Implementation of GMM of Ambraseys 1996 and Sapetta_Pugliese 1996

[mascandola]

The GMMs of Ambraseys 1996 and Sapetta_Pugliese 1996 are implemented as in MPS04

[micheles]

Please update the changelog

[mascandola]

Please update the changelog

Done but results in a new pull request

[mmpagani]

Thank you, @mascandola for this PR.

The magnitude in Ambraseys et al. is Ms. The standard in the OQ Engine is Mw. I suggest adding to your implementation of the Ambraseys a function that converts the input magnitudes (that we assume Mw) into Ms. If you want to be consistent with the original model, you can use their eq. 2 for the conversion.

In the case of Sabetta and Publiese (1996), the homogenisation of magnitude is even trickier as it is supposed to be ML below 5.5 and Ms above this threshold. I suggest adopting a strategy as described above for AMB96 but with conversion equations from literature, perhaps using as a reference https://emidius.mi.ingv.it/CPTI15-DBMI15/data/CPTI15_v4.0_description.pdf

The advantage of standardizing magnitudes to Mw is that it will be possible to use these GMMs interchangeably with more recent ones and compare results.

For the distance metric, SP96 says that R in their eq. 4 is "is distance (fault or epicentral)". This is not completely clear to me but I would be inclined to say that distance is Rjb (and when the rupture dimensions are negligible this inherently equivalent to epicentral).