WIP: Iasp91 P S tables small fix

[salichon]

Make this arrival tables more consistent
Avoid a negative value for the S arrival table

[cla-bot]

We require contributors to sign our Contributor License Agreement, and we don't have @salichon on file. In order for us to review and merge your code, please contact @gempa-jabe to get yourself added.

[salichon]

Hi @gempa-jabe
a rapid follow up on velocity table fix.
Happy to sign the cla

cheers
jerome (GNS)

[gempa-jabe]

Hi @salichon, I won't merge that easily. I need to confirm at least with @jsaul and even better, with tests. Those tables were added decades ago and I don't know the origin of those values.

[jsaul]

I don't know the origin/reason of these offsets but t(0,0) != 0 cannot be correct. The offset does appear to affect other distances, too, but AFAICS "only" for zero depth. This needs to be fixed. I will look into that in detail ASAP. Thanks to @salichon for pointing it out.

[salichon]

Kia Ora and thanks @jsaul @gempa-jabe
All good to me,
i understand of course this needs some testing and so :)
I made some quick check on my side although not as thoroughly or as lengthy as it could be.
(without the creation of these tables for instance).
This came out of a question about depth comparison for a shallow hypocenter with 2 methods but same iasp91 velocity model.
I wanted to exclude this negative S wave value question from the discussion I had.
Cheers!