Up to Specific issues in real coupled models
Hello, Testing a real regional configuration NEMO (3.6) + WRF (4) + XIOS (1) + OASIS3-MCT (4), I am confused about the use of the coupled restart files. From the documentation, I understand they are needed for LOCTRANS operations, although using AVERAGE should perhaps not require them. Now, looking at SST, at the beginning of the coupled run I provide in input a previous SST field from NEMO (just for testing, variable O_SSTSST). This is in Celsius degrees. In output, the ocean coupled file is overwritten and I find (focusing only on SST) these two SST related variables in the NetCDF: 1) float O_SSTSST(y, x) : This has been changed, and I assume is the SST at the latest timestep, but weirdly is in Kelvin now, and not in degC. Is this what I should expect? 2) double loc000001_O_SSTSST(loc000001_O_SSTSST_ny, loc000001_O_SSTSST_nx) : This is filled with 0 and not sure what is meant for So my questions: is the change of units fine? or perhaps the input SST should be in Kelvin, according to WRF (but no crash occurred, I hardly believe it has been really used...) in a sequence of coupled runs, the variable in input and output will always be in the same file and with the same name, so that in principle I could leave the restart file there and O_SSTSST will be updated continuously? Thanks in advance, Bests, Andrea
Hi Andrea, Concerning the LOCTRANS operation and the restart file, you have information in the User Guide https://oasis.cerfacs.fr/wp-content/uploads/sites/114/2021/02/GLOBC_TR_oasis3mct_UserGuide_4.0.pdf, section 4.1 . "The coupling restart file allows the partial time transformation to be saved at the end of a run for exact restart at the start of the next run. When LOCTRANS transformations are specified (for all of them except INSTANT), the initial coupling restart file should not contain any LOCTRANS restart fields. For the following runs, it is mandatory that the coupling restart file contains LOCTRANS restart fields coherent with the current namcouple entries." This is why you have nothing to put for your first run, but that you have the field "double loc000001_O_SSTSST" written by OASIS in the restart file at the end of the run if you asked for a LOCTRANS transformation different from INSTANT in your namcouple for O_SSTSST. The number "000001" corresponds to the LOCTRANS transformation you asked for. Sometimes the loctrans restart field loc000001_O_SSTSST equals 0 because nothing have to be added to the field O_SSTSST at the beginning of the next run, sometimes it is different from 0 and it will be added to the field at the beginning of the next run. Anyway, it is calculated by OASIS. I do not understand why you get an SST in Kelvin in your restart file if you send it in DegC. Do you have the same in your O_SSTSST debug output files (ie a kelvin field instead of DegC) ? (to have a debug file put option EXPOUT in the namcouple). By the way could you send to me your namcouple at oasishelp@cerfacs.fr ? Thanks, Best regards, Laure
Well, I found that both diagnostic files from NEMO and WRF (when using EXPOUT) have at the 1st record SST in degC and for the remaining records in Kelvin! Maybe the first record comes from my restart file and it is actually not used but reported anyway in the diagnostics? And all others are in Kelvin, as I assume should be for the bulk formulas. I should dig in the code, though, Thanks and cheers, Andrea
The fact that the temperature SST is in Kelvin in the debug output files and the restart file at the end of the run come from the models. Is the temperature in degC in your initial restart files ? If so I think you should change it to Kelvin values to be coherent with the calculations in the models. Best regards, Laure
Thanks for the feedback Laure, Yes I will modify the initial restart, Bests, Andrea
