Scientists at Caltech and the U.S. Geological Survey are warily examining new evidence that may indicate a major earthquake is about to strike Southern California.

While no one has yet said when the big one is coming, seismologists living in the Los Angeles area are removing heavy objects from their shelves and tying down water heaters at home in anticipation of a temblor of a magnitude of 7 -- or greater -- on the Richter scale.

There are, one scientist says, "things strange" happening near the San Andreas Fault.

Pasadena has drifted about 9 inches west in the past six months. Land that should be compressing is stretching instead. Exotic gases are bubbling out of the earth.

This series of unusual events occurring here since June has led some scientists to believe a major quake is building up along a long-dormant section of the fault near Los Angeles.

A quake measuring 7 or greater could cause widespread death and destruction in the Los Angeles area.

A February 1971 earthquake centered in the northern San Fernando Valley measured 6.6 on the Richter scale. It killed 64 persons, injured 1,000 and caused an estimated $1 billion in property damage. The 1906 San-Francisco quake virtually demolished that city and killed about 700. It is now estimated to have measured 7.9.

Yet because no one is certain what the new evidence means, and because no procedure exists for announcing scientists' unproven fears, no clear warning has been issued to the public.

Barry Raleigh, a scientist at the Geological Survey regional headquarters in Menlo Park, Calif., said scientists "now have the first really well-documented observations that may be related to a forthcoming earthquake" in Southern California.

Raleigh cautioned, however: "Nothing we see could warrant making a prediction. A prediction implies there is going to be a large earthquake of some magnitude range in a particular place at a particular time. We're not prepared to do that."

The Geological Survey, charged by law with warning state authorities of geological hazards, has asked earthquake scientists working in Southern California to turn in research data to the Menlo Park headquarters for current evaluation, rather than biannually.

The agency has also held two meetings of earthquake scientists -- one in October and one last month -- to discuss their findings on the situation.

Three major anomalies -- unusual events -- have been observed since June:

Unexplained increases in radon gas emissions from three water wells in the San Gabriel Mountains. First observed in California in June, similar radon anomalies were associated in 1978 with three major earthquakes in the Soviet Union.

Scientists have found increases of up to 400 percent in the amount of radon, a rare radioactive gas, emitted from the wells.

The anomalies "do not tie in to any of the smaller earthquakes [recorded in Southern California lately], so we have to assume that what is going on is tied in to some bigger picture," said California institute of Technology researcher Thomas Tombrello. "Whether this has any connection to some larger earthquake I just couldn't guess."

Other scientists have reported apparently unprecedented changes in the direction and type of strain surrounding the San Andreas Fault near Los Angeles.

Instead of compressing in a north-south direction, which would be the normal or "comfortable" situation because it would thus help hold the east-west running fault section immobile, the land just north of Los Angeles is stretching on an east-west axis -- roughly parallel to the fault line.

That section has not shifted since a major earthquake hit there in 1857. Scientists believe that it moves, on average, every 140 years.

In November, sophisticated measurement of the distance between Pasadena and the Owens Valley showed that Pasadena has, since April, moved about 24 centimeters -- or 9 inches -- to the west and south, relative to the northern part of the state. The measured movement crossed the San Andreas Fault.

"There's a possibility that we're seeing some phase of the earth's [normal] behavior," said Pete MacDoran, project manager of Caltech's Astro Radio Interferometric Earth Survey, which uses high-frequency radio waves to measure long distances to within a few inches.

"But it's been an article of faith in modern geophysics and plate tectonics [the current theory of how earthquakes happen] that the ground has to deform to store energy for future earthquakes."

In addition to these undisputed findings, other researchers have reported variations in the earth's magnetic field, ground tilt and the timing of smaller earthquakes around Southern California.

All this together could -- but does not necessarily -- add up to "an earthquake above [magnitude] 7," said Caltech seismologist Kate Hutton.

"There are definitely things strange happening; nobody would doubt that," she said. "I don't think we know enough to be specific, but I think there might be some danger.'"

Other scientists expressed concern in more conservative terms.

"I don't believe I should go any further than the data allow me," said Caltech's Tombrello.

"It's tempting to try to see this [situation] in terms of some overall tectonic effect," he said "And it's pretty exciting in terms of science. But the relevance of the science to earthquake predition is tenuous."

Carl Johnson, studying earthquake patterns for the Geological Survey at Caltech, said he is "troubled by the lack of a proper explanation" of the three anomalies.

"In my own mind, it's quite clear we are observing a transient phenomenon that involves rapid straining of the [earth's] crust," Johnson said.

"The sorts of things we are observing here in various combinations have been observed before larger earthquakes. The possibility this may be as- sociated with activity . . . leading up to a larger earthquake cannot be ruled out."

A key problem that researchers face is that most of the data in which they are now interested in has been recorded for only the last 10 years or less.

That means that no one can say when an earthquake might be coming, if in fact, one is imminent.

"Since really major earthquakes occur on much larger time intervals than that, it's very hard to relate [the current situation] to a specific earthquake," said a scientist studying radon emissions. "It's interesting, but this could go on every 10 years without anything happening."

But the uncertainty of the situation has illustrated a different problem for the fledging science of earthquake prediction: No scientist really wants to be the one to predict a quake, whether the forecast is accurate or not.