We wanted to include the effect of vertical velocities from, e.g., Ekman pumping and other processes. So I modified the code (omega.F), replacing omega with prescribed vertical velocity (~10^-6 m/s). We noticed that this artificial vertical velocity would induce some non-conservation problem, so we also tested modifying the omega only in step3d_t. Both two cases showed similar problem:
The problem is that, the modification resulted in unrealistic MLD deepening.
![Image](http://ww4.sinaimg.cn/mw690/653aa167gw1exryp4x5pdj20k80erwfl.jpg)
![Image](http://ww3.sinaimg.cn/mw690/653aa167gw1exryiqjgp7j20gi0awwh0.jpg)
problematic akv (larger magnitude) and temp. profiles (cyan curve is the MLD):
![Image](http://ww1.sinaimg.cn/mw690/653aa167gw1exryp3sdlzj20ka0f375b.jpg)
![Image](http://ww3.sinaimg.cn/mw690/653aa167gw1exryugvuloj20it0c7taz.jpg)
original akv and temp. profiles:
It is related with the vertical mixing scheme because a diagnostic plot (~150 m) shows that the temperature decreases is due to vertical mixing(red), not vertical advection (green, it is closely following the prescribed w) we added in.
![Image](http://ww4.sinaimg.cn/mw690/653aa167gw1exrzhldv29j20jr0f6aak.jpg)
We are using the LMD vmixing scheme, with:
Code: Select all
# define LMD_RIMIX
# define LMD_CONVEC
# define LMD_SKPP
# define LMD_BKPP
# define LMD_NONLOCAL
![Wink :wink:](./images/smilies/icon_wink.gif)