1. Simulated Deformations of Seattle High-Rise Buildings from a Hypothetical Giant Cascadian Earthquake Tom Heaton Jing Yang (Ph.D. thesis) Caltech Earthquake Engineering Research Laboratory Funded by NSF and the Caltech Tectonics Observatory Thesis available at heaton.caltech.edu

2. M 9+ with an average repeat ≈ 400 years Deep basin that amplifies long periods No strong motion records, no high-rises near a M 9 Think Anchorage Alaska in 1964, but with high-rises PSHA only hides the key issues

3. Empirical Green’s Functions as a method to extrapolate to very large magnitudes Repeat of the giant (M>9) Cascadia earthquake of 1700 Simulate rock ground motions with 2003 Tokachi-Oki M8.3 rock records as empirical Green’s functions Include effect of the Seattle basin by a transfer function derived from teleseismic S-waves transect (Pratt and Brocher, 2006) Wide modelNarrow modelMedium model

4. Ground Motion Recordings of the M 8.3 Tokachi-oki earthquake Many strong motion records Site conditions are known Largest well recorded subduction earthquake

5. Giant Cascadian earthquakes assumed to be similar to Sumatra (2004, M9.2) Tokachi-Oki (2003 M8.2) provides Green’s functions for simulating teleseismic records of Sumatra (2004 M 9.2) and strong motion records for Sumatra and Cascadia

6. Simulation of teleseismic records of Sumatra Two models differ in the down-dip width of the rupture

7. Average P-wave Spectra for the Largest subduction Earthquakes in the past 20 years Sumatra and Tokachi-Oki have similar spectral envelopes so Tokachi-Oki might be an acceptable source of empirical Green’s functions

8. Teleseismic Tokachi-Oki P-waves Use records that are median in amplitude

9. Simulated teleseismic P- waves compared with typical Sumatran record Focus on models that best match in the 1 to 10 sec period band YSS is a median P- waveform for Sumatra All records are on the same scale Tiling Sumatra with 20 Tokachi-Oki earthquakes seems to work acceptably

10. Cascadia Rupture Models Coastal co-seismic subsidence in 1700 suggest narrow model Wide model extends down dip to approximate location of the boundary with slow slip events Wide modelNarrow modelMedium model

11. Seattle Basin transfer function for teleseismic S-waves Based on a study by Pratt and Brocher (2006)

12. Simulated rock and basin ground motions for medium rupture

13. John Hall’s design of a 20-story steel MRF building Building U20 1994 UBC zone4 Stiff soil, 3.5 sec. period Building J20 1992 Japan code 3.05 sec period Both designs consider Perfect welds Brittle welds

14. Pushover Analysis Special attention to P- delta instability Story mechanism collapse Frame 2-D fiber-element code of Hall (1997)

15. Roof Displacement U-20 Brittle welds

16. Roof Displacement U-20 Perfect welds

17. Model Name RockSoil C-Wide-23C-Med-15C-Narrow-13C-Wide-23C-Med-15C-Narrow-13 DirectionEWNSEWNSEWNSEWNSEWNSEWNS PGA cm/s 2 max42841615313488102666742358357311263 med2672301531615077419578172203131226 min208181475135382763001031524966 PGV cm/s max607839 38227 103222127131 med55432114201812129052844882 min32242016569210341541825 U20B IDR (%) max3.02.32.12.62.0 CO med1.62.30.60.40.60.4CO 1.5CO2.3CO min0.60.7 0.40.1 2.1CO2.1CO0.51.3 U20P IDR (%) max1.01.20.61.00.81.4CO 3.0CO med1.41.80.40.30.40.3CO 1.22.41.22.9 min0.40.50.40.30.1 1.51.71.02.20.40.6 J20B IDR (%) max2.82.11.12.20.92.5CO med2.32.60.40.30.50.3CO 1.9CO2.14.4 min0.90.50.40.20.1 1.6CO2.44.30.70.3 J20P IDR (%) max1.01.10.5 0.7CO 6.25.0CO med1.01.20.40.30.40.34.2CO1.11.21.00.9 min0.50.4 0.20.1 1.91.21.11.00.50.3 U6B IDR (%) max1.82.01.31.10.90.7CO med1.61.20.30.2 0.4CO 2.13.51.73.5 min1.51.00.20.30.1 3.94.71.33.40.30.6 U6P IDR (%) max1.62.11.00.8 0.5CO 4.7CO med1.0 0.30.2 0.4CO 1.31.91.11.8 min1.10.90.20.30.1 4.64.71.21.80.30.5 J6B IDR (%) max1.11.50.70.50.40.3CO 5.4CO med0.80.60.2 0.10.2CO 0.81.90.61.9 min0.60.30.10.20.1 3.93.50.72.00.30.4 J6P IDR (%) max0.91.40.60.50.40.3CO 3.13.82.32.4 med0.60.50.2 0.10.22.4CO0.8 0.71.2 min0.50.30.10.20.1 2.4 0.71.10.30.4 Table 6.3: PGA and PGV of simulated strong ground motions in station SEA and performance of 20- and 6- story buildings shaken by these motions.

18. Conclusions Presence of brittle welds significantly degrades performance (2-8 times more likely to collapse) Amplification by the Seattle basin could be a very serious issue Reliable prediction of down dip limit of rupture is a critical issue Simulating building behavior for a hundred cycles of yielding/damaging motion is a critical issue Cannot conclude that Seattle high-rises will perform acceptably in a giant earthquake Don’t hide this issue in PSHA!