Skip to content
← Missions

Black Pluto

Turning open forest data into a 3D map of what a wildfire has available to burn — parcel by parcel, in megajoules of potential fire energy.

Geospatial · environment · dual-useAdvanced Data Engineering2023
Aerial imagery of a cleared field meeting dense Mediterranean forest on the Sainte-Baume massif, Var, with the canopy layer switched off.The same aerial view with the canopy layer on: every tree replaced by an extruded block whose height is its measured LiDAR canopy height, so the forest edge becomes a wall of grey volumes against the bare field.
Drag the handle to compare. One viewpoint, canopy layer toggled. Extruded 3D Tiles in CesiumJS; canopy heights from IGN LiDAR. Basemap imagery via Cesium ion.

What Black Pluto is

Black Pluto turns open geospatial data into a 3D map of the fuel available for a wildfire to burn. It fuses satellite imagery (Sentinel-2), airborne LiDAR (IGN) and French national forest-inventory data into a single terrain model showing — parcel by parcel, at 2-metre resolution — how much biomass is standing, where, and what that means in megajoules of potential fire energy.

The output is a navigable 3D scene that lets a fire-service planner, a forest manager or a terrain analyst answer three questions in seconds:

  • How much fuel is on this hillside, right now?
  • How does it change between visits — after clearing, after growth, after a burn?
  • Where should clearing be prioritised to break the propagation chain?

Why it exists

Existing forest-management tools operate on annual updates and hectare-scale approximations, too slow and too coarse for prevention decisions taken in a season. Black Pluto’s contribution is fusing public datasets at a granularity they are not normally combined at, and rendering the result in 3D at a cost compatible with municipal and regional budgets.

The same modelling carries dual-use value: terrain-planning teams need to know how vegetation cover affects movement, line of sight and concealment. That is the same fuel-load calculation viewed through a different lens.

What we built

  • ETL pipeline — FME orchestration ingests Sentinel-2 tiles, IGN LiDAR point clouds and forest-inventory tabular data; aligns them to a common coordinate reference system; computes NDVI, canopy height and biomass per 2×2 m pixel.
  • Storage — per-parcel time series, so biomass evolution across satellite passes is queryable rather than recomputed.
  • Analytics core — Python modules converting fuel load to megajoules using peer-reviewed biomass-to-energy coefficients per species class.
  • 3D visualisation — the result served in a browser: pan and zoom the terrain, click a parcel, get its fuel-load number.

Where it was proven

Sainte-Baume massif, Plan d’Aups (Var). A 2×2 km parcel processed end to end — raw open-data ingest to 3D output — in ~8 minutes per km². Cross-checked against on-the-ground forest-inventory samples, biomass estimation fell inside the ±10% error band the project targeted.

Current status

TRL 4 — validated on POC data rather than in the field. The work is continuing, and the profile of a first operational partner is clear: a fire service, a regional forest-management authority, or a defence prime working on terrain intelligence.

Why it sits alongside the client work

Black Pluto is not a client engagement, and it is listed as what it is. It earns a place here because it demonstrates the same operating model, with our own money and no client to absorb the risk:

  • Data engineering across incompatible sources: three open datasets, one coherent model, three different coordinate and sampling regimes reconciled.
  • Aerospace-grade discipline applied elsewhere: the same coordinate-reference rigour, configuration management and calibration habits, aimed at forestry.
  • Small team, whole stack — ETL through storage, analytics and 3D visualisation, built by the core team without a consultancy-sized project organisation.