The current state of the practice in earthquake ground motion modeling, to carry out probabilistic seismic hazard analysis for a specific region, is to develop ground-motion models (GMMs) based on combined data from similar tectonic environments worldwide. The main reason for combining the regional and international data is the limited number of ground-motion recordings in the region of interest. In this process, there is no difference between temporal and spatial variabilities. Thus, the aleatory variability of a GMM is assumed to apply to any location within the tectonic category (Lavrentiadis, et al., this issue of BLEE). Moreover, the variability coming from a GMM is assumed to be aleatory. In ground-motion modeling, this is generally called the ergodic assumption.

With a rapid reduction of the cost of ground motion instrumentation and consequently an exponential growth in the number of installed instruments around the world, as well with the access to powerful ground motion simulations, researchers can now investigate the differences between the local characteristics of ground motions relative to models that were developed globally. The variability components (aleatory and epistemic) can also be better assessed. This is carried out by developing partially or fully non-ergodic GMMs, an emerging and promising research topic in seismic hazard analysis.

This Special Issue focuses on a collection of papers on non-ergodic GMM modeling and its applications. The Special Issue starts with a comprehensive overview paper by Lavrentiadis, et al., followed by four papers on methodology and analysis issues, followed by five papers on applications of non-ergodic modeling to various parts of the world.

We would like to thank the Bulletin of Earthquake Engineering, especially Prof. Atilla Ansal, for the cooperation in publishing this Special Issue. The efforts and cooperation of all authors of the papers are gratefully appreciated.

Yousef Bozorgnia, Christine Goulet, Jorge Macedo.

Guest Editors.