Territories · TerraScan

Territorial risk, read from satellite data.

TerraScan turns open European earth observation into a risk score for every administrative unit, with reports and alerts written for the people who have to act on them.

Start an evaluation What it does

Public demonstration · Romania, county by county · fire on real data

Who it is for

For those who answer for a territory.

Local and county administrations, emergency services and forest managers. European satellite data is open and continuous, but it arrives as scenes and spectral bands. Almost nobody turns it into something an administration can act on. TerraScan does that work once, for every unit, on the same data.

What it does

Four risk modules on one pipeline.

  • 01

    Fire

    Fire weather, vegetation dryness, drought, active hotspots and what is exposed around them.

    Implemented
  • 02

    Air

    Air quality and the exposure of the people inside the affected area.

    Implemented
  • 03

    Vegetation

    Vegetation health, water stress and drought, followed as a series.

    Implemented
  • 04

    Water

    Water-related risk for each territory, scored on the same scale as the others.

    Implemented
  • 05

    Composite score

    One figure from 0 to 100 per territory, in classes, that opens into the indicators behind it and the share of the largest ones.

    Implemented
  • 06

    Reports and alerts

    Executive reports in administrative language, downloadable, and rules that raise an alert when a threshold is crossed.

    Implemented

Implemented describes the function. The data behind the public demonstration differs by module: fire runs on real data, while air, vegetation and water run on indicative data. What is real on this page

The product, in its public demonstration

Why this county scores 46.

A manager asks why one territory ranks above its neighbour. The report opens the score into the indicators that produced it, and shows the share of the five largest.

TerraScan demonstration report for a county: a wildfire risk score of 46 out of 100, with bars showing the share of the five largest of its ten indicators, 28, 28, 27, 6 and 5 per cent, and a 30 day trend.
TerraScan public demonstration: a map of Romania with every county scored for wildfire risk, a list of counties by score, the sources and the time of the last update.
Public demonstration · fire · real dataThe fire report for one county, on real data: the score, and what raises it. The bars are the five largest of ten indicators and add up to 94 per cent. The other five indicators carry the rest.Open the demonstration (opens in a new tab)View full-size

How to read the percentages

What a percentage isA relative contribution. The fire score is built from ten indicators. Each is normalised to a value between 0 and 1 and multiplied by its weight, which gives its points. The percentage is that indicator's points divided by the points of all ten, rounded to a whole number.
Why they add up to 94The report lists the five largest contributions: 28, 28, 27, 6 and 5 per cent. The remaining 6 per cent, within rounding, belongs to the five indicators it does not list: surface temperature, forest cover, active fires, critical infrastructure nearby and the season.
What they are notThey are not parts of the 46 points. The score also carries an adjustment for the trend of the last seven days, of at most 8 points up or down, which is not an indicator and has no bar.
The data behind this reportThis is the fire report, the one module the demonstration runs on real data. The report is labelled as generated on real data and gave a confidence level of 90 per cent when it was captured, on 7 October 2026. Its inputs are live, so the score and the shares on the demonstration today can differ from the capture.

What the user does

Switches between the four modules, opens a county, reads which indicators carry the score and compares it with another county. The report can be downloaded as a document from the demonstration report (opens in a new tab).

The result, and its limits

A ranking of where to look first, with the reason attached. The score orders territories for attention. It does not predict a specific event, and detail below county level is shown for a limited set of counties.

In daily work

From a satellite pass to a line in a report.

  1. 01

    Collect

    Optical, radar, thermal and atmospheric products, fire and emergency services, weather feeds.

  2. 02

    Process

    Indices and exposure are computed per administrative unit, not per satellite scene.

  3. 03

    Score

    Each unit gets its figure and its class, with the components visible.

  4. 04

    Report and alert

    Staff read a summary, compare units and receive an alert when a rule fires.

Data and integrations

What it needs in order to work.

Satellite sourcesCopernicus services and the Sentinel-1, 2 and 3 missions. Coverage of the land monitoring layers is European, so the product is currently European in scope.
TerritoryThe list of administrative units to score, with their boundaries.
AI modelReports can be generated with an external language model or with the built-in deterministic generator. The choice is made per deployment.
AccountsRoles for administrators, analysts and viewers, per organisation.
Sentinel-1, 2 and 3Copernicus land and atmosphere servicesEuropean fire informationEmergency management serviceWeather feeds

Results and limits

What you get, and where it stops.

What you get

  • Units compared on one screen, on the same scale
  • A score that traces back to its indicators
  • Reports and funding justifications ready to edit
  • Alerts on thresholds you set

Limits

  • A risk score orders territories for attention. It does not predict a specific event
  • Resolution and refresh depend on the satellite revisit and on the source
  • Outside Europe the land monitoring layers it relies on are not available
  • It does not replace national warning services

What is real on this page

How to read the labels

Implemented means the function exists and runs in the product. It does not mean it has been validated in a customer's operation.

Functions

The four modules, the composite score, the reports and the alerts exist and run in the product. That says nothing about which data a given demonstration report runs on. The data is stated separately, module by module, below.

Fire
Real data

In the public demonstration the fire report is labelled as generated on real data, with a confidence level of 90 to 95 per cent, the level the product gives to live inputs. The two screens on this page are the fire report and the fire view of the map.

Air, vegetation, water
Indicative data

In the public demonstration these three reports are labelled as generated on indicative data, with a confidence level of 55 per cent, the level the product gives to demonstration inputs. Their figures show how the report works. They do not describe the county.

Checked

The labels were read on the demonstration report on 7 October 2026. The demonstration can change, and each report states its own data and confidence level.

Integration at the customer

Detail per locality, accounts, alert rules and report recipients are configured for the territory of each deployment.

Start an evaluation

Name a territory and the risk you answer for.

We walk you through the public demonstration, then tell you what the pipeline would score for your territory, on which sources, and where coverage runs out.

[email protected] All products Deployment and data

The evaluation uses the public demonstration. Your own data is only used once a deployment is agreed.