- Ground-motion model is that of Campbell and Bozorgnia (2008b) (see Section 2.275).
- Use same data as Campbell and Bozorgnia (2008b) (see Section 2.275).
- Modify the random-effects regression method of Abrahamson and Youngs (1992) to account for spatial
correlation defined by a pre-defined empirical model dependent on separation distance or derived during
the regression analysis. Prefer the use of a pre-defined empirical model for various reasons.
- To provide baseline model for comparison, refit model of Campbell and Bozorgnia (2008b) using
random-effects regression ignoring spatial correlation. Find minor differences with reported coefficients of
Campbell and Bozorgnia (2008b), which relate to manual coefficient smoothing.
- Find intra-event σ increases and inter-event σ decreases but total σ remains roughly the same when
spatial correlation is accounted for. Provide theoretical justification for difference in σs if spatial correlation
between records is considered or not.
- Do not report coefficients, only provide graphs of σs.
- State that, because regression coefficients are not significant different if spatial correlation is accounted
for, the regression procedure can be simplified.
- Discuss the implications of findings on risk assessments of spatially-distributed systems.