Using Surface Creep Rate to Infer Fraction Locked for Sections of the San Andreas Fault System in Northern California from Alignment Array and GPS Data

Abstract

Surface creep rate, observed along five branches of the dextral San Andreas fault system in northern California, varies considerably from one section to the next, indicating that so too may the depth at which the faults are locked. We model locking on 29 fault sections using each section's mean long-term creep rate and the consensus values of fault width and geologic slip rate. Surface creep rate observations from 111 short-range alignment and trilateration arrays and 48 near-fault, Global Positioning System station pairs are used to estimate depth of creep, assuming an elastic half-space model and adjusting depth of creep iteratively by trial and error to match the creep observations along fault sections. Fault sections are delineated either by geometric discontinuities between them or by distinctly different creeping behaviors. We remove transient rate changes associated with five large (M ≥ 5:5) regional earthquakes. Estimates of fraction locked, the ratio of moment accumulation rate to loading rate, on each section of the fault system provide a uniform means to inform source parameters relevant to seismic-hazard assessment. From its mean creep rates, we infer the main branch (the San Andreas fault) ranges from only 20% 10% locked on its central creeping section to 99%-100% on the north coast. From mean accumulation rates, we infer that four urban faults appear to have accumulated enough seismic moment to produce major earthquakes: the northern Calaveras (M 6.8), Hayward (M 6.8), Rodgers Creek (M 7.1), and Green Valley (M 7.1). The latter three faults are nearing or past their mean recurrence interval. Online Material: High-resolution fault system map, and tables of creep observations and geographic coordinates of fault model.

Long‐Term Afterslip of the 2004M 6.0 Parkfield, California, Earthquake—Implications for Forecasting Amount and Duration of Afterslip on Other Major Creeping Faults

“…1, upper inset;Lienkaemper et al, 2014;McFarland et al, 2016). Our Parkfield afterslip observations were made on seven key alignment arrays across the main trace of the SAF, recovered from among the 19 arrays established by the U.S. Geological Survey in the Parkfield region, each selected for site and monument stability and for covering the full fault-zone width ( Fig.…”

“…We speculate that, like the Parkfield fault section, the least locked of these faults (Hayward and Green Valley), which also release a major portion (∼40%) of their long-term slip through interseismic creep, might have longer afterslip duration than the more locked faults. Conversely, because it is mostly locked with shallow (∼2 km depth, Lienkaemper et al, 2014) interseismic creep, surface slip in the Rodgers Creek fault ruptures may tend to be dominantly coseismic (Marone et al, 1991) with relatively rapid (i.e., short duration) afterslip, like the 2014 Napa case. However, surface ruptures in a section northward of Santa Rosa will propagate through intermediate (0.5-2 km) depth basin deposits, and thus may locally increase the proportion of postseismic to coseismic slip.…”

Interseismic coupling and refined earthquake potential on the Hayward‐Calaveras fault zone

“…The HCFZ considered here encompasses the HF, the northern CF, and the central CF from the junction with the northern CF to Gilroy (Figure ). The 2 to 8 mm/yr range of creep rates along the HF is well documented by over 30 alignment arrays (AAs) and 5 creepmeters covering the 60 km distance between Point Pinole and Fremont [ Lienkaemper et al , ; Bilham et al , ]. Additionally, the distribution of locking and shallow slip due to creep in space and time has been studied along the urban sections of the HF relying on inversion of InSAR (interferometric synthetic aperture radar) data [ Schmidt et al , ; Shirzaei and Bürgmann , ].…”

“…Creep on the north and central CF is not as well constrained because the fault traces largely lie in vegetated and hilly terrain, limiting remote sensing capabilities, and only nine AAs exist on this fault over a length of ~100 km. The AAs suggest slow and episodic (<2 mm/yr) creep on the northern section of the northern CF, increasing to steady creep at ~4 mm/yr near the transition with the central CF [ Lienkaemper et al , ]. Small aperture networks suggest creep rates between 3 to 7 mm/yr in the northern section of the central CF [ Prescott and Lisowski , ] and AAs suggest creep rates of ~10 mm/yr to the south [ Lienkaemper et al , ].…”

“…The AAs suggest slow and episodic (<2 mm/yr) creep on the northern section of the northern CF, increasing to steady creep at ~4 mm/yr near the transition with the central CF [ Lienkaemper et al , ]. Small aperture networks suggest creep rates between 3 to 7 mm/yr in the northern section of the central CF [ Prescott and Lisowski , ] and AAs suggest creep rates of ~10 mm/yr to the south [ Lienkaemper et al , ]. We aim to refine the along‐HCFZ creep rate variability by using a large InSAR data set and a modified time series method directly accounting for the spatial coherence [ Chaussard et al , ], allowing for creep estimations not only in urban areas [ Schmidt et al , ; Shirzaei and Bürgmann , ] but also on vegetated hillsides.…”