gridded files updates

[gidden]

all

• [x] SSD still seems to be missing (verify after rerunning on 8-16-2023)
• [x] check which implementation is correct
  • [x] separate / separate.mean(self.mean_time_dims).sum(self.sum_spatial_dims))
  • [ ] ~separate / separate.sum(self.mean_time_dims + self.sum_spatial_dims))~
• [ ] translate year/month to datetime
• [ ] remove gas, method, unit dims
• [ ] add dataset attributes
• [ ] add variable attributes
• after discussion with BCS
  • [ ] change fname of global sectors to 'short name' (aircraft, shipping)
  • [ ] replace variable names "-" with "_"
  • [ ] change fname of openburning shares to drop 'em' and replace 'shares_emissions' -> 'share'
  • [ ] add bounds dimensions and bound variables
  • [ ] when viewing the ncs, it seems like lat dim is actually cut off (even though it is in the full dataset)

<gas>-em-anthro

• after discussion with BCS
  • [ ] missing dim values: Agriculture for CO2
  • [ ] missing dim values: Shipping (decide to keep shipping sector different or not; this should now be trivial with 'World' added to iso index raster)
  • [ ] translate sector dim to integer (0: Agriculture; 1: Energy; 2: Industrial; 3: Transportation; 4: Residential, Commercial, Other; 5: Solvents production and application; 6: Waste; 7: International Shipping; 8: Negative CO2 Emissions) (decide if needed or not)

<gas>-em-openburning

• [x] missing dims: sector, year
  • [x] sector likely should be missing since it is in the shares - but need to check with final ScenarioMIP data
  • [x] year definitely should still be there, though

<gas>-openburning-share

• [x] seems to have all dims, but needs to be checked
• after discussion with BCS
  • [ ] translate sector dim to integer ( 0: Agricultural Waste Burning On Fields; 1: Forest Burning; 2: Grassland Burning; 3: Peat Burning) (decide if needed or not)

<gas>-em-aircraft-anthro

• [x] aircraft level values are likely causing differences (do we include level in the normalization or not?) - note this could simply be differences in harmonized data values
  • confirmed that global values between new REMIND and ScenarioMIP scenarios are exactly the same bulk emissions (0.013 in 2020)
  • discovered the issue which is that global data was not unit converted in workflow.ipynb
  • confirmed we are within 1.5% when correcting for units

<gas>-em-shipping-anthro

• [x] difference in values observed note this could simply be differences in harmonized data values
  • discovered the issue which is that global data was not unit converted in workflow.ipynb

CO2-em-removal

• seems good, CDR is negative, emissions are positive
• [ ] update proxy maps with distance to CCS + other drivers

cc: @coroa

[coroa]

translate year/month to datetime

In [34]: def convert_to_datetime(da):
...:     da = da.stack(time=("year", "month"))
...:     dates = pd.DatetimeIndex(da.indexes["time"].map(lambda t: date(t[0], t[1]+1, 1)))
...:     return da.drop_vars(["time", "year", "month"]).assign_coords(time=dates)

[gidden]

Closing in favor of more targeted issues of #10, #11, and #12