Breaking: fix problems with ncep types#98
Conversation
|
@tiemvanderdeure if you have time to review this it would be really helpful. I was lax on the initial review and let in some slop. I think this is a much better structure but would be good to get a second opinion before making the breaking version. |
tiemvanderdeure
left a comment
tiemvanderdeure
left a comment
There was a problem hiding this comment.
Code looks good, just commented on docstring :)
| Data from the NCEP/NCAR Reanalysis 1 (1948–present) and NCEP/DOE Reanalysis 2 | ||
| (1979–present) datasets. |
There was a problem hiding this comment.
Apparently discontinued per the 9th of March 2026... Maybe we should mention this or just write 1948-2026?
There was a problem hiding this comment.
Lets just put 2026 instead of present
| - `Period`: `SixHour` (native, default), `Day`, or `Month`. | ||
|
|
||
| `NCEP{Surface, 2}` is the native 6-hourly surface analysis from Reanalysis 2; | ||
| `NCEP{Pressure, 1, Day}` is the daily Reanalysis 1 pressure-level aggregate. |
There was a problem hiding this comment.
This is at pressure levels right? So it would be a 3D dataset with pressure level steps? "pressure levels" in plural would be more correct, I'm not sure if the struct should also be PressureLevels, just because pressure sounds more like air pressure at ground level.
There was a problem hiding this comment.
Good point, yes PressureLevels is more correct
| Data from the NCEP/NCAR Reanalysis 1 (1948–present) and NCEP/DOE Reanalysis 2 | ||
| (1979–present) datasets. | ||
|
|
||
| See: https://psl.noaa.gov/data/gridded/data.ncep.reanalysis.html |
There was a problem hiding this comment.
And https://psl.noaa.gov/data/gridded/data.ncep.reanalysis2.html for Reanalysis 2
2 participants
There are some problems with NCEP. The "SixHourly" prefix on types is misleading. They are all six hourly by default, but have daily and monthly options as well in some cases.
Now we have the types
Surface,PressureandSurfaceFlux.The "gause" in
SurfaceGausewas a desicription of the drid not the data kind. Thats an implementation detail, the files have crs attached. The grouping described in the docs is surface fluxes.We also had dataset revision as a keyword. I've moved all of this to the NCEP type and made it consistent:
NCEP{Dataset,Revision,Period}Where period can be ommitted to mean the standard six hours:
NCEP{Surface,2}but you can doNCEP{SurfaceFluxes,1,Month}This is a breaking change because these types are already in the wild. @mrke