- Ground-motion model is:
where A is in cm∕s2, a = 0.59, c = -0.003, d = 0.005, e = 0.5, c = 0.01, p = -0.209, V ref = 472.8 and σ =
- Use V s,30, estimated from shallower depths if required, to characterise sites. Derive site term based on
residuals of model without this term. Distribution by NEHRP classes: A (8 records), B (114), C (253), D
(83) and E (9).
- Note that 24 broadband instruments are deployed in Vietnam by Institute of Geophysics (Vietnam) and
Institute of Earth Sciences (Academia Sinica, Taiwan) since 2005 and some records from 2001 (Dien Bien
M5.3 and aftershocks) and 5 other events 4.0–4.9. Note that limited data from Vietnam, particularly from
moderate and large earthquakes, (1 mainshock and 16 aftershocks, 17 records, 2.7 ≤ Mw ≤ 5.1) and hence
supplement it with data from Yunnan (China, 1 earthquake, 4 records, Mw6.4), which is adjacent and has
similar tectonics, and particularly Japan (K-Net, 5 events, 3.9 ≤ Mw ≤ 6.9).
- Select records following these criteria: strike-slip event, inland-type earthquake, 3.5 ≤ Mw ≤ 7.0 (the
largest event thought to be possible in northern Vietnam), ground-surface records and from source distance
less than previously-defined M-dependent cut-off.
- Only use data from strike-slip earthquakes because dominant mechanism in northern Vietnam.
- Vast majority of data is from Mw ≥ 5.8.
- Also derive models based on only data from Japan and only data from Vietnam and Yunnan. Find similar
coefficients and σs.
- Compare predicted and observed PGAs and find good fit and no apparent trends w.r.t. region, for example.
- Also compare predictions and observations for 2007 Ninger (Yunnan, China, Mw6.4) earthquake and find
fairly good match.