Closed budget of T or S using TRD

1

Dear All,

The TRD module in NEMO computes trends in T or S for any grid cell, and for each process (advection, vertical diffusion, horizontal diffusion, bbl etc…), as well as the total trend.

Has anyone managed to compute a locally closed budget using this module AND a non-linear free surface (vvl).
What I mean is that let’s say you output each trend over for example 1 year of simulation and for each grid cell of and compute a budget error as

epsilon(i,j,k)=| d(T(i,j,k)*e3t(i,jk))/dt - Trend_in_advection_of_T(i,j,k) - Trend_in_vertical_diffusion_of_T(i,j,k) - etc… |

and then produce a 2D map of this error by integrating over the vertical axis.

Has anyone managed to get an error which is basically of the order of the computer accuracy ?

Thanks for your help,

Robinson

2

Hi Robinson,

A Phd student does actually make such budgets at MOI (Fernand Assene). Here is a link to a paper (under review) from which you may reach him if needed.

3

Merci Jérôme, I will contact him.