Black carbon deposition risk mapping from Arctic shipping routes

Heavy fuel oil combustion deposits black carbon on Arctic sea ice, cutting albedo and accelerating melt. TROPOMI aerosol data correlated with AIS vessel tracks can map emission density and model downwind deposition risk, though bright ice surfaces make retrieval genuinely hard.

Sensors

Sentinel-5P TROPOMI: Provides daily global NO2 columns at 3.5 × 5.5 km (reprocessed to 3.5 × 3.5 km from August 2019) and the Absorbing Aerosol Index (AAI) at the same resolution. AAI is the primary proxy for UV-absorbing aerosol load, including black carbon, but is not a direct BC mass concentration measurement. Retrieval quality degrades significantly over bright snow and ice surfaces where surface albedo assumptions break down.
MODIS Terra and Aqua: Provides aerosol optical depth (AOD) at 10 km resolution with twice-daily combined revisit. The Dark Target algorithm fails over bright surfaces; the Deep Blue algorithm extends retrieval to some bright land surfaces but is not validated for sea ice. MODIS is most useful here for cloud masking and surface reflectance characterisation rather than direct BC detection.
VIIRS DNB (Suomi-NPP and NOAA-20): Day-Night Band provides 750 m panchromatic imagery useful for ship detection in darkness and for identifying exhaust plume structure in low-light conditions. Not a direct aerosol sensor, but complements AIS by providing independent vessel position evidence in data-sparse polar regions.
AIS vessel tracking: Automatic Identification System messages, collected by terrestrial receivers and increasingly by low-Earth-orbit satellites (commercial providers cover polar gaps), give position, heading, speed and vessel identity at intervals typically of seconds to minutes. Vessel type and flag enable fuel type inference. AIS is the backbone for correlating emission locations with atmospheric observations.

Why black carbon in the Arctic is a different problem from elsewhere

Black carbon is not a greenhouse gas in the conventional sense. It is a solid aerosol that absorbs solar radiation directly in the atmosphere and, critically, when it settles on snow or ice it reduces surface albedo from roughly 0.8 to values that can approach 0.4 depending on deposition density. That albedo reduction drives additional melt independent of background warming. The Arctic amplification effect means the region is warming two to four times faster than the global average, so incremental forcing from deposited BC is not a marginal concern.

Shipping is a growing source. The Northern Sea Route and Northwest Passage are navigable for longer seasons than a decade ago, and transit traffic has increased accordingly. Ships burning heavy fuel oil (HFO) emit black carbon at rates that vary with engine load and fuel sulphur content. The International Maritime Organization has discussed HFO restrictions in Arctic waters, but enforcement requires knowing where ships actually are and what their exhaust is doing after it leaves the stack. That is precisely what this use case addresses.

What TROPOMI actually measures, and where it struggles

TROPOMI's Absorbing Aerosol Index is derived from the ratio of UV radiances at 340 and 380 nm. A positive AAI indicates the presence of UV-absorbing aerosols, which includes black carbon, mineral dust and volcanic ash. It is a qualitative to semi-quantitative indicator, not a mass concentration. Converting AAI to BC column density requires assumptions about aerosol layer height, particle size distribution and mixing state, none of which are directly observed by TROPOMI over the open ocean.

Over sea ice the problem compounds. The retrieval algorithm assumes a surface reflectance model. Bright, specular sea ice violates that assumption, introducing systematic errors in both AAI and NO2 columns. ESA and KNMI documentation for the TROPOMI Level 2 products explicitly flags retrievals over snow and ice as having elevated uncertainty. In practice, the AAI signal from a single ship transit is likely below the noise floor of a single overpass. The method becomes useful when aggregating many transits over weeks to months along defined route corridors, where the cumulative signal rises above background variability.

Correlating AIS tracks with atmospheric columns

The analytical workflow begins with AIS. Each vessel position is timestamped and carries a Maritime Mobile Service Identity number that links to a vessel registry. From registry data, gross tonnage and vessel type allow estimation of engine power class. Published emission factors for HFO combustion, such as those in the IMO's Fourth Greenhouse Gas Study, give a plausible BC emission rate per unit fuel consumed, which itself scales with speed and load. The result is a spatially explicit emission inventory along each route segment.

That inventory is then compared against TROPOMI NO2 and AAI fields for the same day and adjacent days, accounting for TROPOMI's overpass time relative to vessel position. NO2 is a co-emitted tracer with a short atmospheric lifetime of hours to a day in the lower troposphere, making it a useful near-field indicator of combustion even when BC itself is not directly retrievable. Where NO2 enhancements align spatially and temporally with AIS-derived ship positions, the correlation strengthens the case that the atmospheric signal originates from the vessel.

This approach has published precedent. Studies using TROPOMI NO2 to identify shipping emission corridors in the North Sea and Baltic have demonstrated detectable signals from individual large vessels under favourable conditions. The Arctic is harder: lower sun angles reduce UV retrieval quality in winter, cloud cover is frequent, and the bright surface adds noise. Honest expectation-setting matters here. The method works best on busy corridors with repeated transits, not on isolated vessels.

Trajectory modelling: from stack to ice surface

Even a confirmed emission location does not tell you where the BC lands. Deposition depends on atmospheric transport, wet and dry removal rates, and the distance to the nearest ice surface. Lagrangian trajectory models, most commonly HYSPLIT from NOAA, use reanalysis wind fields to forward-project air parcels from a ship's position at emission time. The output is a probability distribution of where the plume travels over the following 24 to 72 hours.

Combining trajectory ensembles with a simplified dry deposition velocity for BC particles (typically in the range of 0.1 to 0.3 cm per second for accumulation-mode particles) produces a deposition footprint: a spatial layer showing where BC from a given route segment is most likely to settle. Overlaying that footprint on sea-ice extent data from NSIDC passive microwave products identifies which ice areas are at elevated deposition risk. The uncertainty in this chain is substantial. Trajectory errors grow with time, deposition velocities vary with particle size and surface roughness, and wet removal during precipitation events can dominate over dry deposition. The output is a risk map, not a measured deposition field.

What the data can and cannot support

This method supports route-level emission density ranking, not ship-by-ship enforcement. If a regulator wants to know which of two proposed Arctic corridors poses higher cumulative BC deposition risk to a given ice shelf, the AIS-plus-TROPOMI approach can provide a defensible, quantified answer with appropriate uncertainty bounds. If the question is whether a specific vessel exceeded a BC emission standard on a specific day, the answer from orbit is: not yet, not with current sensors.

Direct BC mass measurement from orbit does not currently exist at useful resolution. TROPOMI AAI is the closest available proxy. A dedicated UV-Vis absorbing aerosol mission with improved spatial resolution and better surface albedo handling would change this picture, but no such mission is operational as of mid-2025. The honest value proposition is probabilistic risk mapping for policy and route planning, not compliance adjudication. Satellize applies this workflow, using open Sentinel and VIIRS data combined with commercial AIS feeds, for clients who need to understand Arctic emission exposure before committing to route infrastructure or insurance underwriting.

The Copernicus programme archives TROPOMI data from April 2018 onward, giving roughly seven years of historical record. That is enough to characterise seasonal route patterns and identify corridors where cumulative NO2 and AAI anomalies are statistically significant above background, which is the foundation any serious deposition risk assessment must stand on.

Typical figures

TROPOMI spatial resolution (NO2 and AAI)	3.5 × 3.5 km (from August 2019); 3.5 × 5.5 km prior
TROPOMI revisit	Daily global coverage; single overpass per day at a given location
MODIS AOD resolution	10 km; twice-daily combined Terra and Aqua revisit
VIIRS DNB resolution	750 m at nadir; daily polar revisit from Suomi-NPP and NOAA-20
AIS positional latency (satellite AIS)	Typically 10 to 90 minutes depending on provider and polar coverage arc
TROPOMI AAI retrieval uncertainty over sea ice	Elevated; flagged in ESA Level 2 quality field. Retrievals over snow and ice require independent validation before quantitative use
TROPOMI archive depth	April 2018 to present via Copernicus Data Space Ecosystem
Trajectory model temporal range	HYSPLIT forward trajectories reliable to approximately 48 to 72 hours; uncertainty grows nonlinearly with time
Minimum detectable NO2 enhancement (TROPOMI)	Approximately 0.5 to 1 × 10¹⁵ molecules per cm² above background under clear-sky conditions; higher detection threshold over bright surfaces
Delivery formats	GeoTIFF emission density grids, GeoJSON deposition footprint polygons, CSV route-segment emission inventories, PDF risk assessment report

Analytics Satellize can run

Route-corridor emission density map	AIS-derived vessel inventory correlated with TROPOMI NO2 column anomalies; multi-week temporal aggregation to lift signal above single-overpass noise	Monthly GeoTIFF grid of normalised NO2 anomaly per route segment, with vessel-count weighting
Absorbing aerosol index time series by corridor	TROPOMI AAI extraction along AIS-defined route buffers; quality-filtered to remove cloud-contaminated and low-sun-angle retrievals	CSV time series with quality flags and seasonal breakdown; suitable for trend analysis
BC deposition risk footprint	HYSPLIT ensemble forward trajectories from AIS emission points; dry deposition velocity parameterisation from published literature for accumulation-mode BC	GeoJSON polygon layer of 50th and 90th percentile deposition probability zones, overlaid on NSIDC sea-ice extent
High-traffic vessel emission inventory	IMO Fourth GHG Study emission factors applied to AIS speed and vessel-type data; HFO consumption estimated from engine power regression	Per-vessel and per-route-segment BC and NO2 emission estimates in CSV, with uncertainty range
Ice-surface albedo change risk index	Deposition footprint intersected with MODIS surface reflectance time series to identify ice areas with prior albedo reduction signals	Ranked risk index table by ice grid cell, suitable for insurance underwriting or environmental impact assessment input
Seasonal route comparison report	Multi-year TROPOMI archive aggregation across defined Northern Sea Route and Northwest Passage corridors; statistical anomaly detection against pre-shipping-season baseline	Annual PDF report comparing emission density and deposition risk across candidate routes, with year-on-year trend

Sources

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.