Skip to content

Lightning NOx

Overview

Computes lightning-produced nitrogen oxide (NOx) emissions based on convective cloud top height and empirical flash rate–yield relationships. Lightning is a significant natural source of NOx in the middle and upper troposphere, affecting ozone chemistry and atmospheric composition.

The flash rate is parameterized as a power law of cloud top height, with separate NOx yield factors for land and ocean regions.

References: - Price, C., et al. (1997), Vertical distributions of lightning NOx for use in regional and global chemical transport models, JGR, 102(D5), 5943–5941.

Registration Names

  • Native C++: "lightning"
  • Fortran bridge: "lightning_fortran"

Configuration Parameters

YAML Key Type Default Description
yield_land double 3.011e26 NOx yield factor for land regions [molecules/flash]
yield_ocean double 1.566e26 NOx yield factor for ocean regions [molecules/flash]
flash_rate_coeff double 3.44e-5 Flash rate coefficient
flash_rate_power double 4.9 Flash rate power-law exponent

Import Fields

Field Name Units Description
cloud_top_height m Convective cloud top height
land_mask dimensionless Land fraction (>0.5 = land); optional

Export Fields

Field Name Units Description
lightning_nox_emissions kg/s/grid_cell Lightning NOx production rate (3D, vertically distributed)

Algorithm

  1. Convert cloud top height to km: h_km = h / 1000.
  2. Compute flash rate: flash_rate = coeff * h_km^power.
  3. Select yield based on land mask (land vs. ocean).
  4. Compute total NOx yield: yield = (flash_rate * yield_factor) * (MW_NO / 1000) / (Avogadro * 1e6).
  5. Distribute uniformly across vertical levels: level_yield = total_yield / nz.

YAML Configuration Example

physics:
  - name: lightning
    config:
      yield_land: 3.011e26
      yield_ocean: 1.566e26
      flash_rate_coeff: 3.44e-5
      flash_rate_power: 4.9

Implementation Notes

  • Available as both native C++ (Kokkos) and Fortran bridge implementations
  • The land mask is optional; if absent, all cells are treated as land
  • Vertical distribution is currently uniform across all levels (Ott et al. proxy); a more sophisticated C-shape profile could be added
  • The scheme produces 3D output (emissions distributed across vertical levels)