Glyoxal column mapping for biomass-burning secondary aerosol exposure
TROPOMI's 3.5 km glyoxal retrievals let analysts trace secondary organic aerosol formation back to specific fire events and vegetation types, though bright surfaces and mixed land cover complicate source attribution.
Sensors
- Sentinel-5P TROPOMI: Primary glyoxal (CHOCHO) sensor. Measures the 420–460 nm absorption band at 3.5 × 5.5 km ground pixel (upgraded from 7 km in August 2019), global daily coverage, ~15:00 local overpass. Retrieval uncertainty per pixel is typically 1–2 × 10⁻¹⁵ mol cm⁻², limiting detection to moderately elevated columns; single-pixel fire signals in sparse vegetation can fall below the noise floor.
- GOME-2 (MetOp-A/B/C): Predecessor UV–visible spectrometer that established the multi-year glyoxal climatology used to validate TROPOMI. Ground pixel 40 × 80 km, so unsuitable for individual fire attribution but valuable for long-term trend analysis and inter-mission calibration. MetOp-B continues operational service.
- MODIS Burned Area Product (MCD64A1): 500 m monthly burned-area maps derived from Terra and Aqua, used to co-register glyoxal column anomalies with confirmed fire scars. Does not measure glyoxal directly but provides the spatial anchor for pyrogenic versus biogenic source separation.
- MODIS MAIAC AOD: 1 km aerosol optical depth retrievals used alongside glyoxal columns to estimate how much of the SOA-forming potential has already converted to particulate mass downwind of the fire. Heavily affected by cloud and optically thick smoke, which also degrades the TROPOMI glyoxal retrieval in the same scenes.
Why glyoxal, not just smoke
Smoke plumes are visible. The health problem is what forms inside them after the fact. Glyoxal (CHOCHO) is a dicarbonyl produced when isoprene and aromatic compounds oxidise in the atmosphere. It partitions readily into aerosol droplets and condenses onto existing particles, contributing to secondary organic aerosol, the fraction of PM2.5 that epidemiological studies associate most consistently with cardiovascular and pulmonary outcomes.
Measuring glyoxal columns from orbit therefore gives a forward-looking signal: where SOA is being manufactured, not merely where smoke is already dense. A fire burning through isoprene-rich tropical forest produces a different glyoxal signature than one moving through dry grass, and that difference matters for estimating downwind PM2.5 exposure hours or days later.
What the TROPOMI retrieval actually sees, and what it misses
TROPOMI retrieves glyoxal via differential optical absorption spectroscopy (DOAS) in the 420–460 nm window. At 3.5 × 5.5 km pixel size, it can resolve individual large fire clusters, which is a genuine advance over the 40 × 80 km pixels of GOME-2. The ESA Level-2 product includes a per-pixel quality flag and a vertical column density with associated uncertainty, typically 1–2 × 10⁻¹⁵ mol cm⁻² under clean-sky conditions.
Three limits matter in practice. First, bright surfaces, particularly arid soils and desert sand, elevate the apparent glyoxal signal through a surface-albedo artefact that is difficult to remove fully with current reference sector corrections. Second, optically thick smoke itself attenuates the upwelling radiance, which means the densest part of a plume, where glyoxal concentrations are highest, is also where the retrieval is least reliable. Third, cloud fractions above roughly 0.3 cause the pixel to be flagged and excluded, and biomass-burning regions in the tropics are frequently partly cloudy. Buyers should expect data gaps of 30–60% on any given day over humid tropical fire zones.
Separating pyrogenic from biogenic glyoxal
Isoprene emitted by living vegetation is the dominant global source of glyoxal under non-fire conditions. In a mixed landscape where fires burn alongside intact forest, the satellite column reflects both sources simultaneously. Attribution requires ancillary data: MODIS burned-area maps to identify active fire pixels, land-cover classifications to estimate isoprene emission potential from the unburned fraction, and ideally wind-trajectory analysis to track air-mass history.
Published approaches use the glyoxal-to-formaldehyde (CHOCHO/HCHO) column ratio as a discriminant. Pyrogenic sources tend to produce higher ratios than pure biogenic oxidation because biomass burning emits aromatic precursors that yield glyoxal more efficiently than isoprene does. The method is not clean: the ratio shifts with fire age, fuel moisture and atmospheric oxidant levels, and HCHO retrievals carry their own uncertainties. It is a useful constraint, not a definitive fingerprint.
Translating column density to population exposure
A tropospheric column in mol cm⁻² is not a surface concentration. Converting it to something a public-health analyst can use requires assumptions about the vertical profile of glyoxal, which peaks near the fire and dilutes rapidly with altitude and distance. Chemical transport models such as GEOS-Chem or CMAQ can ingest the satellite columns as constraints and produce surface-level SOA estimates, but the uncertainty compounds at each step.
A simpler operational approach treats elevated glyoxal columns as a qualitative indicator of elevated SOA-forming potential and overlays them with gridded population data to rank affected communities by exposure burden. This does not produce a PM2.5 number, but it does produce a defensible priority list for ground-monitoring deployment or public-health alert issuance. Satellize runs this kind of column-to-population overlay as part of its air-quality analytics stack, drawing on the same open TROPOMI data that underpins its Tonga crop-estimation work and other open-constellation programmes.
Archive depth and what it enables
TROPOMI has been operational since October 2017, giving roughly seven years of daily glyoxal retrievals at the current resolution. GOME-2 extends the record back to 2007. Together they span several major fire years, including the 2019–2020 Australian fire season and the 2015 Indonesian peat fires, allowing retrospective exposure studies that ground-based networks could not have captured at comparable spatial coverage.
The archive also supports seasonal baseline construction. Knowing the typical glyoxal column for a given region and month allows anomaly detection: a column three standard deviations above the seasonal mean over a populated valley is a signal worth acting on, even if the absolute value remains uncertain. This is the kind of contextualised alerting that turns a research-grade retrieval into an operational public-health tool.
Honest limits for procurement decisions
No satellite glyoxal product currently meets regulatory air-quality standards as a standalone measurement. The retrieval uncertainty is too large, the vertical sensitivity too diffuse, and the cloud gaps too frequent for that. What the data does well is spatial triage at continental scale, identifying which fire events in which vegetation types are generating the most SOA-forming potential, and directing limited ground-monitoring resources accordingly.
Buyers expecting sub-kilometre resolution or reliable retrievals over bright arid surfaces will be disappointed. Those expecting a daily continental-scale signal that is meaningfully better than no signal at all, and that improves with multi-day compositing, will find the product genuinely useful. The gap between those two expectations is where most procurement mismatches occur.
Typical figures
| Primary sensor | Sentinel-5P TROPOMI |
| Glyoxal pixel size | 3.5 × 5.5 km (post-August 2019 upgrade) |
| Revisit / coverage | Global daily, single overpass ~13:30 local solar time |
| Spectral band (DOAS window) | 420–460 nm visible |
| Typical retrieval uncertainty | 1–2 × 10⁻¹⁵ mol cm⁻² per pixel under clear sky |
| Cloud mask threshold | Pixels with cloud radiance fraction > 0.5 excluded; effective data loss 30–60% over humid tropics |
| Burned-area anchor (MODIS MCD64A1) | 500 m spatial resolution, monthly composite |
| Archive depth | TROPOMI from October 2017; GOME-2 from 2007 |
| Latency (ESA L2 product) | Near-real-time stream ~3 hours; offline reprocessed product within days |
| Delivery formats | NetCDF-4 (ESA standard), regridded GeoTIFF or GeoJSON on request |
Analytics Satellize can run
| Daily glyoxal column anomaly map | Deviation from multi-year seasonal baseline derived from TROPOMI archive; z-score per grid cell | GeoTIFF layer with anomaly magnitude and confidence flag, updated daily |
| Pyrogenic versus biogenic attribution index | CHOCHO/HCHO column ratio using co-located TROPOMI formaldehyde product, masked against MODIS burned-area pixels | Per-pixel attribution score GIS layer with uncertainty range |
| Population-weighted SOA exposure ranking | Glyoxal column overlay with gridded population data (e.g. WorldPop); ranked list of affected administrative units | Tabular report and choropleth map, updated at each overpass |
| Fire-event SOA-potential fingerprint | Spatial co-registration of TROPOMI glyoxal plume with MODIS active fire radiative power and land-cover type | Per-fire-event summary report with estimated SOA-forming potential by vegetation class |
| Multi-day composite for cloud-gap filling | 5-day rolling mean of valid TROPOMI pixels; flagged where fewer than 3 valid observations contribute | Weekly composite GeoTIFF with data-density metadata |
| Retrospective seasonal exposure assessment | Full-archive TROPOMI and GOME-2 time series analysis for user-defined region and fire season | PDF report with time-series plots, anomaly years identified, and data-gap summary |
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.