unmatch surface lateral boundary condition data

[corvianawatie]

Dear all,

I wonder if anyone here ever experienced the unmatch "lateral boundary tracer (T and S) data at the surface layer" to the "calculated data which is affected by atmospheric forcing", and how to solve this problem? I found this in my model, higher surface layer temperature than available boundary data, which causes the tracer can't freely radiate out. The adjustment on nudging timescale still not work (I tried TNUDG=360 with OBCFAC=12; TNUDG=360 with OBCFAC=120), FYI: I used monthly atmospheric data from ERA5 and monthly mercator ocean reanalysis for the boundary data. And my model domain has a highly seasonal cycle, where the change of inflow and outflow for every 6 months is very important to be captured.

Do you have any suggestion to solve this problem?
1. Is the use of sponge will help to diffuse the T and S at boundary?
2. Do I have to add the nudging layer to climatology data? Anyone can share the code to nudging both to boundary data and climatology value?

Or if there is any suggestion about this, maybe I missed some important part.

[kate]

I would start by upping the temporal frequency of your atmospheric fields. Monthly is going wash out all the storms. I use three-hourly to resolve diurnal cycles as well.

I haven't tried monthly boundary information since we stopped using monthly Levitus many, many moons ago, but going to more frequent boundary files has been helpful too, whether it is 5-daily SODA or daily HYCOM.

[corvianawatie]

Thanks for your suggestion, Kate. I tried to increase the frequency of atmospheric fields and it helps. Now, I am setting up to increase the frequency of boundary data.