rft: separate vertical and full strain

[haakon-e]

This pull request refactors the computation of strain rate tensors throughout the codebase to clarify the distinction between vertical-only and full (vertical + horizontal) strain rate calculations. It introduces new, more explicit utility functions for these computations, updates all relevant usages to call the appropriate function, and adds improved documentation to guide future development.

Key changes include:

Refactoring and API Improvements:

• Introduced compute_strain_rate_center_vertical and compute_strain_rate_face_vertical functions to explicitly compute strain rate tensors using only vertical gradients, replacing the more generic compute_strain_rate_center and compute_strain_rate_face in all relevant locations. [1] [2] [3] [4] [5] [6] [7]

• Added new functions compute_strain_rate_center_full! and compute_strain_rate_face_full! to compute the full strain rate tensor (including both vertical and horizontal gradients) at cell centers and faces, respectively, with detailed documentation and guidance for use.

Documentation and Usability:

• Expanded and clarified docstrings for all strain rate computation functions, making it clear which functions use only vertical gradients and which compute the full tensor, and providing usage notes and references for future developers.

• Added a new strain_rate_norm utility function to compute the norm of the strain rate tensor, with optional projection onto specific axes.

These changes improve code clarity, ensure correct usage of strain rate computations, and provide a stronger foundation for future development and maintenance.

[haakon-e]

This change is part of the following stack:

• rft: separate vertical and full strain #4073 ◀
  • rft: smagorinsky-lilly for future h-v splitting #4074
    • rft: split Smagorinsky diffusive model to horizontal and vertical components #4054

_{Change managed by git-spice.}

[szy21]

Thanks! It would be good to add tests for the full strain rate too.