The Problem
Fuel load is invisible
until it isn't🔥
California's highway corridors carry millions of people through some of the state's highest-risk wildfire terrain. The fuel accumulating in those corridors: dry grass, ladder fuels, dense canopy. It is rarely mapped, rarely measured, and rarely acted on before a fire starts.

When a fire ignites adjacent to critical infrastructure, the window to respond is measured in hours. The data to act fast has always existed. It just wasn't assembled.
6 High-risk segments, Hwy 101
67 Moderate-risk segments, Hwy 101
15 mi Corridor assessed, 100m resolution
2 Corridors complete, Hwy 101 + I-280
Our Position
Public data.
Public safety.
No tollbooth.

Every dataset powering this tool is public, free, and maintained by the institutions we collectively fund. No subscriptions. No API keys. No vendor lock-in.

Our job is to use public data well, which will underscore its value and show that it's worth maintaining.

Fire and police organizations have been burned by vendor dependency: pricing changes, deprecated endpoints, procurement cycles that outlast the fire season. That ends here.
The Data Foundation
Five sources.
All public.
All current.
USGS 3DEP LiDAR: terrain, canopy height, ladder fuel structure. QL2 density, 2023 coverage.
3D Elevation Program — U.S. Geological Survey
National airborne LiDAR at QL2 density (≥2 pts/m²). Provides bare-earth DEM, canopy surface model, and intensity returns used to derive canopy height, ladder fuel index, slope, and aspect — four of the seven scoring inputs. 2023 coverage verified across both assessed corridors.
LANDFIRE Fuel models, canopy cover, canopy base height. Annual USGS/USFS release.
Landscape Fire and Resource Management Planning Tools — USGS / USFS
Nationally consistent annual fuel and vegetation database. Provides Scott-Burgan 40 fuel models, canopy base height, and canopy bulk density calibrated to fire behavior standards. The same data used in FARSITE and FlamMap — the operational fire behavior models run by CAL FIRE and the USFS during active incidents.
Sentinel-2 NDVI + NDMI monthly composites, live fuel moisture. ESA / MSPC.
Sentinel-2 Multispectral Instrument — European Space Agency
Twin-satellite constellation, 10-day revisit, 10m resolution VNIR bands. Monthly composites processed via Microsoft Planetary Computer. NDMI (Normalized Difference Moisture Index) departure from seasonal baseline is the most validated remotely-sensed proxy for live fuel moisture content at corridor scale — the leading indicator of fire ignition risk and rate of spread.
CAL FIRE Fire Hazard Severity Zones. State + Local Responsibility Areas.
California Department of Forestry and Fire Protection
Fire Hazard Severity Zones (FHSZ) are state-mandated under AB 38, classifying land by fuel load, slope, fire weather, and ember production potential. Moderate, High, and Very High tiers. The highest-weighted input in our composite score (0.25) — because regulatory classification represents decades of accumulated fire science and enforcement history.
USGS SILVIS Wildland-Urban Interface boundaries. Interface, Intermix, Influence zones.
Spatial Index for Landscape Visualization and Synthesis — Univ. of Wisconsin-Madison / USDA Forest Service
National WUI dataset derived from census housing density and NLCD vegetation cover. Interface, Intermix, and Influence zones define where structural ignition risk and suppression complexity converge. WUI proximity is a primary predictor of civilian casualty risk, suppression cost, and insurance exposure — which is why it is in the room with every agency that has a budget problem after a fire.
The Assessment
Seven inputs.
One score.
100m resolution.
Every 100-meter segment of corridor receives a composite risk score from 0 to 1, calibrated, explainable, and reproducible.
0.25 × FHSZ Classification
0.20 × Ladder Fuel Index   ← LiDAR-derived
0.15 × Canopy Height
0.10 × Slope Factor
0.10 × Aspect Factor   ← SSW exposure peak
0.10 × Fuel Curing   ← Sentinel-2 NDMI departure
0.10 × WUI Proximity
Weighted Linear Combination (WLC) — Multi-Criteria Evaluation
This scoring model is a Weighted Linear Combination, a standard form of Multi-Criteria Decision Analysis (MCDA). Each input is normalized to a 0–1 scale, multiplied by its assigned weight, and summed into a composite index. The same structural approach underlies CVSS in cybersecurity, habitat suitability models in ecology, and watershed risk indices in land management.

The weights are subjective and preliminary. They reflect current engineering judgment and published fire science literature, not empirical calibration against historical fire outcomes. This is a starting point, not a final answer. We look forward to working with fire department operations, GIS teams, and field personnel to challenge, refine, and ultimately replace these weights with values grounded in institutional knowledge and lived experience. The model is designed to be transparent and adjustable precisely because that collaboration is the point.
The Pipeline
Six steps.
Fully automated.
01
⇓
Ingest
5 public sources
LiDAR · FHSZ
LANDFIRE · Sentinel‑2 · WUI
▶
02
✎
Prepare
Clip to corridor
Reproject · Resample
100 m segment grid
▶
03
Σ
Score
7‑input formula
Weighted composite
0 – 1 per segment
▶
04
◆
Classify
Tier thresholds
Low → Moderate
High → Critical
▶
05
✈
Generate
WPML KMZ auto‑built
High + Critical only
Waypoints · Gimbal · Speed
▶
06
◇
Deploy
Field collection
LiDAR gap fill
Score refinement
The Tool
From risk score
to drone mission
in one click.
Every High and Critical segment has a mission-ready flight plan generated automatically: waypoints, altitude, speed, gimbal angle, imaging mode.

Draw a survey area. Set parameters. Download a WPML KMZ ready to fly. No manual mission planning. No transcription errors.

Missions serve a dual purpose: operational assessment of high-risk areas, and LiDAR gap collection, where public coverage is missing, overflights fill the knowledge gap and sharpen the analysis.
6 Mission-ready KMZ files, Hwy 101 High-risk
3-Sensor LiDAR · thermal · visible
Already Live
History knows
where fire starts.
CAL FIRE has mapped every significant fire perimeter in California. 22,000+ records back to 1878. That layer is live in this map, right now.

When we overlaid it on the Hwy 101 corridor today, the High-risk segments our pipeline identified correlated directly with a historic burn area adjacent to the corridor.

We did not tune the model to fit historical fires. We built it from first principles. The fires found the same places we found.
23K+ CAL FIRE perimeters, live in the map
0 New data sources required
The Community
The corridor runs through
someone's neighborhood.
The people most affected by a corridor fire are not in this room. They live adjacent to the risk. They know the terrain. They remember the last fire. And right now, they have no visibility into what we can see.

The intelligence this pipeline produces does not have to stop at the agency level. Activated communities are better-prepared communities. Evacuation routing informed by live risk scores. Neighborhood preparedness tied to fuel load maps. Early warning surfaced from the same data that drives the drone missions.

Local knowledge also flows back in. The field observations, the access points, the micro-terrain details that no dataset captures. That is a feedback loop worth building.
Predictive Intelligence
Where is
the next fire?
Combining historic ignition patterns, current fuel moisture, wind direction, temperature, and relative humidity, we can model a probability surface for fire origin across any corridor.

Not a prediction. A probability. And high enough to act on.

Pair that with fire spread modeling and you know not just where it starts — but where it goes, and how fast.
Structure Risk
Some things
cannot be rebuilt.
Power substations. Water treatment facilities. Hospitals. Schools. Historic landmarks. The places that define a community's identity and enable its survival.

When fire starts, these are the coordinates that matter most. We map them, score them by fuel exposure and proximity, and surface them first, so the people making decisions in the first hours know exactly what they are protecting.

Some of what we protect carries a value beyond any dollar figure. The places that hold the memory and pride of a community. Those matter too.
Response Optimization
Fight the fire
that hasn't started
yet.
Pre-positioned resources. The optimal attack direction before the fire makes that choice for you. Task-level assignments that close information gaps in the chaos of the first hours.

The difference between containment and catastrophe is often the quality of the first decision, made before the fire has established momentum, before it has chosen its path, before it has found the infrastructure it will take with it.

We are building the tools to make that first decision the right one.
The Mission
The data exists.
The tools exist.
The time is now.

No subscriptions. No gatekeepers. No waiting for the next budget cycle to access information that has always been public.

Assembled with purpose. In service of the communities that fund it. Built to protect the people, the places, and the things that cannot be replaced.

This is what it looks like when the information gap closes.
Closing the Data Gap
The sensor exists.
The payload doesn't.
SWIR · 900–1700 nm
Live Fuel Moisture
Direct leaf water content, not a chlorophyll proxy. The band fire behavior models actually need. Unavailable on any sub-$50K turnkey payload today.
LiDAR · 360° point cloud
Fuel Structure
Canopy height, ladder fuel density, understory volume. The only sensor that sees below the canopy. Satellites and cameras cannot substitute.
LWIR · 8–14 μm
Thermal Signature
Surface temperature anomalies, pre-ignition stress patterns, structure exposure. Active surveillance between fire events.
Target mass: < 500 g Power: < 10 W Output: GeoTIFF · open pipeline Architecture: field-swappable · dock-compatible