Notes on emiprep architecture

This section contains notes on the architecture of emiprep.

The general emiprep architecture is best captured by this mind map (you’ll need the XMind application to edit the file):

Warning

This mind map is out-of-date!

The actual flow should be something like

1. For each emission dataset:
   1. read data
   2. geo-reference emission source
   3. vertical distribution
   4. unit conversion
   5. species split
   6. geographical interpolation to model grid
2. For each time step:
   1. For each emission dataset:
      1. Apply time factors
   2. Merge inventories / Aggregate sectors
   3. Write output file

Following are more detailed discussions about specific choices made in the emiprep architecture:

• Reading grid information from WRF files
• Re-gridding to WRF grids
  • Deciding which external interpolation library to use
  • Using CDO for regridding