"Tremors seem to be extremely sensitive to minute stress changes," said co-author Professor Roland Bürgmann of the UC Berkeley faculty of earth and planetary science.

"Seismic waves from the other side of the planet triggered tremors on the Cascadia subduction zone off the coast of Washington state after the Sumatra earthquake last year, while the Denali earthquake in 2002 triggered tremors on a number of faults in California."

"Now we also see that tides -– the daily lunar and solar tides -– very strongly modulate tremors."

It is this extreme sensitivity to stress that lead researcher to conclude that highly pressurised water underground was acting as a lubricant.

"The big finding is that there is very high fluid pressure down there, that is, lithostatic pressure, which means pressure equivalent to the load of all rock above it, 15 to 30 kilometers (10 to 20 miles) of rock," said co-author and seismologist Robert Nadeau of the Berkeley Seismological Laboratory.

"Water under very high pressure essentially lubricates the rock, making the fault very weak."

UC Berkeley seismologists began looking for tremors five years ago in seismic recordings from the Parkfield segment of the San Andreas Fault.

Using eight years of tremor data, researchers correlated tremor activity with the effects of the sun and moon on the crust and with the effects of ocean tides, which are driven by the moon.

Althogh tidal activity is not known to cause quakes directly, it has been known to cause clusters of deep tremors.

Researchers are still trying to develop a coherent model to explain the behaviour of the San Andreas Fault.

"These tremors represent slip along the fault 25 kilometers (15 miles) underground, and this slip should push the fault zone above in a similar pattern," Professor Bürgmann said.

"But it seems like it must be very subtle, because we actually don't see a tidal signal in regular earthquakes."

"Even though the earthquake zone also sees the tidal stress and also feels the added periodic behaviour of the tremor below, they don't seem to be very bothered."

But Nadeau emphasises: "It is certainly in the realm of reasonable conjecture that tremors are stressing the fault zone above it."

"The deep San Andreas Fault is moving faster when tremors are more active, presumably stressing the seismogenic zone, loading the fault a little bit faster."

"And that may have a relationship to stimulating earthquake activity."

Nadeau is certainly encouraged by the new study and feels that more research is warranted.

"There is still all lot to learn about tremor and earthquakes in fault zones."

"The fact that we find tremors adjacent to a locked fault, like the one at Parkfield, makes you think there are some more important relationships going on here, and we need to study it more."