Geologists in the the Soviet Union predicted a major earthquake in central Asia the day before it happened Nov. 1, according to reports reaching th U.S. Geological Survey.

While it was the second quake forecast a day ahead of time, the Soviet quake may have been the first predicted with the kind of precision preferred by scientists in evaluating the precursors of earthquakes. The Chinese predicted an earthquake the day before it struck Hai Cheng Feb. 5, 1976, evacuating people from homes in the region on Feb. 4.

"What impressed us about the Soviet prediction was the large number of measurements they had over more than a year's time," said Dr. Rob Wesson, chief of the Office of Earthquake Studies in the USGS. "Then in mid-October last year they began to see some abrupt changes in seismicity and the day before [Oct. 31] saw even more abrupt changes in water levels."

That day the water levels in 10 Artesian wells fell suddenly in regions within 100 miles of where the quake's epicenter was recorded the following day. The water level in a well one mile deep fell abruptly after having risen continuously for weeks.

"Based mainly on the water well data, the Soviets made their prediction the afternoon before the earthquake." said Dr. David Simpson of Columbia University's Lamont-Doherty Observatory. "There is a growing belief among geologists that in the final stages before an earthquake there is a general swelling in the earth's crust that opens up cracks that allow ground water to seep downward."

A quake that registered 6.7 magnitude on the Richter scale struck a region in the Alai Mountains of south-central Kazakhstan Nov. 1. Its epicenter was in an uninhabited region at an elevation of 10,000 feet and about 100 miles south of Andizhan and 100 miles east of Garm.

"It was a large quake; nobody was injured but it was felt as far away as Tashkent 700 kilometers from the epicenter," said Simpson. "I know I was in Dushanbe [about the same distance as Tashkent] at the time and I felt it."

Aware of the attempts Soviet geologists were making to predict earthquakes in central Asia, Simpson returned to the United States and wrote a detailed report of what he was told of the Nov. 1 prediction for USGS and its Office of Earthquake Studies. Two weeks ago White House science adviser Frank Press returned from a trip to Moscow, where he was told of the Soviet prediction.

Soviet geologists told Simpson they had expected for some time a large earthquake in the region where the Nov. 1 quake struck. There had been four quakes from 1976 until March 1978 in regions surrounding the Alai Mountains, suggesting that the strain in the earth's crust that was relieved by those four quakes was being transferred elsewhere.

Seismic instruments at four locales in Kazakhstan measure changes in the earth's tilt at the same time they noticed new deformations in the earth's crust north and south of where the quake struck. In the months before the quake, there was a sudden decrease in seismic shocks north of where the quake occurred.

"The Soviets even noticed an increase in radio static in the air before the quake and recorded it as pertinent data," Simpson said. "They measure everything they possibly can, and it works, they use it."

In any case, the Soviets finally predicted that an earthquake would strike the Alai Mountains the day before it happened when they saw the falling water levels in the Artesian wells. Their prediction came within a few hours of the quake, was off in size by about one-half a magnitude on the Richter scale and missed the epicenter about 100 miles.