- Ground-motion model is: where Y is in cm∕s2, c1 = 0.447, c2 = 0.597, c3 = 0.00335, c4 = -0.00330, c5 = -0.334, c6 = -0.373,
$a1 = -6.73 × 10-5 and a2 = 2.09 × 10-2 [taken from Kanno et al. (2006)], σν = 0.206 (inter-event),
σϵ = 0.262 (intra-event), σT = 0.333 (total) and σc = 0.058 (component). Take G from Uchiyama and
- Use V s,30, sometimes estimated using empirical relation between V s,20 and V s,30, to characterise sites.
Distribution shows peak at about 300m∕s with few records > 800m∕s.
- Use data from KiK-Net (only surface instruments) and K-Net from before 08/2008. Use these networks
because of their relatively high spatial densities, detailed site information and wide variety of source types
(e.g. crustal and subduction) and sizes.
- Focal depth, roughly: 0 ≤ H ≤ 145km. No clear correlation between Mw and H.
- To focus on motions for engineering applications, select data with: Mw ≥ 5.5, H < 200km, epicentres
within 30∘–46∘N and 130∘–148∘E, M-rrup cut-off defined by Kanno et al. (2006) to ensure sufficient
dynamic range and 100 ≤ V s,30 ≤ 1000m∕s. Also generally eliminate events with < 5 records.
- Develop model for to study spatial correlation of ground-motion parameters. Similar to studies of Hong and
Goda (2007) and Goda and Hong (2008) (see Section 2.281). Spatial-correlation models developed using
inter- and intra-event residuals w.r.t. ground-motion model. This aspect of the model is not summarised
- Baseline correct records and apply 4th-order acausal Butterworth low-cut (fc = 0.05Hz) filter after padding
- Use correction function A, where Rtr is shortest distance from site to Kuril and Izu-Bonin trenches,
of Kanno et al. (2006) to model anomalous ground motions in NE Japan from intermediate and deep
earthquakes occurring in the Pacific plate due to unique Q structure beneath the island arc.
- Modify geometric decay term to model steeper decay for rrup < 75km and make additional adjustments
to reduce biases seen in preliminary models and to achieve homoscedasticity in residuals, particularly for
large H and small rrup.
- State that model should not be extrapolated to M < 5.5, M > 8 and rrup < 10km, where data scarce or