Hi all

I'm running OpenIFS CY40R1 at T159 resolution and writing data every 6 hours. The output includes heat fluxes (SSR, STR, SSHF, SLHF) and water fluxes (E,TP).
I've also got this in my namelist 

&NAMPPC
LRSACC=true ! turn of accumulation of fluxes
/

which I think means that accumulated fluxes are reset each output step, so instead of "E" being total evaporation since model start it is total evaporation since last output. 

This works well for all steps, except for the second output step! 
When the model first runs, everything is fine because SSR(t=0) = 0, and then SSR(t=6h) = SSR(t=6h) + SSR(t=0), and after that I don't have cumulative fluxes. 
But when I restart the model after 8760 hours, I get SSR(t=8766h) = SSR(t=8766h) + SSR(t=8760h), which gets very big since the previous step is not zero anymore. 

I'm enclosing three plots showing the first step (from the restart, t=8760h), the second output step (t=8766h), and the third output step (t=8772h). You can see that SSR(t=8766h) looks incorrect. 

Has anyone else experienced this problem as well? Do the accumulated fluxes not reset for the second output step? 

Many thanks

Joakim


 

3 Comments

  1. Unknown User (nagc)

    Hi Joakim,

    For the restart, do you ask for output at timestep 0?  If you don't then there will be no reset and you will effectively get two output interval's worth of fluxes i.e. 12hrs worth of accumulation instead of 6. Your middle plot suggests the values are roughly double the other two, so this could be it?

    Also, just for clarification, rather than SSR(t=8766h) = SSR(t=8766h)+SSR(t=8760h), this is actually a sum of the fluxes for each timestep between t=8766h & t=8760h.

    Best regards,

                        Glenn

  2. Unknown User (de3j)

    Hi Glenn

    Thanks for the reply. Yes you're right that it is all the steps between t=8766 and t=8760. 

    How can I get output for the first time step? I just give OpenIFS an output frequency (NFRPOS=6, NFRHIS=6) and that gives me output of the first step for the initial run, but for the restart runs the first output is after 6 time steps. 
    So for the initial run, the first output file is ICMGGKCM2+000000 and the last is ICMGGKCM2+008760, while for the restart the files are from ICMGGKCM2+008766 to ICMGGKCM2+017520. Is it possilble to make the restart write ICMGGKCM2+008760 as well to force the accumulated fields to reset, and then I can just delete that file to avoid duplicates? 

    Cheers

    Joakim 


  3. Unknown User (jstreffi)

    Interesting observation. Thank's for the heads up! I had a look and found that I have a similar error in OIFS-FESOM2. I also reset accumulated fluxes and I also have 6hourly output. I run at 45min tstep so the first year has 11680 steps. After the restart it is the first, not the second output that is messed up for me. Here are in order, 11688, 11696 and 11704:

    Best regards,

    Jan