Should I set diffu_hgtf = 1 in a simulation using non-hydrostatic core?

[jingxiangchung]

Dear Graziano,

I am trying to conduct a 5km simulation over the Peninsular Malaysia using RegCM4-NH. I am using the MM5 non-hydrostatical core for this task. Upon completing the simulation, I noticed that the model tends to produce very weird precipitation pattern (see figure,).

Upon playing around with some options, I found out that "diffu_hgtf" (Add topographic effect to diffusion) is the culprit. By turning the "diffu_hgtf" on, the simulated precipitation looks more appropriate. Nevertheless, I dare not to jump on the conclusion that I should conduct the simulation with "diffu_hgtf" turning on as it is stated that "Dynamical core parameters : Use defaults, be on your own otherwise!", and the default for using MM5 non-hydrostatical core is that the "diffu_hgtf" should be turned off.

This put me in an unsure situation on how to proceed with my work, and hence I decided to seek your opinion on this issue. Could I had done anything wrong in setting up the simulation? Please do advice me.

Thank you very much.

Figure: comparison between two simulations where diffu_hgtf = 0 and diffu_hgtf =1:

[graziano-giuliani]

Perfectly safe to use that option, which was created exactly to diffuse topographic precipitation, i.e. precipitation concentrated on the topography. The rationale behind is to increase the diffusion coefficient when the topographic gradient is high.