THE 2015 GORKHA EARTHQUAKE SEQUENCE: STRONG GROUND MOTION FEATURES

The 2015 Gorkha Earthquake and its aftershocks caused a major disaster in Nepal causing widespread casualties, injuries, and physical damage. Ground motion from the mainshock and the largest aftershocks recorded at Kathmandu Valley are studied in detail in this contribution. Horizontal ground motion due to the mainshock has a very peculiar energy content, with spectral ordinates as high as 0.5g at a period of about 5s. while peak ground acceleration is relative low at around 16% of acceleration of gravity. This implies an amplification factor of about 3 at a vibration period of 5s, which is 10 times larger than that implied by modern design codes such as the Eurocode 8. Spectral shapes of horizontal and vertical motion due to the mainshock and the aftershocks are very poorly represented by those recommended in modern design codes as well as uniform hazard spectral shapes obtained from standard probabilistic seismic hazard assessment. Although peak ground acceleration is slightly larger in the vertical motion than in the horizontal motion, peak ground velocity is much larger in horizontal motion. This indicates higher frequency content in vertical motion than in the horizontal motion, partly due to possibly the effects of soft and deep sediments of the valley which damped out high-frequency motion but amplified the lower frequencies. Although ground motion is available until now at only one station in Kathmandu Valley, careful examination of the available data provides a lot of insight into its nature. The most important insight being the fact that standard design spectra recommended in modern codes greatly underestimate spectral ordinates at periods of vibration longer than about 0.5s. This has a serious consequence that use of such spectral shapes for design of high-rise buildings in the Valley may result in serious under-estimation of expected seismic action. It is also apparent that uniform hazard spectra based on ground motion prediction models calibrated from data recorded elsewhere are not applicable in the Valley. These observations necessitate development of suitable ground motion models as well as geotechnical site response models. The results indicate that high-rise (more than about 5 storeys tall) buildings in the Valley should be very carefully designed. Until proper ground-motion models are calibrated a lot of uncertainties exists in the seismic action expected in such structures during future earthquakes.