Issues with evolving base state for the Urca problem

[biboyd]

When using evolve_base_state=true, there is a large amount of noise in the convectively stable region outside the convection zone. Below is a comparison between two simulations, the left keeps base state constant while the right evolves the base state. The evolution scheme (ie advection vs average) yields similar noisy behavior.

The noise starts near the initial convective boundary (where the temperature profile transitions from adiabatic to isothermal). Smoothing this transition helps some, but after enough evolution the noise eventually appears and dominates the outside regions.

Inputs file for this specific are attached. Uses the Exec/science/urca problem.

inputs_models.tar.gz

[zingale]

A good place to start might be with the test_basestate unit test. I am not sure the current state of this, but this was used in the previous Maestro to compare the 1D base state expansion to a 1D Castro simulation with the same initial model to make sure we were doing the expansion correctly. See, e.g., Figure 7 here: https://iopscience.iop.org/article/10.1088/0067-0049/188/2/358/pdf

you might try that same test out with your model.