How to construct a three-dimensional geological model of strata with lens bodies in strata using gempy

[hanppyyong]

I want to build a three-dimensional geological model. This model has four strata from top to bottom, but there is a soil layer in the middle of these strata. This soil layer is a lens. How should I use the gempy model to build this three-dimensional geological model? How to extract directional features and formation surfaces? In particular, how can the information features and direction vectors of the surface be extracted from the stratigraphic boundary lines where the lens contacts other strata?

[NilsChudalla]

Hi @hanppyyong ,
that sounds like something you should be able to model in GemPy! I assume you understand the basics of GemPy and will reference terms associated with it. If you are new to GemPy, I suggest that you check out our YouTube Tutorial (Series)[https://www.youtube.com/watch?v=R-vUld4V-OQ&list=PLaOXHTg4AuvOngAjIOTLa-dmcPaoZMRDF]

Now, your questions:

• Model the lens:
  Not knowing the geological complexity your modeling in more detail, I would try to keep it simple. Try to model it as ersional surfaces (default). Your structural frame could look something like

{
"Top_series": ["Youngest_layer", ...], 
"Lens_series": "Lens_layer",
"Bot_series":["Layer_below_lens", ...]
}

• Orientation data:
  Generally, the underlying algorithm in GemPy has to receive one orientation measurement per SERIES. Orientations are usually great to improve the structural complexity. But you can also use horizontal orientations in "neutral" positions or even outside of the model domain and rely on the surface points producing the geometry you require. In this case, you should have a solid coverage with surface points.

• Surface data:
  Sample the bottom contact of your layers. This can be in real world coordinates from well logs or in a pseudo coordinate system (extracted from image software, e.g. using the inbuild ruler)

If this reply does not help you, please share some more info (e.g. a sketch of your expected structure and info on the available data in your project)

[hanppyyong]

Thank you very much for your answer! I still have some doubts. For instance, in the example I uploaded, should both the upper and lower surfaces of the lens body len1 be formed as len1? Is it correct to select the direction of the arrow I drew in the direction data as the direction? Especially, is the direction of the lens body len1 correct?

[javoha]

Hi @hanppyyong,
so the setup you show in your image would not work properly - while it is possible to have downward facing orientations (e.g for overturned folds like in this simple example) in your case it would probably break the logic of the stratigraphic pile.

I see two main options to tackle this scenario:

A) We get a little creative with the subdivision of the model: In your scenario I would split the "geo2" element into two separate structural elements, one below and one above the lens (A1 and A2 in the sketch below). This way you can basically treat you lens like a pinchout layer and define the bottom of each elemtn as shown in the sketch.
( I marked the smaller lens at the right with a question mark because if you want to model it the same way you would also need to make it its own element and maybe subdivide what is currently the basement).
This solution makes sense if you have a rather simple lens situation like here that you can easily resolve with a little bit of subdivision.

B) If you have a scenario where you have many (small) lenses within a layer you might need to create multiple hierarchies and gempy might only be part of the solution. This would mean creating a structural model only for the "geologic units" with gempy (excluding lenses) and then using som other geostatistical method or object based method to interpolate the lenses afterwards within the different gempy elements.

I hope this helps.
Cheers,
Jan

[hanppyyong]

Thank you very much for your answer! I understand what you mean. Your answer has been very helpful to me. At present, I am modeling multiple small lenses in a stratum. If there is any progress in the future, I will publish it for everyone to discuss together. Thank you again.