Two-Way Nesting with Different Wind Resolutions in WW3

[AlexCrls]

Hi,

I was running the model with two-way nesting and I have a question.
Does WW3 support using two different wind input datasets? Specifically, I want to run the parent grid using 9 km wind data and the nested grid using 3 km wind data, both generated by WRF.

If that's possible, what is the correct procedure to follow?

My current understanding is that I need to run ww3_prnc separately for each wind field (one for each grid), but I'm unsure how to indicate in ww3_multi that each grid should use its respective wind input, especially if the files have the same name. Is there a way to assign them properly, or is this approach not supported?

Best regards,
Alejandro

[mickaelaccensi]

Hi,

you can do it, you need to preprocess the 9km wind using the mod_def of your 9km grid and rename wind.ww3 to wind.9km.
then preprocess the 3km wind using the mod_def of your 3km grid and rename wind.ww3 to wind.3km
Once you have your 2 wind files preprocessed, run ww3_multi and specify 'native' for the wind forcing for your 9km and 3km grids

Best,

Mickael

[AlexCrls]

Hi Mickael

Thanks for your support!

I managed to run the model successfully using both preprocessed wind fields (9km and 3km).
However, when I generate the NetCDF output using ww3_ounf, I notice something unexpected:

In the parent domain output, the child domain appears overlaid.

In the child domain output, there is a visible strip or overlap where the parent domain seems to intrude.

Have you seen this behavior before?
Do you know what might be causing it, or how to prevent it?

Best regards,

[AlexCrls]

I attached an example to illustrate what I’m trying to explain.

In the parent domain, the child domain appears overlaid and clearly visible.

In the child domain, you can see a visible strip at the top where it overlaps with the parent

Best regards,

Alejandro

[mickaelaccensi]

you can mask the parent grid with option in ww3_multi.nml: DOMAIN%FLGHG2 = T

if your wind forcings are too different, it can produce inconsistencies which is apparently your case

[AlexCrls]

Hi, thanks for your response!

I'm using the ww3_multi.inp file (not .nml), and I found this line that seems related to what you mentioned:

$ Specific multi-scale model settings (single line).
$ Flag for masking computation in two-way nesting (except at
$ output times).
$ Flag for masking at printout time.
$
F F

1. Is DOMAIN%FLGHG2 = T in the .nml file equivalent to setting T T in the .inp?

Regarding the wind fields, both come from the same WRF run but at different resolutions (9 km and 3 km). Is it necessary for all grids to use the same wind field to avoid issues at the transition zone?
If using different resolutions is allowed, can WW3 handle the transition internally, or should some preprocessing (e.g., smoothing at the boundaries) be done beforehand?

Thanks again for your help!

Best,
Alejandro

[mickaelaccensi]

you have different solutions :
-running with ww3_multi with the same 9km wind
-running with ww3_multi with 9km and 3km wind
-running with ww3_shel your 9km grid with 9km wind and output spectral boundaries for the 3km domain
-then run ww3_shel on the 3km grid with 3km wind and spectral boundaries from the 9km grid (ww3_bounc)

about your issue, you should verify the consistency between your 9km and 3km wind at several time steps or try yo reduce the 3km domain to see if it solve the issue at domain boundaries