Emissions

The emissions to be transported are defined in the emissions section of the configuration file. The section is structured as follows: 

emissions :
    tracer : [tracer_i, tracer_j, ...] # List of tracers 
    tracer_i:
        region :        # Grid definition for tracer "tracer_i" 
        interval :      # transport model temporal resolution
        categories :
            cat1 :      # "origin" of the emission files for cat "cat1"
            cat2 :
                origin : # "origin" of the emission files for cat "cat2"
                ...    : # other settings, speficic to cat2
            ...
        metacategories :
            metacat1 : 
            metacat2 :
        ...            :  
    tracer_j:
        ...

This section of the configuration file is essentially used to construct a single LUMIA emission file, containing the emissions for the simulation (i.e. emissions of all categories and all tracers, covering the entire length of the simulation), based on category and tracer-specific pre-processed emission files (see File formats section below).

The yaml file keys determine the following settings:

  • Mandatory keys:
    • list of emission tracers: emissions.tracers key (should be a list!)
    • path of the pre-processed files to be read: path, prefix and origin keys.
    • grid definition of the emission files : emissions.{tracer}.region key
    • temporal resolution of the emissions: emissions.{tracer}.interval key
  • Optional keys:
    • name of the netCDF variable to be read: field key
    • whether the pre-processed emission files should be resampled from a different temporal resolution: resample_from key.
    • path to a rclone-mounted emission archive: archive key
    • definition of metacategories: emissions.{tracer}.metacategories key(s)

The path, prefix, field, resample_from and archive keys can be provided at the tracer level (e.g. emissions.{tracer}.path) or at the category level (emissions.{tracer}.categories.{catname}.path). The origin key should be provided at the category level (emissions.{tracer}.categories.{catname}.origin), but if all the other keys are provided at the tracer-level, it is possible to simply use emissions.{tracer}.categories.{catname} as an "origin" key.

Pre-processed emission files

The assembling of pre-processed, category-specific emission files into a single, simulation-specific emission file is performed by the lumia.Data.from_dconf method.

The files should contain three coordinate variables (time, lat and lon), one (or more) emission fields defined on the same coordinates. It is also recommended to include a (lat, lon) area field for convenience, but this is not read or required by LUMIA. The lat and lon variables should refer to the center of the grid cells, while the time coordinate points to the start of each time period. The time coordinate should contain integer, and have a units and a calendar attribute, allowing conversion to a numpy.datetime64 type. The recommended method for creating these files is to use the xarray library.

The path of the pre-processed files is determined by the path, prefix and origin keys: The file names follows the pattern {path}/{tres}/{prefix}{origin}.*.nc. Here prefix is meant as a tracer-spefic prefix (e.g. "co2_flux") and origin refers to the origin of the data in the file (e.g. "LPJ"). The {tres} key refers to the temporal resolution of the pre-processed emission files. By default it is identical as emissions.{tracer}.interval, but can be set at a lower temporal resolution using the resample_from key (in which case the emissions will simply be rebinned by LUMIA).

For instance, if a biosphere category should read emissions from /data/LUMIA/M/co2_emis.LPJGUESS-v20.2018.nc, then:

  • emissions.co2.path (or emissions.co2.categories.biosphere.path) should be set to /data/LUMIA
  • emissions.co2.prefix (or emissions.co2.categories.biosphere.prefix) should be set to co2_emis.
  • emissions.co2.categories.biosphere.origin (or emissions.co2.categories.biosphere) should be set to LPJGUESS-v20
  • if emissions.co2.interval is not set to M, then a emissions.co2.resample_from (or emissions.co2.categories.biosphere.resample_from) key should be defined and set to M.
  • if the file contains any other variable than the coordinates (lat, lon, time) and the variable containing the emission themselves, then a emissions.co2.categories.biosphere.field variable should be set.

Note that at no point we specify the time component of the filename: LUMIA will load all the files matching the pattern {path}/{prefix}{origin}.*.nc as a multi-file netCDF Dataset.

Archive

The emission files need to be on a local file system for LUMIA to read them, however that can be just a temporary folder, with the files being stored on a remote permanent storage, accessed via rclone. The path to the archive should be provided by the emissions.{tracer}.archive key, with the pattern rclone:{rclone_remote}:path/to/remote/dir, where {rclone_remote} is a rclone remote path defined in your rclone.conf file (see the relevant rclone documentation).

Meta-categories

It is possible to define meta-categories, as a linear combination of other categories. For instance in the (fictional) example below, we defined one "NEE" meta-category, as the sum of the GPP and respiration categories. We defined a "fossil" category by subtraction of "agri_fires" from the "anthropogenic" category, and we created a "fires" category combining the "natural_fires" one and the "agri_fires" one, but the latter with a 1.2 scaling factor.

By default, the meta-categories are not transported (i.e. the transport model is unaware of them), but their impact on the concentrations is calculated. On the other hand, the categories used to build the meta-categories are transported, but their impact on concentrations is ignored.

emissions :
    co2 :
        categories:
            GPP : LPJ_GPP
            respiration : LPJ_resp
            anthropogenic : TNO
            natural_fires : GFAS
            agri_fires : 
                origin : EDGAR
                field : agriwasteburning
        meta-categories :
            NEE : GPP + respiration
            fossil : anthropogenic - agri_fires
            fires : natural_fires + 1.2 * agri_fires

LUMIA emission file

raise SectionYetToBeWrittenError ;-)