Seasonal rain and snow trigger small quakes on California faults

California’s winter rains and snow depress the Sierra Nevada and Coast Ranges, which then rebound during the summer, changing the stress on the state’s earthquake faults and causing seasonal upticks in small quakes, according to a new study by University of California, Berkeley seismologists.

water loading stresses California's faults

Stresses on California’s earthquake faults change from month to month because of water loading in the mountains surrounding the Central Valley. In particular, the San Andreas fault sees peak stress in midsummer/early fall, while the faults east of the Sierra Nevada show peak stress in late spring/early summer.

The weight of winter snow and stream water pushes down the Sierra Nevada mountains by about a centimeter, or three-eighths of an inch, while ground and stream water depress the Coast Ranges by about half that. This loading and the summer rebound – the rise of the land after all the snow has melted and much of the water has flowed downhill – makes the earth’s crust flex, pushing and pulling on the state’s faults, including its largest, the San Andreas.

The researchers can measure these vertical motions using the regional global positioning system and thus calculate stresses in the earth due the water loads. They found that on average, the state’s faults experienced more small earthquakes when these seasonal stress changes were at their greatest.

The central San Andreas Fault, for example, sees an increase in small quakes – those greater than magnitude 2 – in late summer and early fall as the water load diminishes in the mountains. Most people can’t feel quakes below about magnitude 2.5.

The faults along the eastern edge of the Sierra Nevada see an uptick in late spring and early summer due to this seasonal unloading, the researchers found.

“It’s not that all earthquakes happen in September. There is no earthquake season,” said Roland Bürgmann, a UC Berkeley professor of earth and planetary science and the senior author of a paper appearing this week in the journal Science. “It all depends on details of the loading, the location of the fault and the geometry of the fault.”

While the impact of this annual up and down movement of the mountains surrounding the Central Valley is small – increasing the chance of earthquakes by a few percentage points at most – the study gives seismologists information about how faults rupture and what kinds of stresses are important in triggering quakes.

“This study supports the notion that the state’s faults are critically stressed so that these small perturbations can affect the earthquake cycle and sometimes promote failure,” said first author Christopher Johnson, a UC Berkeley graduate student. “It is advancing the clock on these different faults.”

Earthquake stress

Previous studies have shown that daily stresses caused by the ebb and flow of ocean tides don’t seem to trigger small or large quakes in California. Rarely, though, extremely large earthquakes – megaquakes – can trigger large quakes thousands of miles away. An 8.6 magnitude quake that occurred in the Indian Ocean in 2012 triggered 16 large quakes greater than magnitude 5.5 around the world.

water loading 2009-2014

Month-by-month water storage in California from 2006 through 2014, estimated from the regional GPS network. The scale is the average water-layer thickness on a 25-kilometer grid patch. Blue represents higher than average water load, while red is lower than average. The summer months experience a loss of water in the Sierra Nevada and Central Valley that is replenished in the mountains during the wet winter months as reservoirs fill and snow accumulates. In the Central Valley, water storage has been continuously decreasing due to large scale agriculture. The seasonal change in mass deforms the crust and the associated stress changes modulate regional seismicity.

The amount of stress generated by seasonal water loading in California is similar to the stresses induced by the seismic waves from distant megaquakes, Johnson said.

“We don’t see an increase in large-magnitude earthquakes from these low-amplitude stresses caused by seasonal water storage,” Johnson said. “What these results are showing, however, is that we do see a correlation with small earthquakes from low-amplitude stresses.”

Johnson and Bürgmann, members of the Berkeley Seismological Laboratory, looked at 3,600 earthquakes over a nine-year period, 2006-2015, and correlated their occurrence with the calculated peak stress on the fault where they occurred. The stress was calculated from the amount the mountains deformed, as measured by a GPS system, using models of rock mechanics that predict stress changes on faults.

“We are finding that on the central San Andreas, the late summer months are when we see most seismicity, and that correlates with the larger stress changes,” Johnson said. “It is not during the rainy season; it is more of the unloading that is resulting in the larger stresses, for that one fault.”

Interestingly, only shear stress – that caused by back and forth sliding motion – triggered an excess of quakes, not changes in compression that clamp or unclamp the fault.

The researchers also looked at all historic large quakes greater than magnitude 5.5 since 1781, and saw somewhat more earthquakes when water unloading stresses are high than when the stresses are low.

“We look at historic records for larger events, and we do see this seasonality, but we are not at the point that we can provide further evidence to hazard estimates that would say that during these periods of time we would expect more large earthquakes to occur,” Johnson cautioned.

peak stress from water

In 2010, the San Andreas fault (left image, red and orange lines) experienced peak stress in September after snowmelt and runoff allowed the mountains to rebound. Faults east of the Sierra Nevada (right, red and orange) experienced peak stress in March of the following year.

The studies are designed to better understand “what makes earthquakes go,” Bürgmann said. “Looking at the responses to these periodic stresses is like running a fault mechanics experiment at the scale of all of California.”

The work was supported by the U.S. Geological Survey National Earthquake Hazards Reduction Program and the Southern California Earthquake Center, as well as a National Science Foundation Graduate Research Fellowship to Johnson. Yuning Fu, of Bowling Green State University, is also a co-author of the paper.