Explore property records
and wildfire research.
Try sample inspection records, 3D geometry, and experimental exposure scenarios. These demonstrations do not provide a validated property risk score, safety determination, or certification.
Capture roof geometry, edges, and adjoining vegetation. Attach close-up photographs and note details that need field verification.
Document the exterior.
Connect the observations.
The proposed workflow combines suitable aerial imagery, a field review, and 3D documentation. This sample shows how observations could be recorded; it does not issue a certification or establish a property’s wildfire safety.
Record visible conditions and inspection limits.
The planned field review covers accessible vents, roof edges, exterior materials, and vegetation. Photographs, visible product markings, and notes would connect each observation to the property record.
What goes into an inspection record?
- Capture the propertyAerial imagery, accessible exterior details and a record of survey limits.
- Build & verify the modelAlign the geometry with site measurements and flag details that still need review.
- Check ember-resistant ventsVerify product labels or listings, installation, condition and gaps against the applicable criteria.
- Document next stepsRecord open questions and proposed improvements. Any independent designation requires a separate program’s eligibility, inspection, and approval.
Explore vegetation patterns
and model assumptions.
Compare historical aerial imagery, mapped features, and experimental spread scenarios. These demonstrations do not measure the risk or treatment needs of a customer property.
Not a wildfire forecast or risk score
Choose the assumptions, then run the model.
Blend the classified grid with its source imagery.
Follow the scenario.
- Current model step
- —
- Cells reached
- —
- Spotting arrivals
- —
Model details & data sources
Landscape composition
Inspect the active landscape or controlled test plot.
Reach over the run
Cumulative reached cells on the model timeline.
Imagery: USDA NAIP aerial photography captured July 9, 2022, at 60 cm native resolution. This historical study area is not a drone inspection of a customer property. The original project has separate Sentinel-2 layers; they are not inputs to this run.
Synthetic landscape: Near-infrared/red image values identify vegetation candidates. The precomputed pipeline aggregates them to a 243 × 268 grid at approximately 5 m and overlays OpenStreetMap building and estimated road footprints. Candidate vegetation is not verified fuel type. “Other mapped cover” means no spectral vegetation trigger; it can still contain dry fuels. Imagery and later mapped footprints may not align.
Calculation: The original shortest-travel-time engine runs locally in a background worker. Wind direction is the direction of travel, clockwise from north. Strength and moisture change relative travel rates; they are illustrative coefficients, not weather observations. The default scenario allows travel only through mapped vegetation. The additional-ground-fuel scenario gives other cover and roads 10% and 5% of the vegetation travel rate. Neither option establishes actual fuel conditions. Building ignition is disabled and unknown cells are excluded, not classified as safe.
Controlled test plots: These examples replace the Auburn map with a synthetic 81 × 89 cell fuel bed. The no-fuel example adds a 30 m band around the central ignition patch. Its cells have zero travel rate, so adjacent-cell spread cannot cross or cut their corners. The connected example removes that band. These are demonstrations of fuel connectivity, not a treatment design or a property forecast.
Ember spotting: When enabled, the seeded engine uses three attempts per reached vegetation cell, a 60 m maximum range, a 20% attempt coefficient, and seed 1. “Spotting arrivals” counts reached cells whose selected arrival path came from a spotting jump, not all airborne embers. These settings are experimental and are not probabilities of wildfire loss.
Playback: The timeline displays calculated arrival times in arbitrary model steps. Earlier reach and recent reach describe arrival ordering, not burn severity or active combustion. When no reachable fuel remains, the calculated arrival sequence ends; this engine does not calculate flame extinction or fuel consumption. Recent reach means arrival within the preceding 5% of this run’s total duration. Models with different settings can have different final times; compare the step values, not slider positions. Aerial blending changes only the display.
No validated heat transfer, terrain, canopy fire, suppression, damage, or real elapsed time is calculated. Results do not establish treatment effectiveness, insurance eligibility, or the safety of a property.
Sources: USDA NAIP via Microsoft Planetary Computer; © OpenStreetMap contributors, ODbL. Original imported data and engine are unchanged.
Import record ↗ · Imagery sources ↗Experimental wildfire-exposure
modeling.
Explore a research example of a synthetic building exposed to a prescribed heat source. The platform is in development; this example excludes embers and does not produce a validated property risk score.
Drone imagery, 3D / CAD geometry, building details, and surrounding vegetation.
Explore how fire conditions interact with the building and landscape.
Research comparisons need suitable field data and validation before they can support property decisions.
Explore the research
in three dimensions.
Rotate the house, play the timeline, and explore the smoke and heat-release layers.
Saved research run · loads on requestHow this research fits our platform
Development status: A1 Ember Terrain’s property-documentation and exposure-research platform is in development. This viewer demonstrates how a modeled exposure can be examined in 3D; it does not produce a validated property risk score.
This study: A precomputed Fire Dynamics Simulator (FDS) research run uses a synthetic house and a prescribed heat source derived from a public QUIC-Fire example. FDS and QUIC-Fire are external research tools; this example does not demonstrate an A1 Ember Terrain wildfire prediction model. This is a separate dataset from the Auburn landscape inventory and the paired house photographs.
Scope: This single scenario excludes embers and uses simplified materials and geometry. It plays back a prescribed heat source rather than recalculating fire spread through changing fuel. Smoke and heat-release values are displayed as thresholded, quantized point samples, not photorealistic flames. It does not establish mitigation effectiveness, insurance pricing, or the safety of a real property. Property-level measurement needs field data and model validation.
Read the study results and assumptions