COAWST - ROMS+WRF Coupled Strange BIAS Atlantic

[william248]

Hello everyone, I'm doing a coupled run of the ROMS with the WRF and I noticed that when I do the coupling, a very strange BIAS appears, practically fixed in the Atlantic. I send the attached image and files. What do you think it could be? Any configuration problem in physics?

This bias persists throughout the model integration time (40 days), with a gradient of almost 10 degrees in temperature at 2 meters that does not change. In rounds without coupling this does not appear.

https://ibb.co/9yDs0BD

[jcwarner]

not sure what i am looking at here.
what does the SST in roms look like?

[william248]

not sure what i am looking at here.
what does the SST in roms look like?

this is the temp variable in the outputs? Because i can only see potential temperature in the history file.

Thanks!

[jcwarner]

yes i understand what T2 is.
I see an image, with lots of colors, i dont know geographically where it is.
I dont know where the 'anamoly' is, etc. Do the temp issue appear right away, is it on land, on the ocean,
I do not have enough info to understand the issue.

[william248]

yes i understand what T2 is.
I see an image, with lots of colors, i dont know geographically where it is.
I dont know where the 'anamoly' is, etc. Do the temp issue appear right away, is it on land, on the ocean,
I do not have enough info to understand the issue.

Sorry for not being more detailed about the problem, jcwarner.

The Study region is between 10N and -10S, over the oceanic region of the Atlantic near the coast of Brazil. I verified that the sensible and latent heat fluxes are very strange over this region, so I believe that I may have forgotten to activate some flag in the set.h file attached above. The anomaly would be over this oceanic region of the Atlantic in red in the figure below showing unreal values over this region. I don't know what could be happening, I believe it's something I forgot to activate in the flags.

https://ibb.co/4ffzFNc

[jcwarner]

ok, that figure looks better and i can see what you are talking about.
Suggest you look at the different components of the heat budget: lwrad, swrad, latent, sens, and compare the coupled to uncoupled of those 4 terms.
Try undef ANA_SRFLUX, although i dont think that is the problem.
i am not familiar with the LMD mixing options, maybe try GLS?
Is this an older version of code? looks like the cppdefs are older style.
At some point we modified the way that ROMS does the heat flux computations, there was nothing wrong, it was just the way the arrays were keeping track of the total heat flux.
We have updated all of that and maybe you should try an updated version, so that i can help better.
if you do an update, then undef ANA_SSFLUX. this was necessary in the older way, but not needed now.
see the Sandy test case, but also undef the ana_srflux.

[william248]

ok, that figure looks better and i can see what you are talking about.
Suggest you look at the different components of the heat budget: lwrad, swrad, latent, sens, and compare the coupled to uncoupled of those 4 terms.
Try undef ANA_SRFLUX, although i dont think that is the problem.
i am not familiar with the LMD mixing options, maybe try GLS?
Is this an older version of code? looks like the cppdefs are older style.
At some point we modified the way that ROMS does the heat flux computations, there was nothing wrong, it was just the way the arrays were keeping track of the total heat flux.
We have updated all of that and maybe you should try an updated version, so that i can help better.
if you do an update, then undef ANA_SSFLUX. this was necessary in the older way, but not needed now.
see the Sandy test case, but also undef the ana_srflux.

Thank you Warner! The problem is that I was running the roms alone first since 2012 and then coupling to WRF in 2017. I tried the above options but I keep having this bias occurring.

[jcwarner]

Can you look at the different components of the heat budget: lwrad, swrad, latent, sens, and compare the coupled to uncoupled of those 4 terms?