Enable reading from Pierre's unified BGC dataset#274
Conversation
…o new-bgc-dataset
NoraLoose
changed the title
| return bgc_data | ||
|
|
||
|
|
||
| def compute_missing_surface_bgc_variables(bgc_data): |
There was a problem hiding this comment.
@abigale-wyatt, could you review the code for this function as well and check for the right units?
There was a problem hiding this comment.
@abigale-wyatt did you have a look at this function as well? I think the units are wrong (see my comments in this function).
abigale-wyatt
left a comment
abigale-wyatt
left a comment
There was a problem hiding this comment.
I added values for NH4 and DOC. The units on everything are correct.
Awesome, thanks so much. I addressed your comments in the commit d79d0da |
|
@abigale-wyatt I left one more question for you in the code. |
|
@abigale-wyatt I will merge this PR to avoid future merge conflicts, and open a separate issue to get the default values of NOx and NHy right. |
2 participants
This PR enables reading from Pierre's "unified" BGC dataset for the classes
InitialConditionsBoundaryForcingSurfaceForcingSince the unified BGC dataset does not contain all necessary MARBL BGC variables, the missing BGC variables need to be filled in (see #273).
Internal changes
InitialConditions,BoundaryForcing,SurfaceForcingto accommodate optional variable names (which the unified BGC dataset requires)UnifiedDatasetclass (+ two subclasses,UnifiedBGCDatasetandUnifiedBGCSurfaceDataset) with the required post-processing stepsDatasetclass:choose_subdomainmethod to handle concatenation in longitude direction across both ends (upper + lower) correctlyneeds_lateral_filltoDatasetclass, which is set toTrueby default. Pierre's unified BGC dataset has land values already filled in, so we setUnifiedDataset.needs_lateral_fill = FalseDatasetto avoid duplicationChecklist
pre-commit run --all-filesdocs/releases.mddocs/api.rst