Methods Notes

These notes follow the manuscript's scientific argument: first construct a common response from the ordered history of observed growth, then project candidate explanations onto it. The corrected v2 methods and audit are authoritative for quantitative claims. Current shape-only PCA, dated transitions, nested same-cohort models, and matched pairs are generated in analysis/v2/. Historical lakehouse tables remain read-only source inputs, not corrected traits.

These notes connect the manuscript methods to the scripts and data products needed for reproduction.

Workflow Overview

1. FIRED event and observation records define the observed fire population.
2. Mapped area increments are ordered and their date support and gaps audited.
3. VASEs encode normalized cumulative area through relative developmental time.
4. Shape coordinates and traits summarize timing, concentration, persistence, recurrence, and taper without using weather to define the axes.
5. gridMET weather is attached afterward as a first external explanatory layer; the corrected analysis uses day-specific exposure where required.
6. State models and matched comparisons ask when exposure adds information and where the present explanation remains insufficient.
7. Stability, null, uncertainty, and provenance checks bound every manuscript claim.

Data Products

Core schemas are stored in schemas/.

The main lakehouse tables are:

• fire_catalog.parquet
• fire_traits.parquet
• vase_slices.parquet
• vase_climate_exposures.parquet
• processing_manifest.parquet
• processing_failures_climate.parquet

Developmental morphology products are written under:

scratch/fire_vase_developmental_morphology/

The figure validation tables are stored under:

figures/main/derived_stats/
analysis/claim_audit_stats/
analysis/climate_revision_stats/

Reproducible Methods Code

• Source/cache preparation: cache_gridmet_years.py
• Fire VASE lakehouse: fire_vase_lakehouse_pilot.py
• Climate attribution: fire_vase_build_climate_tables.py and fire_vase_build_perimeter_climate_tables.py
• Developmental morphology: fire_vase_developmental_morphology_analysis.py
• Manuscript revision analysis: fire_vase_climate_revision.py
• Manuscript figure rendering: manuscript_figures

Interpretation Boundary

The VASE is a standardized observed history, not a geographic silhouette, a natural fire type, or a process model. Weather associations are retrospective and do not establish causal control. Mismatch examples generate testable cases for additional layers; they do not identify the omitted mechanism or estimate its ecological prevalence.

External Data

The manuscript uses FIRED fire records and gridMET daily climate data. See the FIRED dataset note for local cache and citation guidance.

The shared Fire VASE data lake is available through the CyVerse Fire_Vase folder.