Hello,
In Nemo 3.6, it used to be possible to run a 1D configuration, and specify in the namelists the exact same forcing files as for a regular (3D) configuration. For example, atmospheric forcing fields were 2D fields (or 3D when including time), and initial conditions were 3D fields.
When specifying a (longitude, latitude), Nemo would automatically run the 1D model at the right location.
In Nemo 4.2 (and probably 5.0?), this does not seem to work. One has to provide data exactly at the specified location. I haven’t tried, but I suppose it could also be possible to use the 2D or 3D forcing files and use weights (interpolation) to extract just the right point or column at run-time.
In any case, it is more cumbersome than what used to be possible in Nemo 3.6: to use exactly the same namelist_cfg in 1D as in 3D configs, except that in 1D the (longitude,latitude) of the 1D model is also specified.
Cheers,
Luc
Hello Luc,
I was not aware of this 1D feature from Nemo 3.6. After some svn digging, this dedicated routine called domc1d.F90 disappeared somewhere between nemo 3.6 and nemo 4.0, when the domcfg.90 routine has been transformed into a dedicated pre-processing tool (domaincfg).
I agree with you that using dedicated interpolation weights or extracting local data from global simulations (this is what I do using cdo tool) offline are not as convenient as what was possible to do online with nemo 3.6.
But I also see that jpizoom and jpjzoom variables, which were setup before in domc1d.F90, are still available in the domaincfg tool, so perhaps it’s worth it to give a try to check if you can reproduce nemo 3.6 feature using domaincfg tool as a first attempt ?
If yes, perhaps we can think about adapting the old domc1d.F90 routine to work somehow with the domaincfg tool.
Thanks for your feedback,
Guillaume