External Emission Data Process - Integration Guide¶

Overview¶

The External Emission Data Process is a special type of process in CATChem that handles external emission data management rather than computing emissions directly. It provides a standardized interface for loading, validating, interpolating, and accessing external emission inventories from files.

Key Differences from Regular Emission Processes¶

1. Purpose¶

Regular Emission Processes: Compute emissions based on meteorological conditions, chemical concentrations, and parameterizations
External Emission Data Process: Manages pre-computed emission data from external inventories (e.g., EDGAR, CEDS, regional inventories)

2. No Traditional "Schemes"¶

Instead of computational schemes, this process has data handler schemes:
surface_2d: For 2D surface emission data
elevated_3d: For 3D elevated emissions with vertical distribution
temporal_profiles: For time-varying emission profiles and scaling factors

3. File I/O Delegation¶

File reading is delegated to the driver, not handled directly by the process
The process provides data structures and validation for loaded data

Architecture¶

Core Components¶

ExtEmisData_Mod.F90 (existing core module)
ExtEmisDataType: Main container for external emission data
ExtEmisCategoryType: Groups emissions by source category
ExtEmisFieldType: Individual emission fields
Generated Process Files
ProcessExternalEmissionDataInterface_Mod.F90: Main process interface
ExternalEmissionDataCommon_Mod.F90: Common data structures
ExternalEmissionDataProcessCreator_Mod.F90: Factory for creating process instances
Data Handler Schemes
Surface2D: Handles 2D surface emission data
Elevated3D: Handles 3D elevated emissions with vertical distribution
TemporalProfiles: Manages temporal scaling and profiles

Integration with Existing ExtEmisData_Mod¶

The generated process works alongside the existing ExtEmisData_Mod.F90 module:

! The existing module provides the core data structures
USE ExtEmisData_Mod, only: ExtEmisDataType, ExtEmisCategoryType, ExtEmisFieldType

! The new process provides the process interface
USE ProcessExternalEmissionDataInterface_Mod, only: ProcessExternalEmissionDataInterface

Usage Patterns¶

1. Driver Integration¶

! Driver loads external emission files
call driver%load_emission_files(emission_files, ext_emis_data, rc)

! Process validates and manages the data
call external_emission_process%initialize(config, ext_emis_data, rc)
call external_emission_process%validate_data(rc)

2. Emission Process Integration¶

! Other emission processes can access external data
emission_rate = ext_emis_data%get_emission_rate('NOx', i, j, k)

! Or get column data for vectorized operations
column_ptr => ext_emis_data%get_column_ptr('anthropogenic', 'CO', i, j)

3. Temporal Interpolation¶

! Update time indices for all external data
call ext_emis_data%update_time(new_time_idx, rc)

! Access temporally interpolated data
call temporal_profile_scheme%apply_scaling(base_emissions, scaled_emissions, rc)

Separation of Concerns¶

Current Implementation¶

The current approach combines external emission data management with computational emission processes, which can lead to: - Mixed responsibilities in emission modules - Difficulty in testing and validation - Coupling between file I/O and emission calculations

With External Emission Data Process¶

Data Management: Handled by the external emission data process
Computational Emissions: Handled by dedicated emission processes (biogenic, dust, etc.)
File I/O: Delegated to the driver
Validation: Standardized across all external data sources

Future Extensions¶

1. Separate Processes for Specialized Emissions¶

Elevated Emissions Process¶

Handle point source emissions with stack parameters
Integrate with plume rise calculations
Manage temporal profiles for industrial sources

Plume Rise Process¶

Calculate effective emission heights
Handle meteorological dependencies
Integrate with 3D emission distribution

# Future plume_rise_config.yaml
name: plume_rise
class_name: PlumeRise
description: Plume rise calculations for elevated point sources
process_type: emission
schemes:
  - name: briggs
    description: Briggs plume rise formulation
  - name: holland
    description: Holland plume rise formulation

2. Enhanced Data Management¶

Species Mapping¶

Map inventory species to model species
Handle speciation profiles
Support multiple chemical mechanisms

Spatial Processing¶

Regridding and interpolation
Coordinate system transformations
Boundary condition handling

Configuration Example¶

# external_emission_data_config.yaml
name: external_emission_data
schemes:
  - name: surface_2d
    parameters:
      interpolation_method: bilinear
      time_interpolation: true
      missing_value_handling: zero_fill

  - name: elevated_3d
    parameters:
      vertical_interpolation: true
      plume_rise_coupling: false

categories:
  - name: anthropogenic
    source_files:
      - EDGAR_v6_CO_2015.nc
      - EDGAR_v6_NOx_2015.nc
    scaling_factor: 1.0

  - name: biogenic
    source_files:
      - MEGAN_isoprene_2015.nc
    temporal_profiles: true

Testing Strategy¶

Unit Tests¶

Data structure validation
Interpolation algorithms
Memory management
Error handling

Integration Tests¶

Driver integration
Process factory creation
Multi-category data loading
Time interpolation accuracy

Performance Tests¶

Memory usage optimization
Large file loading
Spatial interpolation performance
Column access patterns

Best Practices¶

Modular Design: Keep data management separate from emission calculations
Validation: Always validate external data before use
Memory Management: Monitor memory usage for large datasets
Error Handling: Provide clear error messages for file I/O issues
Documentation: Document data sources, units, and processing steps