Context & Analysis

ABL restart not handled by XIOS when using ln_xios_read & nn_wxios

Fix

adapt what is done for ice to abl

Context

I introduced a nasty bug between release 4.0.3 and 4.0.4 It's a typo that changed drastically the distribution of solar and non-solar fluxes over ice covered areas. As a consequence, the ice cover and volume increased and some processes had wrong seasonality. In particular the lateral melting.

Since this bug has been introduced while correcting other bugs, and given the importance of it on the ice simulations, I feel uncomfortable keeping the last revision 4.0.5 untouched. Fortunately the official announcement of this revision has not yet been done.

So, I wonder if I could erase the current 4.0.5 and redo it. I know it is not the way we should work but I do not want people to use this version and not be aware of that. What say you?

Besides documentation, here are the only changes in 4.0-HEAD vs 4.0.5: NEMO/releases/r4.0/r4.0-HEAD@14359:14435 It concerns a bug fix in iceberg module (ticket #2581) and a small incompatibility oasis-agrif (ticket #2621)

Fix

in iceupdate.F90, replace this line:

            qns(ji,jj) = qt_oce_ai(ji,jj) - zqsr

by this one:

            qns(ji,jj) = qt_oce_ai(ji,jj) - qsr(ji,jj)

In the current version of NEMO/trunk/src/ICE/ice.F90 (last changed at [14252]), in subroutine pnd_TOPO, the melt pond depth h_ip is calculated at line 777 as v_ip/a_i. The denominator a_i looks wrong and differs from the denominator a_ip used in similar calculations of h_ip elsewhere in the same subroutine (lines 616, 680, 704, and 738).