The next catastrophic earthquake on California's San Andreas Fault might be triggered by a moderate jolt that ruptures a little-known fault hours earlier, says a study by New York scientists.

"We have noticed a pattern of earthquakes, which, if it continues, could trigger the southern San Andreas" to snap in a "major" quake measuring 7 on the Richter scale or a "great" quake of magnitude 8, said geologist Ken Hudnut, of Columbia University's Lamont-Doherty Geological Observatory.The next quake in the pattern of northward-migrating temblors is likely to occur on the Extra Fault 90 miles northeast of San Diego, Hudnut and seismologists Leonardo Seeber and Javier Pacheco said in this month's issue of Geophysical Research Letters.

A magnitude-5.5 temblor on the Extra Fault "could be all it takes to start a large-to-great earthquake on the San Andreas Fault," which the Extra Fault intersects, Hudnut said by phone from Palisades, N.Y.

"It's a reasonable scenario," U.S. Geological Survey seismologist Lucile Jones said Wednesday.

Hudnut's theory, along with others about how California's long-dreaded "big one" could begin, soon will be reviewed by a USGS committee co-chaired by Jones and California Institute of Technology geologist Kerry Sieh.

"We want to consider ahead of time how we would respond to various smaller earthquakes and other physical phenomena that might be precursors to a great earthquake," Sieh said.

Jones said the committee will decide whether and when public warnings should be issued if ominous signs of an impending "big one" are detected.

Such warnings could spur evacuations of dangerous buildings, preparations of emergency crews and equipment, and precautions by citizens, said Mike Guerin, spokesman for California's Office of Emergency Services.

A USGS study last summer concluded that a major or great quake is at least 60 percent likely within 30 years somewhere on the southern San Andreas.

If Hudnut's theory is correct, the quake would start where the Extra Fault intersects the southern end of the San Andreas' Coachella Valley segment, an area already considered overdue for a big jolt.

If the quake was limited to that 60-mile segment, which stretches from the Salton Sea northwest to near Palm Springs, it would be a magnitude-7 quake, Jones and Sieh said.

However, the San Andreas might continue to rupture through an additional 120 miles, on its San Bernardino Mountains and Mojave segments, which are closer to Los Angeles. That would be a magnitude-8 "great" quake, nicknamed "the big one," Jones and Sieh said.

A 1980 federal report estimated such a quake would kill 3,000 to 14,000 people, seriously injure 12,000 to 55,000, and cause up to $17 billion in damage.

There is no indication when the Extra Fault might snap. But the theory it might trigger "the big one" is supported by computer calculations of fault behavior and a pattern in which quakes on certain "cross-faults" have migrated north since 1979, Hudnut said.