Changes between Version 24 and Version 25 of ticket/0863

Timestamp:

2011-10-25T15:16:54+02:00 (13 years ago)

Author:

mlelod

Comment:

--

ticket/0863

@@ -134 +134 @@
 
 === Isopycnal diffusion ===
-'''CAUTION''': Not coded yet.
-
-For the computation of isoneutral slopes, the density of neighbor cells has to be computed at the same depth.[[BR]]To that purpose, an argument is added in the `eos` routine that gives the depth at which density is calculated.[[BR]]For isoneutral slopes, `fsdept_0` is used. For the pressure gradient, the actual depth `fsdept_n` is used.
-
-=== Thickness weighted outputs ===
-In vvl case, the "toce" variable in the netcdf file does not contain temperature anymore but the heat content fse3t_n * tn.[[BR]]The same thing is done for salinity and horizontal velocities (only in flux form advection but I'm not sure if it has to be done in any vvl case).[[BR]]3 other variables are created in netcdf outputs:
+The computation of potential density is done at a certain depth. In the vvl case, neighbor cells potential densities are not computed at the exact same depth, which causes an error in the computation of isopycnal slopes. To fix this problem, the depth has been added as an argument in the `eos_insitu` and `eos_insitu_pot` routines.
+
+When the `eos` interface is called before computing the isopycnal slopes, `fsdept_0` is used. When it's called before the computation of the hydrostatic pressure gradient, the current depth `fsdept_n` is used.
+
+=== Thickness weighted outputs (active and passive tracers + velocities) ===
+In vvl case, the "toce" variable in the netcdf file does not contain temperature anymore but the heat content fse3t_n * tn.[[BR]]The same thing is done for salinity and horizontal velocities (only in flux form advection but I'm not sure if it has to be done in any vvl case).[[BR]]The same treatment is also applied to passive tracers.[[BR]]3 other variables are created in netcdf outputs:
 
  * `e3t_n` is the vertical scale factor at t-points (`_n` suffix indicates that it is different from the background scale factor).[[BR]]e3u and e3v and e3w are not stored, they can be recomputed off-line.
- * `dept_n` is the depth at t-points. Depth at u, v and w points should be computed off-line also.
+ * `dept_n` is the depth at t-points with a 0 reference when the ocean is at rest (ssh is removed).[[BR]]Depths at u, v and w points should be computed off-line also.
  * `e3tdef` is the squared deformation ratio of `e3t` given in %!^2.[[BR]]During the post treatment, the square root of this quantity gives the variance of the level thickness in %.
 
@@ -148 +148 @@
       z_e3t_def(:,:,:) = ( ( fse3t_n(:,:,:) - fse3t_0(:,:,:) ) / fse3t_0(:,:,:) * 100 * tmask(:,:,:) ) ** 2
       CALL iom_put( "e3t_n"  , fse3t_n  (:,:,:) )
-      CALL iom_put( "dept_n" , fsdept_n (:,:,:) )
+      CALL iom_put( "dept_n" , fsde3w_n (:,:,:) )
       CALL iom_put( "e3tdef" , z_e3t_def(:,:,:) )