2D gravity inversion for the relief of a basin using a polygonal parametrization

[leouieda]

Implements the PolygonalBasinGravity class in the gravmag.basin2d module. This class performs the inversion for the relief of a basin by representing it as a polygon. The x coordinates of the vertices are evenly spaced and their depths are the parameters inverted for. Fixes two extra vertices to the surface at the edges of the model space.

This class is a good tool to show off and teach gravity inversion and regularization. The different effects of damping, smoothness, and total variation can be clearly seen. It has very rudimentary functionality:

• The top of the basin is assumed flat at a constant height.
• The range of the basin in the x dimension is assumed to be xmin and xmax of the data.
• Density is constant in the whole basin.

Below is the result of the cookbook recipe:

Checklist:

• Make tests for new code
• Create/update docstrings
• Code follows PEP8 style conventions
• Code and docs have been spellchecked
• Changelog entry
• Include new dependencies in docs, requirements.txt, README
• Documentation builds properly
• All tests pass
• Can be merged

[eusoubrasileiro]

How about one of this with density increasing with depth by any function?
Would be difficult to calculate the answer?
On Nov 26, 2014 9:40 AM, "Leonardo Uieda" notifications@github.com wrote:

Merged #149 #149.

—
Reply to this email directly or view it on GitHub
#149 (comment).

[leouieda]

@eusoubrasileiro there is even some matlab code for that: http://software.seg.org/2014/0001/index.html Though the license for it is a bit weird, I don't think it restricts any kind of reuse (with attribution). If you want to give that a try, I can help with discussion 😄

One thing to think about: we treat physical properties as floats (or arrays in case of magnetization) in a dictionary, right? So how would we treat depth-varying density? One idea is to pass a function. But that would break all the plotting code for prisms, etc. Maybe pass 'density' as a float (this would be the starting/reference density) and another parameter like 'density-coeff' for the decay coefficient. That wouldn't break any of the existing code and the special functions that calculate with depth-varying density could look for this special property in the prism/polygon/etc.

[leouieda]

@eusoubrasileiro maybe we could open an issue for this discussion? Would like to do that, or should I?

[eusoubrasileiro]

@leouieda master I have not forgotten that, so I opened the issue and copied the discussion #152 to keep track.