openairclim.core.attribution

Provides attribution functionality for non-linear species.

class openairclim.core.attribution.AttributionFunc(*args, **kwargs)[source]

Bases: Protocol

Defines type protocol for attribution function inputs.

openairclim.core.attribution.apply_attribution(func: AttributionFunc, diff_func: AttributionFunc, method: str, species: str, sub_dict: dict[str, ndarray], full_dict: dict[str, ndarray], **kwargs: Any) → dict[str, ndarray][source]

Applies attribution methodology using the function func.

Dispatches to the attribution function matching method for sub_dict.

Parameters:

  • func (AttributionFunc) – A callable that maps a dict of species arrays to another such dict. The first argument must be a dict[str, np.ndarray].

  • diff_func (AttributionFunc) – A callable that calculates the derivative of func with respect to the input. Must have the same shape as func. Can be None if no attribution method is used that requires a derivative.

  • method (str) – Attribution method. Choice of: “none”, “residual”, “marginal”, “proportional”, “differential”.

  • species (str) – Name of the species to consider (e.g. ‘CO2’).

  • sub_dict (dict[str, np.ndarray]) – Dict of species arrays representing a part of the full dict (e.g. one aircraft identifier).

  • full_dict (dict[str, np.ndarray]) – Dict of species arrays representing all sub-dicts. Must include sub_dict!

  • **kwargs – Additional keyword arguments to pass to func and diff_func.

Raises:

ValueError – If an invalid method is given.

Returns:

func value attributable to sub_dict

Return type:

dict[str, np.ndarray]

openairclim.core.attribution.differential_attribution(diff_func: AttributionFunc, sub_dict: dict[str, ndarray], full_dict: dict[str, ndarray], species: str, **kwargs: Any) → dict[str, ndarray][source]

Calculates the func value attributable to sub_dict (differential).

Uses a differential methodology for species species. The dict full_dict, representing all other sources (i.e. background plus other aircraft identifiers), must include sub_dict.

Parameters:

  • diff_func (AttributionFunc) – A callable that calculates the derivative of func with respect to the input. Must have the same shape as func.

  • sub_dict (dict[str, np.ndarray]) – Dict of species arrays representing a part of the full dict (e.g. one aircraft identifier).

  • full_dict (dict[str, np.ndarray]) – Dict of species arrays representing all sub-dicts. Must include sub_dict!

  • species (str) – Name of the species to consider (e.g. ‘CO2’).

  • **kwargs – Additional keyword arguments to pass to func and diff_func.

Returns:

func value attributable to sub_dict

Return type:

dict[str, np.ndarray]

openairclim.core.attribution.marginal_attribution(diff_func: AttributionFunc, sub_dict: dict[str, ndarray], full_dict: dict[str, ndarray], species: str, **kwargs: Any) → dict[str, ndarray][source]

Calculates the func value attributable to sub_dict (marginal).

Uses a marginal methodology for species species. The dict full_dict, representing all other sources (i.e. background plus other aircraft identifiers), must include sub_dict.

Parameters:

  • diff_func (AttributionFunc) – A callable that calculates the derivative of func with respect to the input. Must have the same shape as func.

  • sub_dict (dict[str, np.ndarray]) – Dict of species arrays representing a part of the full dict (e.g. one aircraft identifier).

  • full_dict (dict[str, np.ndarray]) – Dict of species arrays representing all sub-dicts. Must include sub_dict!

  • species (str) – Name of the species to consider (e.g. ‘CO2’).

  • **kwargs – Additional keyword arguments to pass to func and diff_func.

Returns:

func value attributable to sub_dict

Return type:

dict[str, np.ndarray]

openairclim.core.attribution.proportional_attribution(func: AttributionFunc, sub_dict: dict[str, ndarray], full_dict: dict[str, ndarray], species: str, **kwargs: Any) → dict[str, ndarray][source]

Calculates the func value attributable to sub_dict (proportional).

Uses a proportional methodology for species species. The dict full_dict, representing all other sources (i.e. background plus other aircraft identifiers), must include sub_dict.

Parameters:

  • func (AttributionFunc) – A callable that maps a dict of species arrays to another such dict. The first argument must be a dict[str, np.ndarray].

  • sub_dict (dict[str, np.ndarray]) – Dict of species arrays representing a part of the full dict (e.g. one aircraft identifier).

  • full_dict (dict[str, np.ndarray]) – Dict of species arrays representing all sub-dicts. Must include sub_dict!

  • species (str) – Name of the species to consider (e.g. ‘CO2’).

  • **kwargs – Additional keyword arguments to pass to func.

Returns:

func value attributable to sub_dict

Return type:

dict[str, np.ndarray]

openairclim.core.attribution.residual_attribution(func: AttributionFunc, sub_dict: dict[str, ndarray], full_dict: dict[str, ndarray], species: str, **kwargs: Any) → dict[str, ndarray][source]

Calculates the func value attributable to sub_dict (residual).

Uses a residual methodology for species species. The dict full_dict, representing all other sources (i.e. background plus other aircraft identifiers), must include sub_dict.

Parameters:

  • func (AttributionFunc) – A callable that maps a dict of species arrays to another such dict. The first argument must be a dict[str, np.ndarray].

  • sub_dict (dict[str, np.ndarray]) – Dict of species arrays representing a part of the full dict (e.g. one aircraft identifier).

  • full_dict (dict[str, np.ndarray]) – Dict of species arrays representing all sub-dicts. Must include sub_dict!

  • species (str) – Name of the species to consider (e.g. ‘CO2’).

  • **kwargs – Additional keyword arguments to pass to func.

Returns:

func value attributable to sub_dict

Return type:

dict[str, np.ndarray]