Notes:
For implementation passing in just Mean fD into the bradley model to adjust. Then get the sigma value at the end and combine that with the fD sigma to get the overall sigma.
NL soil is most important at lower periods and directivity at higher periods, see an interesting mix in the middle.
Plan:
1: High Vs30 760 for all 3 GMM approaches for 9 sites
2: Lower Vs30 180 for all 3 GMM approaches for 9 sites
3: Plotting sigma values as well as mu
4: Spatial ratio plots with a given set periods for 10,000 sites
Test variables:
Fault: AlpineK2T
Model: Br_13
Number of Hypocentres: 100
Method: Latin Hypercube
Spatial Periods: (1,2,5,10)
Summary:
Compared 3 different approaches to producing mu and sigma values for 9 sites around a fault for periods 0 - 10.
Results below show the 3 methods and their IM values for the given periods where Vs30 = (180, 760).
All results seem very similar, so ratio plots were made to compare them.
Both mu and sigma were compared.
Mu had 1% is changes in the 1 - 3 T range for some sites closer to the fault for Vs30 180 and gradual increase for Vs30 = 760 for some sites.
Sigma changes were very small 0.2% being the largest change for Vs30 of 180 and 0.01% for Vs30 of 760 between uncoupled and fully coupled.
Spatial plots were then looked at for given periods (1,2,5,10) to see spatially how mu changes between the different methods and if there are patterns.
Most results seemed really small ~1% and random spatially apart from Vs30 180 Uncoupled / fully coupled which for lower periods T = (1, 2) there is some correlation between the difference in mu and how high / low the mu value is.
Sigma was also done spatially to see what the difference was between the methods for all sites at given periods.
Only difference in values was between Uncoupled / Fully Coupled which seemed to affect only 1 - 3 T periods and were very small values as expected from the 9 site plots.
Another test that was executed was how much sigma changes when you add directivity for each method.
Below shows the 9 site plots as a ratio of the Sigma GMM / Sigma Total. This shows the impact directivity has on sigma as given periods for these 9 sites.
Spatial plots were also done for both Vs30 values (180, 760).
