Sensors
- Meteosat SEVIRI: Spinning Enhanced Visible and InfraRed Imager on Meteosat Second Generation. Full-disc scan every 15 minutes (rapid-scan sector every 5 minutes). Spatial sampling 3 km at nadir in thermal IR channels, degrading toward limb. Channels at 10.8 µm and 12.0 µm are the operational split-window pair for ash BTD retrieval. Covers Europe, Africa, the Atlantic and Indian Ocean VAAC regions.
- GOES-16/18 ABI: Advanced Baseline Imager on NOAA's geostationary pair. Full-disc imagery every 10 minutes, CONUS sector every 5 minutes, mesoscale sectors every 30 to 60 seconds during events. 2 km resolution in thermal IR at nadir. Bands 11 (8.4 µm), 14 (11.2 µm) and 15 (12.3 µm) support both split-window ash detection and SO2 proxy retrievals. Covers the Washington and Anchorage VAAC regions.
- MetOp IASI: Infrared Atmospheric Sounding Interferometer: a hyperspectral sounder with 8,461 channels across 3.6–15.5 µm, 12 km circular footprint at nadir. Polar orbit gives two overpasses per day per location, so it cannot match geostationary revisit for real-time tracking. Its strength is quantitative retrieval of ash column mass loading and SO2 column amounts, which constrain dispersion model inputs. Published detection limits for SO2 are on the order of a few Dobson Units.
- Sentinel-5P TROPOMI: TROPOspheric Monitoring Instrument on Sentinel-5 Precursor. 3.5 km × 5.5 km pixel (upgraded to 3.5 × 3.5 km from August 2019), daily global coverage. UV-visible SO2 column retrieval is operationally validated; the instrument can detect SO2 plumes from moderate eruptions and is used as a corroborating chemical signal when the BTD ash signature is ambiguous. Near-real-time products available within 3 hours of overpass.
What a split-window brightness temperature difference actually measures
The detection method rests on a physical quirk of silicate ash particles. In clear sky, the 10–11 µm channel records a higher brightness temperature than the 12 µm channel because water vapour absorbs more strongly at the longer wavelength. Ash particles invert this relationship: their refractive index causes stronger absorption at 11 µm than at 12 µm, producing a negative brightness temperature difference (BTD). A value below roughly minus one kelvin, sustained across multiple pixels, is the operational flag for ash presence.
The magnitude of the BTD signal depends on ash particle size, concentration and the temperature contrast between the plume and the surface below. Thin, high-altitude plumes over cold ocean backgrounds produce weaker signals than dense, low-altitude plumes over warm land. Operational thresholds used by VAACs are not universal; they are tuned by region and season. This is not a weakness of the physics, it is an honest constraint that practitioners work around with local climatology.
The ice-cloud problem: when the atmosphere lies
High-altitude ice cloud produces a BTD signal that overlaps substantially with volcanic ash. Both can drive the split-window difference negative. Distinguishing them matters enormously: a false positive closes airspace unnecessarily, a false negative puts aircraft in danger.
Several disambiguation strategies are used in combination. Temporal coherence helps: ash plumes originate from a point source and evolve in a pattern consistent with known eruption timing, whereas meteorological cloud fields move differently and lack a volcanic source. Colour ratio methods using visible and near-infrared channels add texture information. Most decisively, SO2 detection breaks the tie. Ice cloud contains no SO2; a co-located SO2 enhancement from TROPOMI or IASI alongside a negative BTD is strong evidence for a genuine ash event. The ambiguity is not fully resolved by any single channel or instrument, which is why VAAC forecasters treat multi-sensor corroboration as mandatory, not optional.
How VAACs turn pixels into airspace decisions
Nine Volcanic Ash Advisory Centres, designated by ICAO, share global responsibility for issuing Volcanic Ash Advisories (VAAs) and the associated Volcanic Ash Graphic (VAG). The operational chain moves fast. A VAAC receives satellite imagery, pilot reports and ground observatory alerts, runs a trajectory dispersion model (the London VAAC uses NAME, the Washington VAAC uses HYSPLIT), and issues advisories at six-hourly intervals or more frequently during active events.
The 2010 Eyjafjallajökull eruption exposed how much economic damage a poorly constrained ash cloud can cause: European airspace closures grounded roughly 100,000 flights over six days. The response was a systematic upgrade of satellite data assimilation into dispersion models and the introduction of quantitative ash concentration thresholds (low: below 2 mg/m³; medium: 2–4 mg/m³; high: above 4 mg/m³) that replaced the earlier binary close-or-open approach. Satellite retrievals from IASI and SEVIRI directly feed these concentration estimates.
Latency is the operational pressure point. A geostationary sensor delivers a new image every 5 to 15 minutes. Processing and advisory issuance add further minutes. An eruption column can reach cruising altitude in under 30 minutes. The system works, but it has no slack.
SO2 as a leading indicator, not just a confirmation
Sulphur dioxide is co-emitted with ash during explosive eruptions, but the two do not always travel together. Larger ash particles settle gravitationally within hours to days; SO2 remains in the atmosphere for days to weeks and disperses over continental scales. This means SO2 can serve as a leading indicator of eruption intensity before the ash plume is fully resolved, and as a long-range tracer after the dense ash cloud has thinned below BTD detection thresholds.
TROPOMI's daily global coverage at sub-5 km resolution makes it the current benchmark for SO2 column mapping at the synoptic scale. IASI provides independent quantitative mass loading estimates twice daily. Neither replaces geostationary monitoring for real-time tracking, but both substantially reduce the uncertainty in dispersion model initialisation. For low-level, SO2-poor eruptions (some basaltic Hawaiian-style events), the chemical signal is weak and the BTD method carries more weight.
Honest limits: what the sensors cannot do
Cloud cover is not the primary enemy here, unlike optical use cases, because thermal infrared penetrates meteorological cloud to some degree. The real limits are different. Thin ash above thick meteorological cloud can be masked entirely. Very fresh, coarse-grained ash (particle diameter above roughly 30 µm) produces a weaker BTD signal and may be missed in the first minutes after an eruption before the plume rises and spreads. Geostationary sensors lose spatial resolution toward the limb: SEVIRI's 3 km nadir sampling degrades to roughly 8 km at 60° latitude, which matters for Icelandic or Aleutian events.
Quantitative ash concentration retrievals carry substantial uncertainty. Published studies report retrieval errors of a factor of two or more in mass loading, depending on assumed particle size distribution. VAACs are aware of this and treat satellite-derived concentrations as order-of-magnitude guidance rather than precise measurements. Pilots and dispatchers should understand that the advisory polygons represent the best available estimate under time pressure, not a certified boundary.
Satellize ingests TROPOMI SO2 products and SEVIRI/ABI imagery operationally and can deliver co-registered, multi-sensor ash and SO2 overlays for government aviation authorities or airport operators who need a single integrated feed rather than raw agency products. The analytic workflow draws on the same open-data pipelines used for our Tonga crop programme, adapted for the very different time pressures of aviation hazard.
Typical figures
| Spatial resolution (geostationary IR) | 3 km at nadir (SEVIRI); 2 km at nadir (ABI). Degrades to 6–8 km at high latitudes. |
| Revisit (geostationary) | 5–15 minutes full disc; 30–60 seconds for ABI mesoscale sectors during events |
| Spatial resolution (polar sounder) | 12 km circular footprint (IASI); 3.5 × 5.5 km (TROPOMI, upgraded to 3.5 × 3.5 km) |
| Revisit (polar) | Twice daily per location (IASI); daily global coverage (TROPOMI) |
| Key spectral channels | 10.8 µm and 12.0 µm split-window pair (BTD ash); UV-visible SO2 band ~310–340 nm (TROPOMI); full thermal IR spectrum 3.6–15.5 µm (IASI) |
| SO2 detection limit | Low single-digit Dobson Units for IASI and TROPOMI under favourable conditions |
| Operational latency (VAAC advisory) | Typically 30–90 minutes from eruption onset to first advisory; updated every 6 hours or more frequently during active events |
| TROPOMI NRT product latency | Within 3 hours of overpass |
| Archive depth | SEVIRI: Meteosat-8 onwards (~2004); GOES-16 ABI: 2017 onwards; TROPOMI: late 2017 onwards; IASI: 2007 onwards |
| Ash concentration threshold classes (ICAO) | Low <2 mg/m³; Medium 2–4 mg/m³; High >4 mg/m³ |
Analytics Satellize can run
| Real-time ash plume polygon | Split-window BTD thresholding on SEVIRI or ABI thermal IR imagery, with temporal coherence filtering to reduce ice-cloud false positives | GeoJSON polygon updated every 15 minutes, compatible with aviation GIS and flight-planning systems |
| SO2 column overlay | TROPOMI Level-2 SO2 total column product ingestion and reprojection, co-registered with BTD ash layer | Gridded GeoTIFF or WMS tile, issued within 4 hours of TROPOMI overpass |
| Multi-sensor ash confidence score | Bayesian combination of BTD signal, SO2 presence/absence and plume trajectory consistency with known eruption source | Per-pixel confidence flag (high/medium/low/ice-cloud ambiguous) appended to ash polygon layer |
| Dispersion model initialisation file | IASI-derived ash mass loading and SO2 column used to set source term parameters for NAME or HYSPLIT trajectory runs | Structured input file in format accepted by VAAC dispersion models, with uncertainty bounds |
| Event timeline and eruption chronology report | Retrospective analysis of geostationary archive imagery to reconstruct plume evolution, column height proxy and areal extent over the full event duration | PDF report with time-stamped maps and quantitative plume area and estimated SO2 burden, suitable for post-event safety investigation |
| Automated alert on BTD threshold breach | Continuous monitoring of SEVIRI or ABI near-real-time feeds against user-defined geographic watch zones and BTD thresholds | Push alert via API or email within minutes of threshold exceedance, with thumbnail image and coordinates |
Who does the work
We can get this done for you. Satellize runs its own analyst desk and a strong science team. You do not buy a data feed and work out what it means; our people source the imagery, run the analysis described on this page, and hand you the answer with its confidence limits stated. Discuss this requirement.