Roger Bilham

Our collaborative investigations with Kandilli Observatory, Istanbul, involve measurements of strain and aseismic slip on the northern and eastern Anatolian faults to learn more about the relationship between earthquakes and fault creep. In December 2019 the largest creep event (10 mm) yet recorded on the northern Anatolian fault propagated eastward near Ismetpasa in seven sub-events with velocities of up to 2.3 m/s.  Surface after-slip in the Mw6.8 2020 Sivrice/Elazig earthquake on the eastern Anatolian fault commenced a few days after the mainshock and decayed with a time constant of approximately 20 days.  Unlike in previous afterslip episodes elsewhere, however, slip rates accelerated after local aftershocks and continues at low rates a year after the event. Slip was also triggered by this earthquake on the East Anatolian fault some 40 km east of the rupture zone initiating a creep event that propagated eastward at 0.26 m/s. Slip was triggered on the southern San Andreas fault following the 2019 Ridgecrest earthquakes and propagated southward eventually triggering minor seismicity near the Salton Sea.

As part of an investigation to document coseismic slip during a major earthquake, we have upgraded Hayward fault creepmeters to capture coseismic rupture in a pending earthquake at 1 Hz sample rates. The Hayward fault creepmeters have been found to act as strain-meters in that they record the harmonics of tidal loading in San Francisco Bay.  Gas was vented from the East Anatolian fault following the Sivrice earthquake and experiments to measure potential CO2, H2, CH4 and 222Rn venting during creep events on the northern and southern San Andreas faults have been implemented.  The introduction of multiple axis extensometers across creeping faults is providing surprising insight into their behavior.  In particular creeping faults appear to dilate several mm during aseismic slip in the dry season and to close rapidly during heavy precipitation, a behavior we suspect is attributable to the rotation and lubrication of lozenge-shaped clay clasts in the fault zone.

In a collaborative project with NED University, Karachi, we have investigated Medieval and pre-historic flooding of the Kashmir Valley.  Quite small landslides in the Pir Panjal mountains can dam the river Jhelum, causing the entire valley to flood.  We calculate that breach of a 3 year transient dam released a 14 cubic km catastrophic flood that drained the valley in the ninth century transporting boulders with diameters exceeding 12 m, similar to those observed in the stream bed near Muzafferabad.  Were the river dammed for a century, as occurred in the Pamir landslide that formed Lake Sarez, 400 cubic km of water could be stored in the Kashmir valley flooding Srinagar to a depth of 170 m.  Although ancient shorelines at this level exist in the Kashmir valley, the absence of mega-boulders much larger than 13 m in diameter donstream suggests that floods with volumes >15 cubic km probably have not occurred, and these shorelines were shifted to their current levels by tectonic processes.