Afterslip Behavior Following the M6.0, 2014 South Napa Earthquake with Implications for Afterslip Forecasting on Other Seismogenic Faults.

Abstract:

Surface slip on the 15-km-long 2014 Napa rupture occurred mostly as afterslip on its southern 8-9 km, but mostly coseismically on its northern section, with maximum coseismic slip reaching 0.4-0.5 m. Urban post-earthquake recovery demands rapid and realistic answers to: How much more afterslip can occur? How long will it last? About 69% of all surface slip was postseismic. Cumulative slip approaches 0.3-0.5 m in the middle of the rupture, estimated with program AFTER (Boatwright et al., 1989). At 1-yr post-earthquake, slip should be 96-99% complete on the southern section. However, 0.8 km north of the coseismic maximum (site NLOD), slip is ≤92% complete. Model estimates of total fault displacement (Uf) improved in the weeks following the earthquake. By 3 weeks, uncertainties were ≤±5 cm; by 3 months, 3 sites had stable Uf values and ≤±1 cm uncertainties, except site NLOD, where at 9 months its formal Uf-uncertainty was ±2 cm, but the model is uncertain. To estimate the impact of afterslip following a large future earthquake expected on the Hayward fault in the densely populated San Francisco East Bay, we are analyzing how rapidly we can resolve final displacement (Uf) from other records, e.g., the M6.0, 2004 Parkfield earthquake. Napa uncertainties resolve faster than Parkfield’s by a few months. Although 1-yr solutions are shown to be accurate for the central Parkfield rupture, at the rupture tips, afterslip rates, and thus Uf estimates, increased considerably in later years. Unstable Uf values at NLOD resemble behavior at the Parkfield rupture tips. Unlike Napa, Parkfield afterslip was only ~74% complete after 1 year,~91% after 6 years, and ultimately taking 8-11 yr to reach interseismic creep rates. The Napa and Parkfield afterslip behaviors represent end members, locked-versus-creeping fault sections. Though we might expect the creeping Hayward Fault to behave more like Parkfield, its much larger expected magnitude (M6.8) could make an important difference.