Re: [ferret_users] integrate vertically to a variable level

["William S. Kessler" <William.S.Kessler@xxxxxxxx>]

Hi Elodie -

In general you can't use @DIN except with a definite value.

There might be another way to do this, but one easy way is to set all  
values of u_z to zero above the MLD, then integrate to the surface,  
using a mask:

let zz = z[gz=Temp]
let mask = if zz gt mld then 1 else 0
let u_z_mask = u_z*mask
let um = u_z_mask[z=0:2000@din]   ! note if z is positive down, then  
integrate downward

When doing this, consider the vertical resolution. If the velocity  
increases towards the surface, the result might be strongly dependent  
on exactly where you stop integrating. Does the MLD jump back and  
forth from, say, 20 to 30m? That might make a big difference, and  
therefore cause a spurious variability. In that case you might want  
to interpolate your fields to higher vertical resolution before doing  
this calculation. On the other hand, it might very well be that there  
is little vertical shear within the mixed layer; in that case the  
velocity at its base might be essentially the same as that at the  
surface.

Billy


On May 16, 2006, at 7:22 PM, Elodie MARTINEZ wrote:

> Dear Ferret Users:
> I have trouble to integrate vertically to a variable level.
>
> I would like to obtain the flow field at the base of the mixed  
> layer (defined below and referenced as MLD) assuming the
>
> depth of no motion is at 2000 m, so I have done:
>
> !calculation of the mixed layer depth let Tsurf=Temp[z=0]
> let Tsurf2=T_RUN2_Temp-Tsurf
> let MLD=Tsurf2[Z=@LOC:-0.5]
>
> ! The flow field at the base of the mixed layer
> let um = u_z[z=2000:MLD@DIN]  (with u_z the vertical shear)
>
> in this last equation, ferret doesn't recognize MLD (which depend  
> on latitude, longitude and time)
>
> I wonder if you know how to solve this problem?
>
> Thank you very much   in advance!!!
>
> Best regards,
>
>
> elodie
>
>
>
>
> <elodie.martinez.vcf>