Skip to content

Knowledge

What sits behind the situation picture: which sensors measure, how they measure, what they cannot see and by which methods samoEKO turns raw data into a statement.

Every figure on these pages has a named source: operator, product name and, where one exists, the publication. Where a number is not documented, none is given here.

This section is public and requires no sign-in. It addresses three groups of readers at once: operational and administrative staff who want to know what they are looking at on the map; students working on this data; and funders examining the substance behind the presentation.

Satellites and sensors

Seven data paths, four operators. Each page states what the sensor physically measures, at what resolution and in what cadence, where its limit lies and what samoEKO uses it for.

  • VIIRS: active fire at 375 metres

    375 m · 6 windows/day

    The backbone of the situation picture: three satellites, six observation windows a day, the finest thermal fire detection available.

  • MODIS: the long record

    1 km · archive since 2000

    Coarser than VIIRS and with an end date, but the only source reaching back to the year 2000.

  • Sentinel-2: ten metres

    10 m · every 2-3 days

    Not a fire sensor but the detail image: individual fire edges instead of heat blobs. The only layer whose tiles samoEKO cuts itself.

  • Sentinel-3 SLSTR: the European second opinion

    night pass · no FRP

    Fire detections that FIRMS does not carry: an additional night window from a second operator chain but without an intensity value.

  • Sentinel-5P TROPOMI: trace gases

    3.5 × 5.5 km · daily

    Measures gas columns above a location, not concentrations at ground level. For sulphur dioxide this data path meets a hard limit.

  • Meteosat Third Generation: the ten-minute cadence

    10 min · 500 m

    The only sensor that shows movement: an image every ten minutes, plus fire temperature and lightning at five-minute cadence.

  • CAMS: computation, not measurement

    approx. 40 km · model

    The only layer on the map that shows nothing observed: a model forecast for particulate matter and fire smoke.

Methods

The calculations between raw data and statement: the fire weather index, burnt area mapping including burn severity, emission estimation following the IPCC guidelines, and ground sensing with its corrections.

  1. 1Fire weather index: the Canadian system

    Six chained quantities from weather and fuel moisture. Two operators compute it, with markedly different results.

  2. 2Burnt areas: mapping, minimum size and burn severity

    From point to polygon: how a burnt area is produced, how small it may be, and what can be said about its severity.

  3. 3Emission estimation following the IPCC guidelines

    Four quantities, four disclosed model assumptions and an overall uncertainty of a factor of two, which is why a single number is never given.

  4. 4Ground sensing: particulate matter from low-cost sensors

    What a satellite cannot do: measure continuously where people breathe. The price is corrections that must be documented and versioned.

Reference

Limits of detection
samoEKO is a situation picture with six observation windows per day and explicitly not a real-time fire detection system. The five measurable limits, observation gap, latency, cloud cover, detection threshold and permanent industrial heat sources, are collected on a page of their own. It is part of this knowledge section and is linked from every sensor page.To the limits of detection
Data sources and licences
Who supplies what, under which licence, and which attribution is mandatory. Plus the sources that were examined and rejected on licence grounds.To the data sources
For students
The research API, the rules on reproducibility and the catalogue of thesis topics.To the student section

samoEKO is a brand of samoLabs, Juelich.