Abstract
The uncertainty and unreliability in the probabilistic seismic demand models (PSDMs) can be addressed in the framework of performance‐based earthquake engineering (PBEE). Some characteristics have been introduced and extensively investigated in the earthquake engineering literature for appropriateness (i.e., optimality) of various intensity measures, such as efficiency, effectiveness, sufficiency, practicality, and hazard computability. The reliability of the intensity measures can also be considered as another important criterion for intensity measure (IM) selection, which has not been investigated extensively. In this paper, a framework has been developed and implemented to investigate the reliability of various intensity measures alongside the number of adopted earthquake records. In this regard, the inverse of the mean coefficient of variation (IMCV), which is detected to be in a direct and linear relationship with the number of earthquake records, is adopted as the main criterion to compare the reliability of ground motion intensity measures. A simple relation has been proposed and calibrated for the studied IMs that provide the required number of earthquake records to satisfy an intended level of reliability. As a case study, a significant number of recorded ground motions at Zagros and non‐Zagros zones (two major seismic regions in Iran) with rock and soil conditions are selected, and the most reliable ground motion intensity measure is introduced at various seismic regions and soil types by exploiting the developed framework and the proposed relationship.