inconsistent station propagation coefficients with Pugh et al. 2016 Appendix A

[djpugh]

inconsistent station propagation coefficients with Pugh et al. 2016 Appendix A

[djpugh]

Thanks @hvasbath - I'll need to refamiliarise myself with the formulation as I haven't actually been working in this space for a while to double check - but I believe the difference is because in MTfit the 6 vector is ordered as:
Mxx
Myy
Mzz
sqrt(2)*Mxy
sqrt(2)*Mxz
sqrt(2)*Myz

While in the paper it is ordered as (for consistency with Chapman (?) IIRC)
Mxx
Myy
Mzz
sqrt(2)*Myz
sqrt(2)*Mxz
sqrt(2)*Mxy

Plus there is some expansion of double angles as well
I've added some comments to the code in the PR #67 - please could you check that it makes sense to you - i will try and double check later after work as well

[hvasbath]

Hi @djpugh ,

thanks a lot for the clarification! The double-angles hint resolves everyhing, but I wonder why you do that, because numerically its two times faster to calculate sin(2x) instead of 2sin(x)cos(x).

[djpugh]

TBH, I can't remember why, it's a good point (in fact should probably cache the cos and sin values - I think because it is a one off calculation per run, I probably didn't focus on it

[hvasbath]

FYI, was testing eq A7-A9 against A15-A17.
Found the error. Its in equation A2 Appendix A.
Must be in third component -sin(take-off angles) instead of sin(take-off angles). Its just a typo comparing to Aki and RIchards.
Again thanks a lot!