system of systems,” he said.

A statistical analysis Coyle co-authored and released last month argued that the GMD system’s kill probability is unknown given the limited testing so far and is probably extremely low. Such a low probability would undermine the Pentagon’s claims that simply deploying more interceptors could ensure success. The analysis found that firing large numbers of low-kill-probability interceptors would do little to increase the chances of shooting down all enemy warheads (see GSN, May 14).

The military “has conducted exercises for the media, the purpose of which was to show that more interceptors did help. So clearly those models have loaded into them very high kill probabilities, but they have no basis of asserting those kill probabilities given the basis of what the tests have done so far,” Coyle said in an e-mail.

Kadish and other officials have said they do have an assessment of the capability but would not release it because it is classified.

The Missile Defense Agency also told GSN in a statement that, “Over time we are building up our modeling and simulation capability at the system level to approximate more closely the type of end-to-end testing we would like to have to verify that the system is doing what we want it to do.”

It said the agency has followed the report’s recommendation to appoint a program director for modeling and simulation and has created a new modeling and simulation directorate.

Implications of the reported modeling and simulations deficiencies, according to the Defense Science Board report and the experts, are several and significant. 

For instance, the report says that the Missile Defense Agency is not providing the U.S. Strategic Command, which has overall responsibility for using the system, a modeling and simulation tool for assessing operational effectiveness of each two-year block of new equipment it delivers.

The command could use such capability, it says, to develop regional and global contingency plans incorporating the system, write offensive war plans with defenses in mind, and “assess the likely effectiveness of the deployed system against evolving threats.”

Without the benefit of modeling and simulations assessments, he said, “you don’t have a technical basis for making operational decisions in a real battle,” he said.

“You have to characterize what the system can do before you have confidence that you can really meet the challenge that the adversary can throw at you,” Sowder said.

The report notes a lack of an “end-to-end system engineering and verification tool,” which experts say models the actual performance of the entire system of components.  

It would give operators “an idea of what you can expect when you call the system up,” Postol said.

It is “difficult to understand how BMDS could be put on alert without such a tool,” the report says.

Insufficient modeling capabilities also affect Missile Defense Agency program-planning decisions, according to the report.

Knowledge of system effectiveness, according to Coyle, affects planning for how many interceptors should be launched at a single target — and therefore calculations of how many should be purchased for defense against an estimated threat.

Because agency officials have “stopped flight-intercept testing and also don’t have realistic models, they have no real basis to show the Congress why they need the money they say they need,” he said.

U.S. military officials have said that the planned activation this year would give the country protection when it currently has none.

Back to top
About Newswire  |  Contact National Journal  |  Re-Use Guidelines © Copyright 2008 by National Journal Group, Inc. The material in this section is produced independently for NTI by National Journal Group, Inc. Any reproduction or retransmission, in whole or in part, is a violation of federal law and is strictly prohibited without the consent of the National Journal Group, Inc. All rights reserved.