add mass flux divergence to detrainment

[szy21]

Purpose

Adds a term for mass flux divergence in detrainment. This is not implemented for diagnostic edmf as calculating the mass flux divergence at level k requires mass flux at level k, which we don't have when solving edmf variable at level k. We can get around it by using levels k-1 and k-2 to calculate the divergence, but it's a bit more complicated and we are not focusing on improving diagnostic EDMF now.

I keep both vert_div and massflux_vert_div term, to have no behavior change for diagnostic edmf. We can reevaluate later if vert_div is still needed. I also remove GeneralizedHarmonicsEntrainment and GeneralizedHarmonicsDetrainment, as it's unlikely we will use them.

part of #2705

To-do

Content

---

• I have read and checked the items on the review checklist.

[szy21]

I will update ClimaParams for other environment as well. - done