The costly effort to develop a space-based missile defense is beset by so much scientific uncertainty that President Bush should not decide during his current term in office to proceed with its deployment,

according to an independent congressional study released today.

The General Accounting Office concluded after an 11-month study that several barriers will prevent managers of the Strategic Defense Initiative (SDI) program from giving competent technical advice on deployment to Bush by early 1993, as he has sought. It said this will make a decision to proceed "premature and fraught with high risk."

SDI managers "may be able to justify a decision to delay or cancel phase 1" of the strategic defense system by early 1993, said the GAO report. "However . . . {they} will not be able to support an informed decision to deploy it" because of a recent redesign of the so-called "Star Wars" system, insufficient testing and evaluation, and inadequate congressional funding.

The GAO report was given to reporters on Monday by Rep. John Conyers (D-Mich.), chairman of the House Committee on Government Operations, but was withheld from the Department of Defense to prevent comment from the Pentagon in advance of its public release yesterday. SDI officials said they could not comment on the report because they had not obtained a copy.

During his presidential campaign and in a June 1989 directive to the Pentagon, Bush said he wanted to decide by January 1993 whether to deploy space weaponry. At that time, SDI managers were planning a much more complicated and costly system than now favored, and pursued extensive testing to meet Bush's goal.

But the GAO report suggests that by scaling the system back earlier this year in an effort to cut costs and simplify development, the government has introduced new complications and uncertainties that cannot be resolved for at least four years.

"Much of the current test data and analyses will not be applicable to the new architecture, and many of the tests will likely have to be redesigned and repeated," the GAO said of some $19.8 billion of SDI research performed since 1985.

This was attributed to selection in January of a new plan to orbit thousands of small, autonomous rockets, known as "Brilliant Pebbles," to sense and try to hit a fraction of Soviet nuclear missiles in flight before they reach U.S. territory. An earlier plan would have orbited rockets in large "garages" to attack Soviet missiles with aid from sensors on separate satellites.

The GAO said this shift "highlights the instabilities of {the} design," which have "reduced, changed, or eliminated the need for some of the space-based elements" that figured in key computer simulations and other tests. None of the weapon elements needed for the new scheme, which the Pentagon says will cost at least $44 billion, has been "solidified," the report said.

SDI has never received as much money from Congress as the White House has sought; for fiscal 1991, Bush seeks $4.6 billion and the Senate Armed Services Committee last week proposed $1 billion less.

The report noted that SDI managers have no plans for "end-to-end" testing of an integrated defensive system prior to Bush's decision or the planned start of full-scale development in 1994. Such a test would use weapon and sensor prototypes, actual software and trained personnel making realistic decisions.

Instead, managers plan to simulate a mix of equipment, software and personnel in its early tests. A lack of end-to-end testing was recently cited for the failure to discover a major defect in the $1.6 billion Hubble Space Telescope.

White House spokesman Bill Harlow said he had not seen the GAO report and could not comment directly. But he said there is "every indication that SDI will be able to" provide adequate guidance on deployment within Bush's timetable.