In the early hours of Sunday morning, the Napa Valley region north of San Francisco was shaken by a magnitude 6 earthquake, the largest to hit this region since the magnitude 6.9 Loma Prieta earthquake in 1989. An earthquake in wine country produces dramatic visuals of wine bottles thrown from the shelves and leaking into the street, and there was strong shaking and significant damage in the region close the the epicentre. But in the grand scheme of things, a magnitude 6 earthquake is not particularly large: it may have been strong enough to wake people up in San Francisco (as data from fitness trackers has rather cleverly been used to illustrate), but it didn’t pack enough of a punch to do any real damage there.

There have already been some excellent write-ups of the geology of the earthquake by Mika Mckinnon, Dana Hunter and Garry Hayes: in brief, the root cause of all earthquakes in California is its location on the boundary between two rigid but independently mobile bits of the Earth’s crust: the Pacific and North American plates. The Pacific plate is moving northwest relative to the North American plate at around 4 centimetres a year, and the focal mechanism for Sunday’s quake is consistent with right lateral strike slip on a northwest-southeast oriented fault, which is what we’d predict for a fault at this plate boundary. But interestingly, we are clearly some way (about 50 kilometres) to the east of the San Andreas Fault, which is popularly considered to be ‘the’ plate boundary in California.

Location of Sunday’s earthquake relative to the San Andreas Fault. Yellow lines mark faults thought to have been active in the last 130,000 years, according to USGS mapping. The red line is the rough surface trace of Sunday’s rupture, as mapped by Mike Oskin’s speedy students, which appears to be associated with a previously unmapped strand of the West Napa Fault Zone.

This event actually nicely illustrates the nature of the plate boundary in California: we are not looking at a single line in the crust, where you can step directly from one plate to another, but a broad deforming zone. Nowadays, we have some lovely data that shows this directly, courtesy of the abundance of GPS stations deployed all through the western US. The red arrows in the figure below show contemporary motion of the crust relative to North America measured at seven stations, the most westerly of which is on the seaward side of the San Andreas Fault, and can therefore be considered to represent the Pacific plate. Although there is a significant drop in the rate at which crust is being dragged to the northwest after crossing the San Andreas, the crust in this region is still moving relative to the stable interior of the North American plate at some fraction of the Pacific plate’s velocity. As you move inland, each time you cross one of the mapped fault systems – which are themselves complicated amalgamations of anastomosing fault segments – the measured GPS velocity drops further, indicating that these systems are all actively involved in accommodating plate motions.

GPS velocities (in mm/yr) for selected UNAVCO Plate boundary observatory stations in Northern California.

The astute amongst you will note that even as the GPS velocity seems to stabilise 60-80 km inland from the San Andreas, it is still not zero; even here, the crust is not fully attached to North America. To get to the stable, non deforming bit, we still have to get across the Basin and Range, and before we get to that we have to cross the apparently more rigid Sierra Nevada mountains. The plate boundary in the western US is far from a well-defined line. On the plus side, this makes it much more interesting geologically; on the negative side, a broad zone of many faults makes estimating the seismic risk much more ‘interesting’, too.