- Ground-motion model is:
_{1}= 0.14, c_{2}= -6.25, c_{3}= 0.37, c_{4}= 2.237, c_{5}= -7.542, c_{6}= -0.125, c_{7}= 1.19, c_{8}= -6.15, c_{9}= 0.6, c_{10}= 0.345, b_{v}= -0.24, V_{A}= 484.5; τ = 0.435 (inter-event), ϕ = 0.508 (intra-event) and σ = 0.669 (total) for homoscedastic variability; and τ(M) = s_{1}for M ≤ 7.1, τ(M) = s_{1}+(s_{2}-s_{1})(M_{w}-7.1) for 7.1 < M_{w}< 7.5 and τ(M) = s_{2}for M_{w}≥ 7.5 where s_{1}= 0.28 and s_{2}= 0.04, τ(M,R) = τ(M) + r_{1}for R ≤ 100km, τ(M,R) = τ(M) + (r_{1}- r_{2})(R - 100) for 100 < R < 130km and τ(M,R) = τ(M) + r_{2}for R ≥ 130km where r_{1}= 0.30 and r_{2}= 0.52 for heteroscastic inter-event variability. - Characterise sites using V
_{s,30}. State that valid between 200 and 1300m∕s. - Use B
_{depth}(depth to 1.5km∕s shear-wave velocity isosurface), which ranges from about 100m to about 2400m. Only 353 records have estimates of B_{depth}. Believe model applicable for B_{depth}between 0 and 10km. - Use 3 styles of faulting:
- 1.
- Strike-slip (1120 records) and normal (13 records). F = 1.0.
- 2.
- Reverse (1450 records). F = 1.28.
- 3.
- Reverse/strike-slip oblique. F = 1.14.

- Update of model of Graizer and Kalkan (2007, 2008) (see Section 2.279) to include anelastic attenuation
term based on Q
_{0}, the regional quality factor, and a frequency-dependent sedimentary-basin scaling term based on B_{depth}. Use same database as Graizer and Kalkan (2007, 2008). - Choose c
_{10}= 0.345 based on average Q_{0}= 150 for California. Recommend using model with Q_{0}determined from Lg or coda waves for region of interest. Believe model applicable for Q_{0}≤ 250. - Constrain coefficients of basin-effect terms using data from 1999 M
_{w}7.1 Hector Mine, 1992 M_{w}7.3 Landers and 1989 M_{w}6.9 Loma Prieta earthquakes. - Analyse inter- and intra-event residuals using best-fit lines against various independent parameters and find no trends or large offsets.
- Provide model for pseudo-spectral accelerations. This is not summarised hear because it is not in the style of coefficients for many periods but as a closed-form function.
- Define heteroscastic variability models by dividing standard deviations into 8 magnitude and 10 distance bins.
- Compare observations and predictions for 13 Californian earthquakes and find a good match.