Ground Motions during the 2015 Gorkha, Nepal, Earthquake: An Expected Event that Defied Expectations

Thursday, 17 December 2015: 10:20
305 (Moscone South)
Susan Elizabeth Hough1, Stacey Servito Martin2, Amod Mani Dixit3, Surya Shrestha4, Ramesh Guragain4, Elizabeth S Cochran3, Danielle F Sumy5, Adam T Ringler6, Daniel E McNamara7, Domniki Asimaki8, Roger G Bilham9, David Mencin10, John Galetzka10, James H Luetgert11, Lingsen Meng12, Jean-Paul Ampuero8 and Sudhir Rajaure13, (1)USGS, Pasadena, CA, United States, (2)Earth Observatory of Singapore, Singapore, Singapore, (3)Organization Not Listed, Washington, DC, United States, (4)NSET, Kathmandu, Nepal, (5)Incorporated Research Institutions for Seismology, Seattle, WA, United States, (6)USGS, Albuquerque, NM, United States, (7)USGS National Earthquake Information Center Golden, Golden, CO, United States, (8)California Institute of Technology, Pasadena, CA, United States, (9)University of Colorado, CIRES,, Boulder, CO, United States, (10)UNAVCO, Inc. Boulder, Boulder, CO, United States, (11)USGS California Water Science Center Menlo Park, Menlo Park, CA, United States, (12)University of California Los Angeles, Los Angeles, CA, United States, (13)Tribuvan University, Kathmandu, Nepal
Earthquakes with magnitudes close to, and exceeding, Mw8 have long been expected along the Himalayan arc. A repeat of the 1934 Bihar-Nepal earthquake was expected to cause heavy damage, with as many as 40,000 fatalities in Kathmandu Valley. The 2015 Gorkha earthquake was smaller than the 1934 event, but unlike the 1934 earthquake, involved rupture of the segment of the Main Central Thrust directly beneath the valley. Despite the large magnitude and proximity to the valley, the damage was lower than expected. Throughout Kathmandu Valley as well as the near-field region, European Macroseismic Scale intensities exceeded 8 in only rare instances. The extent of landsliding and liquefaction was also lower than had been expected (Collins and Jibson, 2015). Strong motion data from one conventional (NetQuakes) instrument, several low-cost MEMS accelerometers, and high-rate GPS provide insights into the extent to which mainshock and aftershock ground motions were controlled by source, path, and site effects. Mainshock horizontal peak accelerations recorded in central Kathmandu were ≈0.16g, with shaking strongly peaked near 5 s. The long period character of the mainshock can be attributed in part to source properties (Avouac et al., 2015). The expected basin response resonance (≈1-3-s period) is, however, also absent in the mainshock coda, but present in the coda of large aftershocks. A progressive shift in predominant period is also observed over ≈10 minutes following the mainshock. We conclude that shallow soft sediments underlying Kathmandu Valley responded nonlinearly during the mainshock, such that the predominant period of amplification was shifted to longer periods than the weak-motion resonance, and high frequencies were deamplified. Mainshock ground motions, controlled by site and well as source effects, were therefore concentrated at periods that were too long to significantly impact vernacular structures in Kathmandu Valley, most of which are between 3-10 stories.