Compare the evidence.
Not just the fire.
A structured comparison of what is strongly documented, what remains disputed and what each incident teaches about wildfire, evacuation, urban exposure and the limits of available evidence.
Five fires. Different evidence questions.
Each pilot stresses a different part of the Wildfire Black Box architecture: conflicting official records, chronology, evacuation scale, multi-fire scope or changing review and cause information.
Palisades Fire
Lahaina Fire
Fort McMurray
Black Saturday
Eaton Fire
Same framework. Different evidence.
Dates, units, geographic scope and source definitions travel with every value. Unlike measurements are not silently converted into false equivalents.
| DIMENSION | PALISADES 2025 | LAHAINA 2023 | FORT McMURRAY 2016 | BLACK SATURDAY 2009 | EATON 2025 |
|---|---|---|---|---|---|
| Record focus | Final incident facts, impacts, weather and changing official information | Origin, two-phase chronology, urban progression and systems analysis | Mass evacuation, WUI transition, satellite footprint and recovery monitoring | Multi-fire disaster scope, fire weather, impacts and policy legacy | Nighttime fire behaviour, West Altadena alerts, two official reviews and cause-state revision |
| Primary observation layer | NASA satellite smoke context and agency incident data | Landsat active-fire observation plus a multi-source official chronology | Landsat and MODIS observations at distinct dates and scales | Bureau of Meteorology observations plus Royal Commission fire evidence | CAL FIRE incident snapshots, County radio and alert records, and NASA AVIRIS/FEDS products |
| Main source conflict | Different official structure damage and destruction totals | Minute-level differences in official phase timestamps | More than 80,000, almost 88,000 or approximately 90,000 evacuated | Homes, buildings, sub-fire perimeters and event totals require separate definitions | Different official reviews answer different questions |
| Cause status | Agency label reported; supporting causal reconstruction remains limited | Official accidental-cause determination with explicit legal boundary | Unknown until a retrievable final official determination is processed | No single event-wide cause; cause evidence remains attached to each sub-incident | Electrical in the current CAL FIRE record; earlier states remain preserved |
| Area rule | Final agency area retained with jurisdiction note | No context-free area comparison used | Dated snapshots retained; 5,229 km² is not relabelled as a final perimeter | More than 450,000 ha retained as an event-wide lower bound, not one perimeter | 14,021 final acres controls; 14,117 remains a dated operational snapshot |
| Systems question | How evolving agency records change the public reconstruction | How preparedness, infrastructure and response factors shaped consequences | How situational awareness, routes, weather and smoke information supported mass evacuation | How weather, warnings, shelter decisions, utilities and the built environment interacted | How surveillance loss, alerting, receipt, evacuation and fire arrival operated on different clocks |
| Key unresolved module | Detailed spread, evacuation and causal reconstruction | Definition-controlled building totals and full geospatial progression | Final cause, final perimeter and definition-controlled human-impact module | Fire-specific GIS progression, warning receipt and complete sub-incident registry | Device-specific cause evidence, recipient-level delivery data and reproducible spatial progression |
Where is the evidence strongest?
These profiles count the material claims currently scored in each public record. They reveal where evidence is strong or weak; they do not rank the incidents.
Palisades
Lahaina
Fort McMurray
Black Saturday
Eaton
Recurring evidence patterns
The real value of cross-incident intelligence begins when records built with the same architecture expose recurring problems in wildfire documentation.
Precision has a cost
The most precise-looking figure is not always the most defensible one. Timestamps, structure totals and evacuation counts can remain disputed.
Observation is not explanation
Satellite imagery can document location, smoke, active-fire signatures and dated footprint without proving ignition or responsibility.
Cause is not the whole system
Ignition evidence is different from understanding infrastructure, preparedness, communications, evacuation and exposure.
Event scope is evidence
A disaster name can contain many fires. Event, sub-incident and observation layers must remain separate.
Review scope changes meaning
Two investigations can reach compatible findings because they examine different places, periods and questions.
One canonical system. Two controlled projections.
The same record IDs, claim IDs, source IDs, evidence states and revision classes feed both the public Wildfire Black Box and the private Tecnobosque Core.
Wildfire Black Box
Permanent incident pages, readable material claims, public sources, evidence states, final TB-ECI results, limitations, comparison matrices and visible revisions.
Tecnobosque Core
Canonical structured records, deeper provenance, internal processing, derived variables, dependencies, cross-incident relationships, review intelligence and future licensed products.
From public record to structured product
The public layer is a controlled projection of the same underlying canonical record used by Tecnobosque Core.
| LAYER | PUBLIC VALUE | STRUCTURED / CORE VALUE |
|---|---|---|
| Record | Permanent incident page | Canonical record, version history and future licensed endpoint |
| Claim | Readable statement, TB-ECI result and limitations | Exact scoring components, dependencies, dates and processing notes |
| Source | Retrievable citation and public limitation | Provenance graph, independence groups and archival controls |
| Comparison | Curated cross-incident matrix | Derived variables, controlled queries, benchmarks and licensed datasets |
| Revision | Visible material change log | Machine-readable event history, review queue and compatibility notices |
Comparison without false equivalence
Cross-incident comparison does not turn correlation into causation, TB-ECI into a probability model or technical documentation into a legal finding.
- The five pilot registers contain different numbers and types of material claims.
- Claim-confidence profiles will change as evidence is added, corrected or superseded.
- The current pilot set is not a statistically representative sample of global wildfire.
- Currency, area, evacuation and structure values are not normalized unless the transformation is explicitly documented.
- TB-ECI remains a developing Tecnobosque methodology and has not been externally validated as an official standard.
Open the evidence behind the matrix
Every comparison remains connected to an incident record, material claim, evidence state, source and revision history.