Hello together,

I was recently browsing though the code looking for where the sea ice thickness in uncoupled OpenIFS is set. I found that it is read from the restart file and written to the restart file as ZTHKICE and handed around a bunch as PTHKICE but nothing more. So  I take it we have sea ice thickness persistence? I was half-expecting to find something like the sea ice albedo climatology from MODIS data. Or am I missing something?

Cheers, Jan


10 Comments

  1. Glenn Carver

    Hi Jan,

    I can't answer this. I'll find someone who might be able to help.

    Glenn

  2. Sarah Keeley

    HI Jan,

    The sea ice thickness is set in the surface routine and is fixed at 1.5m and uses the soil levels calculations to do the thermodynamics in the calculations.

    We use an albedo climatology based on SHEBA campaign data

    Let me know if you need more details -response time might be slow - home schooling as well!

  3. Jan Streffing

    Hello Sarah,

    thank you very much for the info. That's all I needed to know. Thank you Glenn for reaching out.

    Cheers, Jan

  4. Joakim Kjellsson

    Hi Jan, Glenn and Sarah

    I'd like to ask a follow-up question then: If I send sea-ice thickness from NEMO to OpenIFS, is it ever used? I was digging through the code and could not see that either ZTHKICE or PTHKICE is used for any calculations. 

    Cheers
    Joakim 

    1. Sarah Keeley

      You can send it to be used in the thermodynamic calculations.  There is a switch to change the surface conductivity to snow if snow thickness is greater than zero in NEMO and with a bit of a hack you could couple the ice thickness by changing the thickness of the bottom soil layer but it is not something that is usually done - this would still set the minimum thickness at 1m unless you also moved the definitions of the sea ice layer thicknesses.

      The flags will only be in the surface routines in surf module if they still exist in the openifs code base


      1. Glenn Carver

        The surf routines should still be there but let me know if something is missing.

      2. Joakim Kjellsson

        Thanks for the reply, Sarah. 

        If I understand this correctly, it means that sending ice and snow thickness from NEMO/LIM to OpenIFS is unnecessary since it's not used anyway? So both the IFS and EC-Earth always assume 1m thick sea ice in the surface flux calculations? 

        The only use of snow thickness I can find is in surf/module/surfseb_ctl_mod.F90 : 

        IF (PSNTICE(JL) > 0.0_JPRB) THEN
        ! For now use same conductivity as snow on land
        ! Possible refinement: take fractional snow cover for thin layers of snow into account
        IF (PTSKM1M(JL,JT) >= PTSRF(JL,JT).AND.PTSKM1M(JL,JT) > ZRTTMEPS) THEN
        ZLAMSK(JL,JT)=ZLARGESN
        ELSE
        ZLAMSK(JL,JT)=RVLAMSK(5) ! Snow tile!!
        ENDIF

        i.e. it is only used to figure out if there is any snow. 1 cm or 1m of snow does not make any difference? 


        1. Sarah Keeley

          The ice is always assumed to be 1.5m thick

          I am not sure if EC-Earth is doing - but operationally we assume 1.5m thickness - it means the surface temperatures can be quite off when the ice is much thicker or thinner and if it is snow covered.

          One option coded by ECEarth which is in the IFS is LESICE switch - this allows you to bypass the IFS thermodynamic calculation and use the temperatures from the ice model - this then works with snow thickness switch as it tells the skin temp calculation whether the temperature seen at ISTL1 which is coupled from the ocean model is an ice temeprature or snow temperature. which is why it doesn't care about the thickness.;

          1. Joakim Kjellsson

            Hi Sarah

            Ah, then it makes sense. This is probably what is going on for me as well. I'm passing skin-temp over ocean and ice which means ice thickness is never used and snow thickness is only used to figure out whether there is snow or not. I'll double-check that the LESICE switch is there. 

            Thanks for explaining what's going on. My knowledge of the surfex modules is about as good as my latin... (wink) 

            Cheers
            Joakim 

  5. Jan Streffing

    Perhaps a closing remark:

    since modifying the sea ice thickness in OpenIFS sounded unwieldy, for AWI-CM3 I switched to coupling the sea ice surface temperature, which is calculated in the sea ice model, where sea ice thickness is well known. The same strategy is employed in EC-Earth, and it makes the sea ice thickness in OpenIFS superfluous in my coupled setups.

    Cheers, Jan