Campbell-Bozorgnia NGA Ground Motion Relations for
the Geometric Mean Horizontal Component of Peak
and Spectral Ground Motion Parameters

Kenneth W. Campbell
EQECAT, Inc.
Beaverton, Oregon
and
Yousef Bozorgnia
Pacific Earthquake Engineering Research Center
University of California, Berkeley

PEER 2007/02
MAY 2007

Abstract
We present a new empirical ground motion model, commonly referred to as an attenuation relationship, which we developed as
part of the PEER Next Generation Attenuation of Ground Motion (NGA) Project. Using a common database of worldwide strong
motion recordings, we selected a subset of ground motion data and predictor variables that we believed were appropriate for use in
developing our model. Consistent with the requirements of the PEER NGA Project, we developed both a median and aleatory
uncertainty model for peak ground acceleration (PGA), peak ground velocity (PGV), peak ground displacement (PGD), and
response spectral acceleration (PSA) and displacement (SD) for oscillator periods ranging from 0.01–10.0 s, magnitudes ranging
from 4.0–8.0, and distances ranging from 0–200 km. We consider these models to be valid for use in the western United States and
in other similar tectonically active regions of shallow crustal faulting worldwide. A comparison of our NGA model with our
previous ground motion models (Campbell, 1997; Campbell and Bozorgnia, 1994, 2003) showed that the biggest differences in
these models occur for sites located at small-to-moderate distances from large-magnitude earthquakes or near reverse faults with
surface rupture, where the NGA model predicts lower ground motion, and for sites located on the hanging wall of dipping strikeslip
and normal faults, where the NGA model predicts higher ground motion. We also found that the standard deviation is no
longer a direct function of magnitude, which increases aleatory uncertainty for large-magnitude earthquakes and decreases it for
small-magnitude earthquakes for stiff sites, compared to our previous models. However, the dependence of the standard deviation
on nonlinear site effects in our new model can lead to less aleatory uncertainty for soft sites even at large magnitudes as compared
to our previous models.

Key Words
NGA, ground motion,
attenuation, near source, response
spectra, site amplification,
hanging wall, basin response