New GMM verification and followup

Created by: pmpowers-usgs

Following preliminary implementations of new GMMs, this issue summarizes open questions and remaining work to be done.

Zhao et al. 2016 (nshmp-haz-v2#225): (model has been deprecated and removed !260 (merged))

• Unit tests.
• Review site amplification model; smoothing between vs30; should hard rock be added as an anchor/key above "Site Class I"?
• 0.75 s not supported; needs interpolation; 7.5s, 10s?
• Check nonlinCrossover() notes; there are edge cases where Zhao recommends using approximate functional forms from Zhao, Hu et al. (2015), but these are difficult to follow

Graizer & Kalkan 2016 (nshmp-haz-v2#224): (model has been deprecated and removed !260 (merged))

• Unit tests
• Review basin term; default at 150m for CA (pers. comm.); interpolating between z1p0 and z2p5; Grazier supplied z2p5 to z1p5 conversion 11/7/17

Atkinson 2010 (nshmp-haz-v2#226): (completed as part of Hawaii NSHM)

• Unit tests
• Review depth dependence once updated HI source models implemented

Wong et al. 2015 (nshmp-haz-v2#227): (completed as part of Hawaii NSHM)

• Unit tests
• Needs siteamp model (so does McVerry for use of volcanic model in HI)

Sub models with CB14 basin (nshmp-haz-v2#256): (completed as part of 2018 NSHM)

• Unit tests
• BcHydro12
• A&M09
• Zhao06 (+ site term smoothing) (not doing)

NGA-East site amplification (nshmp-haz-v2#257):

• Unit tests #406