Generate code & zarr collections to make BRAN2020 U & V "analysis ready" for CARSv2

[Thomas-Moore-Creative]

Generate code & zarr collections to make BRAN2020 U & V "analysis ready" for CARSv2

U & V will be needed for using DIVAnd's "advection constraint.

Advection constraint

The functions DIVAndrun, DIVAndgo and diva3d can also use an advection constraint forcing the analysis to align with a vector field (e.g. a velocity field). The velocity field should be a tuple of n-elements. Every element of the tuple is a gridded array (defined at the same location than the target array) representing a single velocity component. For 3D analysis, the order of the dimensions is typically: longitude, latitude and depth. Like-wise the velocity components are zonal, meridional and vertical velocity. The three velocity components has to be scaled by a constant factor to enhance or decrease this constraint. It is recommended that this parameter is tuned by cross-validation. There are no tools currently in DIVAnd.jl to automate this process.

For the two dimensional case, the velocity has just two components

• [x] generate zarr collection.
• [x] calculate climatology.
• [x] export climatology as zarr & NetCDF.
• [x] rsync data to tube.
• [ ] write intake catalogue for BRAN2020
• [ ] clean up workflow, improve metadata, and document transfer approach.
• [ ] write some supporting docs to access the climatology.

[Thomas-Moore-Creative]

output zarr, rechunked, and calculated climatology

[Thomas-Moore-Creative]

regional, 4D zarr collection on NCI scratch: /scratch/xv83/tm4888/zarr_collections/BRAN2020/AUSregion.BRAN2020.month.UV.zarr

Size = 30GB

[Thomas-Moore-Creative]

regional climatology of BRAN2020 U & V currents size ~ 2GB

/scratch/xv83/tm4888/zarr_collections/BRAN2020/AUSregion.BRAN2020.climatology.UV.zarr

/scratch/xv83/tm4888/zarr_collections/BRAN2020/AUSregion.BRAN2020.climatology.UV.nc

[Thomas-Moore-Creative]

@ChrisC28 & @BecCowley I still need to document things better and possibly automate some things but as a start I've got a regional subset of both BRAN2020 u & v as well as the climatology over on Tube /oa-decadal-climate/work/observations/CARSv2_ancillary/BRAN2020

https://github.com/CARSv2/cars-v2/commit/325e5203e5911e2401aff9c89c106bf1f0abff1a

[ChrisC28]

Thanks @Thomas-Moore-Creative . I'll push the cleaned Julia code to github to perform the mapping today or tomorrow and then we can sit down an workout how we might get the velocities into DIVAnd.

[ChrisC28]

OK, I've pushed the updated "working" (term used partially in jest) example to the repo. Apologies for the delay (struggling with the token security).

[Thomas-Moore-Creative]

I believe @ChrisC28 will be using the updated climatological currents coming from #24 - so closing this here.