Monte Carlo seismic response analysis of permafrost sites: a case study

The horizontal seismic responses in cold regions have not received adequate attentions. In seismic design for frozen ground, computations for unfrozen ground are generally adopted directly. This study uses the Monte Carlo method to verify the applicability of traditional seismic computations in cold regions. Xidatan valley, Kekexili mountain, Chumaerhe plain, and Beiluhe basin along the Qinghai–Tibet railway are selected as the study regions; a series of random soil profiles are generated. Equivalent linear analysis results show that the current Chinese code provides mismatched spectral site amplification estimations at rare seismic intensity level. These estimations can practically be updated by increasing the current seismic motion parameters along the Qinghai–Tibet railway or by amplifying the existing design spectrum with a coefficient. The melting of permafrost layer would increase spectral peak; 70% of the original thickness would result in spectral peak amplification of 120%. Ground temperature model and permafrost dynamic nonlinear performance should be further explored.