The Council on Environmental Quality is proposing to limit the use of "worst-case" scenarios in deciding whether federal projects would wreak unacceptable havoc on the environment.

Under the National Environmental Policy Act, federal agencies must prepare detailed analyses whenever they propose actions that could have a significant environmental impact.

Courts generally have ruled that worst-case analyses are required in environmental impact statements under the 1969 law as a way to gauge the effects of catastrophic accidents, such as a hazardous waste landfill overflowing in a massive flood.

But the CEQ, in a Federal Register notice last week, said that it is preparing new advice for federal agencies that would allow them to consider a "threshold of probability" before including a worst-case analysis.

"We think this is a reasonable interpretation," said the CEQ's general counsel, Dinah Bear. "The worst should be seen in the right perspective, so it doesn't look like it's going to happen."

Environmentalists, however, pronounced the action "a back-door approach" to changing a rule that has thrown a monkey wrench into several major federal projects over the past decade.

"Fortunately, this is not a change in the regulations," said National Wildlife Federation lawyer Norm Dean. "If it were, it would be extremely troublesome. But as it is, it only constitutes bad advice."

The CEQ, which is responsible for regulations under the environmental policy act, said it is concerned that its worst-case rules "are being read to require federal agencies to conduct such analysis for potential effects that may well be highly remote or unlikely."

That, as it happens, is exactly how the 5th U.S. Circuit Court of Appeals read the regulations earlier this year when it put the kibosh on a proposed deep-water port and oil distribution center in Galveston, Tex.

In a 26-volume environmental impact statement, the U.S. Army Corps of Engineers had failed to consider what would happen if a loaded supertanker sank in Galveston Bay. Despite the Corps' argument that such an accident is unlikely, the court ruled that it couldn't issue the necessary construction permits for the port until the Corps calculated the effect that the massive oil spill resulting from such an accident would have on the bay.

Similarly, a district court in Oregon ruled that the Interior Department could not spray forests there with 2,4-D until it had either completed a worst-case analysis or cleared up the scientific uncertainties about the herbicide's potential to cause cancer and birth defects.

The CEQ said that it is "increasingly apparent" that the worst-case analysis "has been subject to a wide variety of conflicting interpretations by both federal agencies and reviewing courts."

But legal advisers for environmental groups said that they don't know of any conflicting court opinions.

The Galveston ruling, which marked the first time the requirement had been interpreted at the appellate level, "was very clear and very logical," according to Rick Middleton, an attorney for the Sierra Club Legal Defense Fund, which filed the case.

"The type of things we're concerned about are major oil spills, nuclear meltdowns--things that we know can happen, not made-up science fiction stuff," he said. "We all admit they do not happen that frequently, but when they do it's catastrophic."

Middleton, like Dean, said he does not believe the CEQ's proposed "guidance" will be legally binding on agencies. "This is just the Reagan administration's advice on how they wish it would be carried out," he said. "But bad guidance will lead to bad decisions. And we'll have to file more lawsuits."

Bear emphasized that the guidance is still in the draft stage and "can change quite drastically."

But she said that the proposal is an attempt to deal with the concerns of scientists, who have to wrestle with the crystal balls. "The scientists have to do it," she said. "All the lawyers have to worry about is when you do it."