I'm working on implementing the overturning streamfunction variables (msftmz, msftyz, msftmrho, msftyrho) for CMIP7, and all four need a basin dimension (Atlantic/Arctic, Indo-Pacific, Global) to be compliant. That means we need a 2D basin mask to know which cells belong to which basin before summing over longitude.
These are not Core variables so this is low priority, but it would be good to get clarity on the mask before we get there
Maybe that is something we want though?
Current state in MOPPy:
msftmz and msftyz are mapped and use calc_overturning_streamfunction, which is algorithmically correct but computes the global basin only — no basin masking, and the output is missing the basin dimensionmsftmrho and msftyrho are in the mapping but have empty calculation blocks, so they're not implemented at all yet
APP4 used lsmask_ACCESS-OM2_1deg_20110618.nc for the basin mask, but ACCESS-ESM1.6 has a different ocean grid so that file can't just be reused.
I came across this forum post by Matt Chamberlain — it's about different variables (htovgyre, htovovrt etc.) but the same underlying need: MOM5 requires a basin_mask.nc to produce any basin-split output. He mentions the ESM1.5 version lives at:
/g/data/vk83/configuration/inputs/access-esm1p5/modern/share/ocean/basins/global.1deg/2020.05.19/basin_mask.nc
but was missing from ESM1.6. The thread was closed Feb 2026 by @ccarouge.
I'm working on implementing the overturning streamfunction variables (
msftmz,msftyz,msftmrho,msftyrho) for CMIP7, and all four need abasindimension (Atlantic/Arctic, Indo-Pacific, Global) to be compliant. That means we need a 2D basin mask to know which cells belong to which basin before summing over longitude.These are not Core variables so this is low priority, but it would be good to get clarity on the mask before we get there
Maybe that is something we want though?
Current state in MOPPy:
msftmzandmsftyzare mapped and usecalc_overturning_streamfunction, which is algorithmically correct but computes the global basin only — no basin masking, and the output is missing thebasindimensionmsftmrhoandmsftyrhoare in the mapping but have emptycalculationblocks, so they're not implemented at all yetAPP4 used
lsmask_ACCESS-OM2_1deg_20110618.ncfor the basin mask, but ACCESS-ESM1.6 has a different ocean grid so that file can't just be reused.I came across this forum post by Matt Chamberlain — it's about different variables (
htovgyre,htovovrtetc.) but the same underlying need: MOM5 requires abasin_mask.ncto produce any basin-split output. He mentions the ESM1.5 version lives at:but was missing from ESM1.6. The thread was closed Feb 2026 by @ccarouge.