# Data Sources — Petrel Ignition Density Layer v0.3

All input data used to train and build the v0.3 vintage, with citations and
licensing notes. Every v0.3 input is commercial-clean (public domain, CC-BY,
or free-reuse-with-attribution) — no share-alike sources are included.

## Anthropic-composer covariates

### 1. GHSL — Global Human Settlement Layer R2023A (`GHS-POP`, `GHS-BUILT-S`, `GHS-SMOD`)

Population count, built-up surface, and settlement classification grids from
the EC Joint Research Centre, 2020 epoch, 1 km.

- **Citation:** European Commission, Joint Research Centre — Global Human
  Settlement Layer Data Package 2023 (GHSL R2023A): GHS-POP, GHS-BUILT-S,
  GHS-SMOD, epoch 2020, 1 km. Publications Office of the European Union.
  https://ghsl.jrc.ec.europa.eu/
- **License:** Free reuse, including commercial, with attribution
  (European Commission / Copernicus open data policy)
- **Used as:** human-presence predictors — population density, built-up
  density, and the ordinal urban/suburban/rural settlement degree

### 2. HYDE 3.3 — History Database of the Global Environment

Cropland and grazing-land area per cell, baseline release, 2020 AD slice, 5′.

- **Citation:** Klein Goldewijk, K., Beusen, A., Doelman, J., & Stehfest, E.
  (2017). Anthropogenic land use estimates for the Holocene — HYDE 3.2.
  *Earth System Science Data*, 9, 927–953 (method reference; v0.3 uses the
  HYDE 3.3 baseline release). https://doi.org/10.5194/essd-9-927-2017
- **License:** CC-BY
- **Acquired:** 2026-07-09
- **Used as:** agricultural-ignition proxies (crop-residue and rangeland
  burning context)

### 3. Beck 2018 Köppen-Geiger climate classification (`KOPPEN`)

High-resolution present-day Köppen-Geiger climate classes, ~1 km.

- **Citation:** Beck, H. E., Zimmermann, N. E., McVicar, T. R., Vergopolan,
  N., Berg, A., & Wood, E. F. (2018). Present and future Köppen-Geiger
  climate classification maps at 1-km resolution. *Scientific Data*, 5,
  180214. https://doi.org/10.1038/sdata.2018.214
- **License:** CC-BY
- **Used as:** biome-context feature in the anthropic classifier, and the
  regionalization for the lightning ignition-efficiency calibration

## Lightning-composer inputs

### 4. NASA LIS/OTD 0.5° High Resolution Full Climatology (HRFC v2.3.2015)

Long-term mean lightning flash density (flashes km⁻² yr⁻¹) from the Optical
Transient Detector (OTD) and the TRMM Lightning Imaging Sensor (LIS).

- **Citation:** Cecil, D. J., Buechler, D. E., & Blakeslee, R. J. (2014).
  Gridded lightning climatology from TRMM-LIS and OTD: dataset description.
  *Atmospheric Research*, 135–136, 404–414. Dataset: LIS/OTD 0.5 Degree HRFC
  V2.3.2015, NASA Global Hydrometeorology Resource Center DAAC.
- **License:** Public (NASA data policy; distributed via NASA Earthdata)
- **Acquired:** 2026-07-09
- **Used as:** the flash-density climatology multiplied by the per-Köppen
  ignition efficiency

## Training labels

### 5. NASA FIRMS active-fire detections (MODIS + VIIRS)

Satellite active-fire detections, 2015–2023, reduced to gridded
fire-presence and lightning-attributable fire-density maps.

- **Citation:** NASA Fire Information for Resource Management System (FIRMS):
  MODIS Collection 6.1 and VIIRS 375 m active fire products. Giglio, L.,
  Schroeder, W., & Justice, C. O. (2016). The Collection 6 MODIS active fire
  detection algorithm and fire products. *Remote Sensing of Environment*,
  178, 31–41. https://doi.org/10.1016/j.rse.2016.02.054
- **License:** Public (NASA data policy)
- **Used as:** binary fire-presence labels for the anthropic classifier;
  fire-density numerator for the per-Köppen p(ignition|flash) calibration

## Geographic masks (post-processing)

### 6. GSHHG v2.3.7 full-resolution shoreline (`gshhg_f`)

Petrel's canonical land mask — the Wessel & Smith Global Self-consistent,
Hierarchical, High-resolution Geography database at full resolution
(~100 m native).

- **Citation:** Wessel, P., & Smith, W. H. F. (1996). A global,
  self-consistent, hierarchical, high-resolution shoreline database.
  *Journal of Geophysical Research*, 101(B4), 8741–8743.
- **License:** LGPL
- **Used as:** the coast clip applied to all three layers (`all_touched` —
  any cell touching real land keeps its value). GSHHG L1 only, so Antarctica
  is excluded by standard.

## Inputs deliberately excluded from v0.3

- **OpenStreetMap road presence** — considered as a human-access proxy but
  **excluded**: OSM's ODbL share-alike terms were unresolved for this
  distribution model, and the GHSL settlement/built-up inputs carry most of
  the same signal. No OSM-derived data is present in v0.3.

## Reproducibility

The exact source paths, model knobs, and promote configuration for v0.3 are
versioned in `configs/ignition_layer.yaml` in the Petrel ML repository. That
config is the source of truth — if anything here disagrees with it, the
config wins.