Defense Secretary Richard B. Cheney yesterday ordered all U.S. short-range nuclear-tipped missiles promptly removed from strategic aircraft on war "alert," pending completion of a special inquiry into risks that the weapons could accidentally explode.

His decision came several weeks after directors of the three U.S. nuclear weapons laboratories unexpectedly told a congressional hearing that Short-Range Attack Missile-As (SRAM-As), should be immediately withdrawn from Air Force "alert" bombers and put into storage because of unacceptable safety risks.

The directors described their longstanding fear that inadvertent aircraft fires could detonate volatile materials in the SRAM-As, widely dispersing radioactive, cancer-causing plutonium at the core of the nuclear warheads on each SRAM-A.

"I believe it is prudent to take this action" until the weapons laboratories and the Air Force finish their study of SRAM-A safety risks later this summer, Cheney said in a statement released late yesterday at the Pentagon, which added that in Cheney's view the weapons pose "no safety hazards to the public."

Officials said the decision affects dozens of "alert" B-1, B-52 and FB-111 bombers at bases in the United States that are loaded with SRAM-As, fueled and ready for takeoff on short notice in event of a crisis. Hundreds of the missiles are to be withdrawn from these planes and put into nearby "secure" storage by Monday morning, the officials said. The numbers are classified. No SRAM-As are deployed outside the United States, and none are used in routine bomber training flights, the officials said. Nuclear bombs will remain deployed on some of the "alert" aircraft.

Pentagon spokesman Pete Williams had initially said that no action on the SRAM-A would be taken until the safety study was completed. But the laboratory directors' testimony provoked calls for action by key legislators, including House Speaker Thomas S. Foley (D-Wash.), whose congressional district has an Air Force base equipped with "alert" B-52s.

Cheney responded by calling a meeting of officials, including senior Air Force officers; Joint Chiefs of Staff Chairman Colin L. Powell, and Energy Secretary James D. Watkins, on the SRAM-A on May 26 to hear the lab directors' concerns.

Watkins, whose department is responsible for nuclear weapons design and production, had pressed for months at the behest of the directors to get high-level attention for the SRAM-A safety problems. But Defense Department officials, who considered the directors' concerns overblown, objected.

Four members of the Senate. Armed Services Committee, including chairman Sam Nunn (D-Ga.) and ranking minority member John W. Warner (R-Va.), told Cheney in a letter released yesterday that the lab directors' testimony had "reinforced concerns this committee has held for some time." The legislators added that "continued deployment of the current SRAM-A, under normal peacetime alert conditions, may not be consistent with a strong and safe nuclear deterrent."

A fifth member of the committee, Sen. Jeff Bingaman (D-N.M.), said separately yesterday that he applauds Cheney's decision and hopes the department "will ultimately decide to make this action permanent."

The Air Force said 1,500 SRAM-As were produced between 1971 and 1975 but officials declined to comment on independent estimates that only 1,100 of the missiles remain in service. Each missile has a range of up to 137 miles and is designed to destroy Soviet anti-aircraft batteries or heavily defended military installations with a nuclear blast equivalent to roughly 200,000 tons of TNT.

Weapons scientists say that the SRAM-As were built under much looser safety standards than those in place today and that their W-69 warheads lack fire- and explosion-resistant materials now commonplace in other nuclear weapons.

Roger Batzel, a Lawrence Livermore National Laboratory director, told a closed hearing of the Senate Appropriations subcommittee on defense several years ago that an accidental fire on an "alert" B-52 bomber at Grand Forks, N.D., in 1980 nearly caused a SRAM-A explosion that would have dispersed plutonium as far as 20 miles away.

"The wind happened to be blowing down the axis of the airplane," he said of the accident. "Had the wind been blowing across, rather than parallel to the fuselage, that whole system would have been engulfed in flames," including the SRAM-A missiles.

"There is a real world out there and those kinds of accidents do happen," Batzel said in what officials described as one of several blunt warnings from the labs. "You are talking about something that in one respect could be probably worse than {the 1986 Soviet nuclear reactor fire at} Chernobyl . . . because you have plutonium in the soil and on the soil, which you have to clean up."

According to declassified 1989 congressional testimony by a senior Air Force official, the SRAM-As were expected to have a "shelf life" of roughly five years but have been recertified every two years after inspections. The official, Lt. Gen. Ronald W. Yates, added that "Today, about 38 percent of the SRAM-As . . . we test fail some avionics or control testing" and that by 1993, roughly half of the missiles were expected to fail some tests.

Although some computer parts for the SRAM-A are no longer produced, the Air Force has been able to keep the weapons in service by repeated refurbishment. A new strategic missile intended to replace the SRAM-As at a cost of $2.4 billion has repeatedly been delayed by technical problems and funding shortages, officials said, and will not begin to be deployed until 1994. The SRAM-As will not be completely retired until 1998, according to current schedules.