Comment on “Total Probability Theorem Versus Shakeability: A Comparison between Two Seismic‐Hazard Approaches Used in Central Asia” by D. Bindi and S. Parolai: Table 1

Abstract

The recent paper of D. Bindi and S. Parolai (Bindi and Parolai, 2015; hereafter referred to as BP15) considers discrepancies between variants of seismic‐hazard estimates for central Asia. BP15 notes that for this territory the levels of hazard estimated in Ulomov et al. (1999; hereafter referred to as Ul99) during the construction of the Global Seismic Hazard Assessment Program (GSHAP) map shows a clear mismatch when compared with similar maps for neighboring areas: hazard estimates of Ul99 are systematically larger. Indeed, when Zhang et al. (1999) who were responsible for aggregating probabilistic seismic‐hazard assessment (PSHA) maps for various territories of Asia into the single final map tried to combine the component maps, the only way to incorporate the results of Ul99 was to decrease its peak ground acceleration (PGA) values by ∼30%. BP15 discusses the assumption that this problem could be caused by deficiencies of Riznichenko’s hazard estimation procedure, believed to be used in the calculating of Ul99 for GSHAP. The first deficiency under discussion is the implicit assumption of Riznichenko (1965) that the level of shaking, when considered as a function of magnitude M and distance r , can be treated in deterministic style, with no allowance for the scatter of individual observations with respect to an assumed mean relationship. Originally, Riznichenko expressed the level of shaking in terms of macroseismic intensity I , but most of the further discussion is applicable to cases when the level of shaking is expressed in any amplitude measure, further generically denoted A . The need to account for the scatter of I or A is evident (see e.g., Keilis‐Borok et al. , 1973). BP15 comes to the conclusion that their initial guess is untrue: the possible assumption of zero …