Convective adjustment with UKI seems to be failing?

[navidcy]

It seems to me that the UKI doesn't converge to the true parameter values. Am I interpreting this correctly?

See figure at the bottom of https://clima.github.io/OceanTurbulenceParameterEstimation.jl/dev/literated/unscented_perfect_convective_adjustment_calibration/

cc @Zhengyu-Huang

[navidcy]

OK perhaps there is a different reason for this. The fact that UKI cannot converge to the true value for the background diffusivity may be implying that that particular value doesn't affect the results that much or that our noise levels are too high?

[navidcy]

I note that it's hard to point to a docs link since dev keeps changing. Another reason to push for #71.
The UKI problem was different when I initially posted the issue. However, still now the UKI results seem cumbersome and puzzling...

[glwagner]

Why does package registration freeze the docs?

[navidcy]

Tagging does that. Nor registration. My bad...

[glwagner]

Ahh