Understanding the physical processes causing surface creep on faults

Since the 1960s, we have observed creep events along the central San Andreas Fault. Despite this, we still do not know (1) when they occur, (2) where they occur, (3) how they grow spatially, (4) how much moment they release, or (5) the driving physics behind them. To address these questions, I use creepmeter and strainmeter data from along the central San Andreas Fault.

Using a cross-correlation approach, I examine 18 USGS creepmeter records between 1985 and 2020 and identify 2120 creep events. I correlate these events between creepmeters and identify many large (> few km long) creep events. Next, I combine creepmeter and strainmeter observations at the northern end of the central San Andreas Fault to identify the along-strike location, depth, and magnitude of creep events. Strainmeters record strain offsets associated with creep events, which I reproduce by modelling the strain produced by a rectangular patch of slip at various fault locations and magnitudes. I find that creep events are not the repeated rupture of a single fault patch but arise from a range of along-strike locations at depths >4 km, with a magnitude of 3–4.

Finally, I use displacement-time evolution of creep events to discriminate between five rheological models for creep events: (1) rate and state friction at a constant state rate near steady state, (2) rate and state friction above steady state, and (3) rate and state friction below steady state (4) power-law flow, and (5) linear flow. Using a basin hopping approach, I minimise the misfit between the observed creep event and the predicted displacement from each model. Most creep events have their displacement evolution best described by power-law flow or rate and state friction below steady state. These surprising results suggest that additional work is required to understand the driving physics behind creep events.