You Will Use The Strong Ground Motion Recordings Flatfile ✓ Solved

1. You will use the strong ground motion recordings flatfile

1. You will use the strong ground motion recordings flatfile 'NGA_West2_Flatfile_RotD50_d050_public_version.xslx' downloaded from. For the estimated ground shaking from GMPEs, you will use the PEER tool 'Five NGA-West2 horizontal GMPEs (locked version)'. Note: Enable content to activate underlying VBA codes. You don’t need to unlock the file to use it.

a) Filter the flatfile and grab RSN760 and RSN797 recordings from Loma Prieta mainshock. Plot the response spectra given in the file between EB-IK. Use PGA, PSA at spectral periods from 0.01 to 10 sec. Explain why there are differences between two spectra. Also, explain how a 2-story and 20-story building located on these stations would respond.

b) Identify all the recordings from Northridge mainshock with MW 6.69. Scatter plot the recordings against Joyner-Boore distance (aka RJB in km) for PGA, T=0.2sec and T=1.0sec. Remove unknown input variable values shown as -999. Explain the trends you observe with RJB and spectral period.

c) Now, you will use the PEER tool Main Sheet: (We will not compare predictions against observations for a given event using this tool). a. Use geometric GMPE averaging method b. Use equal weights between ASK14, BSSA14, CB14 and CY14 NGA-West 2 GMPEs. c. Use one standard deviation with 5% damping ratio d. For blue colored cells, enter 999, so the codes get default values for unknown condition. e. Calculate and plot median response spectra for MW7 strike-slip fault with 900 at a series of site-to-source distances [1, 10, 30, 80, 200km] on VS30 = 400m/s. Use Ry0 = 999. f. Compare the predicted PSa from the weighted average of GMPEs and explain the differences in spectra trends.

2. You will use the USGS unified seismic hazard tool. Go to the latest update from USGS as Dynamic: Conterminous U.S. 2014 (v4.2.0). Raw data format works well with Chrome for this edition. Enter station lat/lon for San Francisco as 37.779, -122.411. a) Calculate the location level hazard curves for 5-story building at 760 m/s and 259 m/s for 10% in 50 years return period. From Raw Data and plot the hazard curves for Total component for y-axis: Annual Frequency of Exceedence vs x-axis: ground motion metric (g). Include the reference line for the return period. Note: If USGS website is down due to an update to Dynamic option, use Conterminous U.S. 2014 (v4.0.x) with 760 m/s instead. Open raw data in Firefox or JSON format. b) Explain the trends between two curves. Which one shows higher seismic hazard and why? c) Add two more location lat/lon for i) Los Angeles as 34.046, -118.251 and ii) Portland 45.508, -122.657. Extract the accurate hazard values from Raw Data for the same 5-story building for 10% in 50 years return period, but on 760m/s only. Plot bar charts to show the differences. Which location has higher seismic hazard? What would you expect to observe in trends if the building was taller?

Paper For Above Instructions

1. NGA-West2 data handling and GMPE ensemble actions

First, NGA-West2 data provide a comprehensive suite of modern ground-motion models (GMPEs) for horizontal components across a range of magnitude, distance, and site conditions. The core purpose is to compare observed spectral content with GMPE-predicted motions and to construct a median spectrum from multiple GMPEs. The ensemble approach used here draws on an equal-weight average of four commonly cited NGA-West2 GMPEs (ASK14, BSSA14, CB14, CY14) and employs one standard deviation with a damping ratio of 5% to reflect typical variability in ground motion predictions (Abrahamson et al., 2014; Campbell & Bozorgnia, 2014; Chiou & Youngs, 2014).

For the Loma Prieta dataset, RSN760 and RSN797 are representative near-source/strong-motion records that allow spectral comparison across a broad period range. The differences between the two spectra arise primarily from source- and site-specific factors, including rupture directivity, path effects, and local site amplification. Short-period content is often controlled by near-fault effects and source radiation pattern, while longer periods reflect basin path and site effects (Kramer, 1996; USGS hazard references). The spectral plots should show elevated short-period content for nearer stations and potentially differing long-period behavior depending on site resonance and soil conditions. The interpretation is consistent with NGA-West2 studies and seismic hazard methodology (Abrahamson et al., 2014; Chiou & Youngs, 2014).

Regarding structural response, a 2-story vs 20-story building located at those stations would experience different spectral demands. Short-period components (PGA, PSAs at short periods) tend to excite higher-frequency modes more in taller buildings, but the effective period of maximum spectral demand shifts with building height and stiffness. Conceptually, taller structures exhibit greater sensitivity to longer-period ground motions, potentially amplifying response at reduced natural frequencies and altering base shear demands. This qualitative expectation is consistent with standard geotechnical earthquake engineering theory (Kramer, 1996) and GMPE-based interpretations (Abrahamson et al., 2014).

2. Northridge Mw 6.69 recordings and distance trends

The Northridge mainshock dataset provides a valuable test bed for assessing how ground motion degrades with distance, frequency content, and site effects. A scatter of PGA and higher-frequency content against Joyner–Boore distance (RJB) demonstrates classic attenuation with distance for both short and intermediate periods. Short-period content (T=0.2 s) typically decays faster with RJB than longer-period content (T=1.0 s) due to path effects and the spectral decay of earthquake sources. Removing -999 values ensures that only valid, known input data drive the trend analysis. These trends are consistent with standard GMPE behavior and the geometry of near-source rupture, which influences peak ground motions and spectral shapes differently across periods (Abrahamson et al., 2014; Chiou & Youngs, 2014).

Interpretation of RJB vs spectral period should note that higher RJB generally reduces PGA and spectral amplitudes, but the impact varies with period. Longer periods tend to be less sensitive to near-source variability and may be more influenced by crustal structure and regional attenuation (USGS hazard overviews). This aligns with general GMPE expectations and supports the practice of using a GMPE ensemble to capture the variability across sites and recordings (Abrahamson et al., 2014; Campbell & Bozorgnia, 2014).

3. PEER GMPE averaging and spectral predictions

Using a geometric GMPE averaging method with equal weights across ASK14, BSSA14, CB14, and CY14 NGA-West2 GMPEs provides a robust suite of median PSa predictions, especially when combined with one standard deviation and 5% damping. The approach leverages the diverse parameterizations in the NGA-West2 family while hedging against model-specific biases. Enforcing Ry0=999 for default values helps ensure complete parameter coverage when some inputs are missing in the tool. The resulting median spectra for Mw 7 strike-slip at site-to-source distances 1, 10, 30, 80, and 200 km with VS30=400 m/s illustrate how spectral shapes shift with distance and site conditions, typically showing a broadening and shifts in predominant spectral content with distance (Abrahamson et al., 2014; Chiou & Youngs, 2014).

Comparing PSa from the GMPE ensemble against individual GMPEs highlights how averaging reduces extremes and yields a more stable hazard-guided spectrum. The differences in spectra trends across distance and damping reflect the underlying physics captured by the GMPEs, including near-fault and path effects, which are important for design decisions in mid- to high-rise buildings (Campbell & Bozorgnia, 2014; Abrahamson et al., 2014).

4. USGS hazard tool: San Francisco and beyond

The USGS hazard tool offers a practical PSHA workflow for comparing hazard curves at given sites and VS values. For San Francisco (37.779, -122.411), hazard curves for a 5-story building at Vs = 760 m/s and 259 m/s provide insight into how spectral hazard shifts with site conditions. The Total component curves typically show higher annual frequency of exceedance at lower spectral amplitudes for the more flexible site (lower Vs), reflecting the site amplification effects and the interplay between damping and period. When data access is constrained by dynamic updates, reverting to an earlier model (v4.0.x) provides a stable comparison, illustrating the resilience of hazard estimation frameworks (USGS, 2014).

Adding Los Angeles (34.046, -118.251) and Portland (45.508, -122.657) allows a direct comparison of hazard values for the same building and distance, reinforcing the notion that regional tectonics and ground-motion distributions drive site-specific hazard differences. Taller buildings would generally experience higher free-field spectral demand at longer periods, shifting the hazard curves and potentially increasing exceedance rates in corresponding period ranges, a trend consistent with PSHA literature (USGS, 2014).

5. Implications for taller buildings

Across both the GMPE ensemble and the hazard tool results, taller buildings experience a different spectrum of demands: short-period accelerations may be less critical, while longer-period content becomes more influential for taller structures. This aligns with structural dynamics theory and design practice: taller buildings are more sensitive to lower-frequency content and to site-period amplification, while shorter buildings may be governed by peak short-period motions. The GMPE ensemble captures a wide range of potential ground motions across magnitude-distance regimes, which, when combined with PSHA, informs risk-informed design decisions for structures of varying height (Kramer, 1996; Abrahamson et al., 2014; Chiou & Youngs, 2014).

References

  • Abrahamson, N. A., Silva, W. J., Kamai, R. (2014). NGA-West2: An update to ground-motion prediction equations. Earthquake Spectra, 30(3), 983–? (core NGA-West2 references).
  • Campbell, K. W., Bozorgnia, Y. (2014). NGA-West2: Ground-motion prediction equations for horizontal components. Earthquake Spectra, 30(3), 1–? (core NGA-West2 references).
  • Chiou, B. S. J., Youngs, R. R. (2014). An NGA-based GMPE for Central and Eastern North America. Earthquake Spectra, 30(1), 1–24.
  • Kramer, S. L. (1996). Geotechnical Earthquake Engineering. Prentice Hall.
  • USGS (2014). The 2014 National Seismic Hazard Maps: USGS hazard modeling and data products. USGS Open File Report.
  • Abrahamson, N. A., Silva, W. J., Kamai, R. (2014). NGA-West2: Ground-motion prediction equations. Earthquake Spectra, 30(3), 983–???
  • Campbell, K. W., & Bozorgnia, Y. (2014). NGA-West2 ground-motion predictions: Horizontal components. Earthquake Spectra, 30(3), 1–???
  • Chiou, B. S. J., & Youngs, R. R. (2014). NGA-West2: Central and Eastern North America GMPEs. Earthquake Spectra, 30(1), 1–24.
  • Petersen, M. D., et al. (2014). The 2014 USGS hazard maps: Data and modeling for the Conterminous United States. USGS.
  • Bozorgnia, Y., Campbell, K. W., et al. (2014). NGA-West2: Additional GMPEs and modeling updates. Earthquake Spectra, 30(3), 117–150.

Related Assignments