Accessing fully interpolated atmospheric forcing

[suhanidalal]

Hi,

I actually asked a similar question a few months back, but I’m running a global Oceananigans simulation (loosely based on this example). I am computing wind speed and gas transfer velocity, so I need to access the JRA55 wind components and x/y momentum fluxes. This is how I currently do it:

u10 = coupled_model.interfaces.exchanger.exchange_atmosphere_state.u 
v10 = coupled_model.interfaces.exchanger.exchange_atmosphere_state.v 
taux = coupled_model.interfaces.atmosphere_ocean_interface.fluxes.x_momentum
tauy = coupled_model.interfaces.atmosphere_ocean_interface.fluxes.y_momentum

Unlike ocean fields such as T, S, and u/v, these outputs are not fully interpolated onto the ocean grid after I run my simulation. Is there an updated way to access the fully interpolated wind and flux fields?

Thank you!

[glwagner]

I thought those were the interpolated results so my understanding may be outdated, cc @simone-silvestri

We need to have docs on this. Right after we finish the metadata docs in #587

[suhanidalal]

My objective is to calculate wind speed and gas transfer velocity using these fields. Currently, when I access and plot these fields (even using FieldTimeSeries), the halo/boundary cells are zeros, so the field is not defined across the full domain. I want values across the entire grid, similar to how U, V, and T are populated. I’m wondering whether this is due to how I accessed the fields during the simulation (which doesn't seem to be the case) or how I’m plotting them. Thanks!

[glwagner]

The cells that are not populated are halo regions and should not be used for computations, right? You can try using fill_halo_regions! if you absolutely need to, but I am concerned about using halo regions for computations unless it is done very intentionally.

[simone-silvestri]

Also note that the field you are showing is on the ocean grid (the tripolar grid)

[suhanidalal]

How are salinity, temperature, and the velocity fields different in terms of their spatial coverage? I was hoping to have atmospheric forcing values (like τx and τy) defined across the entire ocean domain, similar to how salinity, temperature, and velocity appear. The attached image from above (τy) shows where the left and right edges (over ocean) have zeros.

[glwagner]

Can you include your whole script so that we can inspect it to see what is going wrong?