LIM3 future bug fix: heat budget

[clem]

Dear happy LIM3 users. We found a bug from the last revision of LIM3 which impacts the heat budget of the ice-ocean system (it does not concern the previous revisions). The bug has just been cornered but I do not have the time to fix it now. It's time to go home for me. I'll fix it on Monday/Tuesday? and I will add a diagnostics to avoid such waste of time in the future.
Clem

[clem]

We just posted a commit which solves a number of bugs (rev. 5167). It mainly concerns heat, salt and mass exchanges bewteen the ocean and the ice/snow. Details of the changes are described hereafter

1- limthd: wrong non-solar flux was passed to the ocean in coupled case
2- limthd: in case monocat, we fixed a problem of conservation
3- limthd_dh: division by 0 in case ice temperature = melting temperature is now avoided
4- limthd_dh: precipitations and sublimation were not perfectly taking into account
5- limthd_dif: changing the sign of the heat flux resulting from the non-convergence of the scheme
6- limitd_me: some rewritting to be coherent with the rest of the code

Also we removed the useless dimension (nlay_i+1) for a number of variables. It was not a bug but induced some confusion (see ticket #1497).

And finally, we coded clean online diagnostics in limcons.F90 to check whether a simulation in ice-covered area conserves mass, salt and heat or not. The treshold is set up to 1 mm of ice that is spuriously lost/gained over 100 years. You can activate it with the logical ln_limdiahsb but be aware that it is only working with closed boundaries for the ice (no ice-BDY) and no ice damping.

If you are interested by checking conservation in your simulation, you also have helpful ice-ocean outputs in xxx_scalar.nc that can be activated with the logical ln_diahsb.