Estimating unknown site parameters required by some GMMs

[raoanirudh]

Estimating unknown site parameters required by some GMMs

Some ground motion models include an adjustment to the base model to consider the effects of basin depth on ground motion amplitude. In these cases, the site-response model within the GMM typically includes a parameter for sediment depth, most commonly represented by Z_1.0 or Z_2.5. In the process of developing the GMMs, some of the recording stations in the database might not have reported values for these parameters. Thus, the GMM developers typically employ an empirical relationship relating Z_1.0 to V_s30 or Z_2.5 to V_s30 to estimate reference Z_1.0 or Z_2.5 values for sites where these values are not available.

The relationships adopted by the different NGA-West2 GMMs for estimating Z_1.0 from V_s30 are shown below:

---

Chiou, Youngs - 2014 - Update of the Chiou and Youngs NGA model for the average horizontal component of peak ground motion and response

---

Abrahamson, Silva, Kamai - 2014 - Summary of the ASK14 ground motion relation for active crustal regions

---

Boore et al. - 2014 - NGA-West2 equations for predicting PGA, PGV, and 5% damped PSA for shallow crustal earthquakes

---

These equations are similar but not identical. When Z_1.0 is unknown, Boore et al. (2014) recommend turning off the adjustment factor to consider the effects of basin depth on ground motion amplitude.

This raises a couple of questions / considerations:

1. If a user has a GMM logic tree containing more than one of these GMMs, each of which recommends a different empirical relationship to estimate unknown Z_1.0 values, they are currently able to provide only one set of Z_1.0 values in their site model file. Ideally, perhaps the estimated Z_1.0 values will need to differ based on the empirical relationships provided by each GMM, as was recommended by Kaklamanos et al. (2011) for estimating unknown site parameters for the NGA-West1 GMMs (see below). In the above example, only the NGA-West2 GMMs are shown where the three sets of equations are not too different from each other, but this issue will extend to any other GMMs that use adjustments for basin depth.
2. Should the parameters representing sediment depth, Z_1.0 or Z_2.5, be listed as mandatory required site parameters in the implementation of these GMMs in the engine, if in most of the applications, these parameters are likely to be unknown? If Z_1.0 and Z_2.5 are removed from the list of mandatory site parameters for the relevant GMMs, the inbuilt reference relationships coded within each GMM can then be used to estimate these parameters, effectively turning off the basin depth adjustment factor for sites where these parameters are unknown? If the user does have known values for these parameters, then they could be provided through the site model file in the usual way.

---

Kaklamanos, Baise, Boore - 2011 - Estimating Unknown Input Parameters when Implementing the NGA Ground-Motion Prediction Equations in Engineering Practice

[CB-quakemodel]

I will get around to this. We need to in effect turn off the use of the GMM's own empirical relationships if provided in the site model (i.e. only calculate from vs30 within each GMM if missing in the provided site model).