2.450 Shoushtari et al. (2018)

• Ground-motion model is (chosen after testing various functional forms):

  where Y is in cm∕s2, a = 0.4683, b = -0.002159, c = 0, d = 0.6524, S_B = -0.2571, S_C = 0.1464, S_D = 0.3654, S_E = 0.3428 and σ = 0.387.

• Use 4 NEHRP site classes:

  B

  Rock, 760 < V _s,30 ≤ 1500m∕s. 11 records from Malaysia. Use S_B term.

  C

  Hard soil/soft rock, 360 < V _s,30 ≤ 760m∕s. Use S_C term.

  D

  Medium soil, 180 < V _s,30 ≤ 360m∕s. Use S_D term.

  E

  Soft soil, V _s,30 < 180m∕s. 5 records from Malaysia. Use S_E term.

  Classification for Malaysian stations done using geological description.

• Develop model for use in assessing hazard from distant Sumatran earthquakes. Because available data from this region is sparse also use data from Japan, which believe is tectonically similar. Data from 16 stations in Malaysia.
• Partner model to Shoushtari et al. (2016) (see Section 2.413) for intraslab events.
• Consider events with thrust mechanisms on shallow dipping planes with focal depths ≤ 50km as interface.
• Focal depths between 12 and 50km.
• Process records using acausal 4th-order bandpass Butterworth filter after zero-padding and cosine tapering.
• Data roughly uniformly distributed w.r.t. distance and magnitude.
• Examine residuals w.r.t. magnitude and distance and find no trends.
• Compare predictions and observations (including one from Singapore, which is not used to derive model) binned by magnitude interval. Find good match.
• State that in future would recommend using two-stage or maximum-likelihood regression techniques.