Coastal polynya sea-salt aerosol production and atmospheric loading
Coastal polynyas expose open water in the depths of polar winter, generating frost flowers and sea-salt aerosol that load the troposphere with bromine and drive ozone depletion events. Fusing AMSR2 brightness temperatures, Sentinel-1 SAR backscatter and MERRA-2 wind reanalysis makes the flux quantifiable.
Sensors
- AMSR2 (GCOM-W1): Passive microwave radiometer operating at 6.9 to 89 GHz. Spatial resolution ranges from roughly 3 km at 89 GHz to 35 km at 6.9 GHz. Daily global coverage. Brightness temperature polarisation ratios at 18.7 and 36.5 GHz distinguish open water, nilas, and young ice, and elevated emissivity at high frequencies flags frost-flower-covered surfaces. Revisit is effectively daily at polar latitudes.
- Sentinel-1 SAR (C-band, 5.405 GHz): Synthetic aperture radar providing 5 x 20 m resolution in Interferometric Wide Swath mode (250 km swath). C-band backscatter from newly formed nilas ice is anomalously high relative to older ice because brine-wetted frost flowers and a rough saline surface increase volume scattering. Repeat pass at polar latitudes can be 1 to 3 days depending on orbit geometry.
- MERRA-2 reanalysis (NASA GMAO): Modern-Era Retrospective analysis for Research and Applications, version 2. Provides 3-hourly wind fields at 0.5 x 0.625 degree resolution. Used to drive aerosol flux calculations from polynya surface area estimates: wind speed over open water determines sea-spray production rate, and wind direction determines transport pathway into the free troposphere.
- CALIOP (CALIPSO): Cloud-Aerosol Lidar with Orthogonal Polarisation aboard CALIPSO. Operates at 532 nm and 1064 nm with a vertical resolution of 30 m in the troposphere. Provides direct detection of elevated aerosol layers downwind of polynyas, distinguishing sea-salt from other aerosol types by depolarisation ratio. Coverage is limited to the narrow 70 m ground track, so coincident overpasses with active polynyas are infrequent but highly diagnostic.
What frost flowers are and why they matter
When air temperatures drop well below freezing over newly exposed polar ocean, thin nilas ice forms within hours. The surface is brine-wetted and highly porous. Water vapour migrates upward through the ice and deposits as centimetre-scale ice crystals called frost flowers. These structures wick brine from the ice surface by capillary action, reaching salinities of 100 parts per thousand or more, roughly three times that of seawater.
The chemistry matters beyond the physics. Frost flowers are a primary source of reactive bromine in the polar boundary layer. Photolysis of bromine compounds triggers catalytic ozone destruction cycles that can reduce surface-level ozone to near zero within days during polar spring. These tropospheric ozone depletion events (ODEs) were first documented in the Arctic in the 1980s and have since been observed in both hemispheres. The aerosol loading from frost flowers and open-water sea spray also affects cloud microphysics and the polar radiation budget.
What the backscatter anomaly gives away
Mature first-year and multi-year sea ice returns moderate C-band backscatter, typically in the range of -15 to -10 dB in HH polarisation. Nilas ice covered in frost flowers is different. The rough, brine-saturated surface and the volume scattering from the crystal matrix push backscatter values anomalously high, sometimes to -8 dB or above in Sentinel-1 Extra Wide Swath data. This makes frost-flower-covered nilas distinguishable from both open water (low backscatter) and older ice (moderate, smoother backscatter), provided the analyst accounts for wind roughening of open water, which can produce similar high-backscatter returns.
The disambiguation comes from AMSR2. The polarisation ratio at 18.7 GHz is sensitive to ice concentration and ice type. Nilas ice has a distinct brightness temperature signature that separates it from open water and from consolidated ice. Combining the SAR spatial detail (5 to 40 m depending on acquisition mode) with AMSR2 ice-type classification (3 to 35 km depending on frequency) allows mapping of frost-flower extent at sub-kilometre precision, within the limits of the SAR resolution, while the passive microwave data anchors the ice-type interpretation.
Estimating aerosol flux: the method and its honest limits
The standard approach fuses three inputs. First, polynya area and frost-flower extent are mapped from the SAR and AMSR2 fusion described above. Second, MERRA-2 10-metre wind speeds over the open-water fraction drive a sea-spray source function, typically a whitecap-based parameterisation such as that of Monahan et al. (1986) or later refinements, to estimate sea-salt aerosol production rate in particles per square metre per second across the size spectrum from 0.1 to 10 micrometres. Third, frost-flower sublimation flux is estimated from the surface area of frost-flower coverage and ambient vapour pressure deficit, drawing on published laboratory and field measurements of frost-flower brine composition.
The limits are real and should be stated plainly. MERRA-2 wind fields at 0.5 degree resolution cannot resolve the sharp gradients that occur at polynya edges, where katabatic winds can accelerate dramatically over short distances. The sea-spray source functions carry uncertainties of a factor of two or more. Frost-flower coverage fraction within a SAR pixel is inferred, not directly measured; a pixel that appears frost-flower-bright may be partially open water or partially consolidated ice. CALIOP overpasses that happen to cross an active polynya downwind provide the only direct aerosol column validation, and these coincidences are rare. The flux estimates are therefore best treated as order-of-magnitude constraints on the polar aerosol budget rather than precise emission inventories.
The CALIPSO validation problem
CALIOP's 532 nm channel detects sea-salt aerosol plumes with high sensitivity, and the depolarisation ratio distinguishes spherical sea-salt droplets from non-spherical dust or ice crystals. In principle, a CALIPSO overpass downwind of a polynya shortly after a wind event should show an elevated marine aerosol layer in the lowest 1 to 2 km of the troposphere. In practice, the 70 m ground track means that most polynya events are never sampled directly. Studies using CALIPSO data over the Weddell Sea and the Arctic Ocean have documented such layers, but the statistical sample is thin.
This is not a reason to abandon the lidar data. It is a reason to treat every coincident overpass as a high-value event and to archive the geometry carefully. When Sentinel-1 identifies an active frost-flower field and MERRA-2 shows the wind trajectory, a CALIPSO pass within 24 hours and 500 km downwind constitutes a natural experiment. Systematic collocation of these three datasets over multi-year archives is where the most defensible flux estimates come from.
Practical scope and what the analysis cannot replace
Satellite-derived aerosol flux estimates from polynyas are useful for two communities: atmospheric chemists building polar ozone depletion models, and climate scientists assessing aerosol forcing in the polar radiation budget. For the first group, the key output is bromine loading as a function of frost-flower extent and season. For the second, it is the size-resolved sea-salt mass flux and its transport altitude.
Neither group should treat the satellite product as a substitute for in-situ chemistry. Frost-flower brine composition varies with ice age, temperature, and biological activity in ways that remote sensing cannot resolve. The bromine yield per unit frost-flower area is still an active research question. What the satellite fusion provides is the spatial and temporal context that field campaigns, by their nature, cannot: continuous monitoring across hundreds of kilometres of coastline through the polar night, when ships cannot operate and aircraft range is limited. Satellize runs this kind of multi-source fusion on open constellations as a standing capability, with the same data-integration approach used in its Tonga crop-estimation programme applied here to cryospheric inputs rather than agricultural ones.
Typical figures
| SAR spatial resolution (Sentinel-1 IW mode) | 5 x 20 m ground range (multi-looked to ~10 x 10 m in standard products) |
| Passive microwave resolution (AMSR2 at 36.5 GHz) | ~7 x 12 km footprint |
| Passive microwave resolution (AMSR2 at 89 GHz) | ~3 x 5 km footprint |
| Sentinel-1 revisit at polar latitudes | 1 to 3 days depending on orbit geometry and acquisition plan |
| AMSR2 revisit | Daily global coverage; twice daily at high latitudes |
| MERRA-2 wind field resolution and latency | 0.5 x 0.625 degree, 3-hourly; reanalysis available with ~2-month lag; near-real-time GEOS-FP available at similar resolution |
| CALIOP vertical resolution (troposphere) | 30 m; horizontal along-track sampling 333 m; swath width 70 m |
| Minimum detectable frost-flower extent (SAR-based) | Patches of order 100 m x 100 m detectable in backscatter anomaly; subpixel fraction not resolved |
| Archive depth | Sentinel-1: from 2014; AMSR2: from 2012; MERRA-2: from 1980; CALIOP: from 2006 |
| Aerosol flux estimate uncertainty | Factor of 2 to 5 depending on wind parameterisation and frost-flower coverage fraction assumptions |
Analytics Satellize can run
| Frost-flower extent map | Threshold and gradient classification of Sentinel-1 C-band backscatter anomaly, cross-validated against AMSR2 18.7 and 36.5 GHz polarisation ratio to confirm nilas ice type | GeoTIFF raster and vector polygon layer, per SAR acquisition, with ice-type confidence class |
| Polynya open-water area time series | AMSR2 ice-concentration retrieval (Bootstrap or NASA Team algorithm) masked to coastal polynya regions, with SAR-derived edge refinement | Daily area time series in CSV with uncertainty bounds, per named polynya region |
| Sea-salt aerosol flux estimate | Whitecap-based sea-spray source function applied to MERRA-2 10-m wind speed over SAR-mapped open-water fraction; size-resolved particle flux integrated to total mass flux | Gridded NetCDF flux field at MERRA-2 resolution, with per-event summary statistics in tabular report |
| Bromine loading index | Frost-flower area multiplied by published brine salinity and bromine enrichment factors from peer-reviewed field studies; scaled by sublimation rate from ambient vapour pressure deficit in MERRA-2 | Seasonal bromine flux index per polynya region, delivered as tabular report with methodology annex |
| CALIOP coincidence log and aerosol layer extraction | Automated collocation of CALIPSO ground-track geometry against active polynya polygons within 24-hour and 500 km downwind window; extraction of Level 2 aerosol layer products for matched overpasses | Per-event collocation report with CALIOP aerosol optical depth and layer height, linked to corresponding SAR and AMSR2 imagery |
| Wind-transport trajectory overlay | MERRA-2 3-hourly wind fields used to compute 48-hour forward trajectories from polynya centroid using standard kinematic trajectory method (analogous to HYSPLIT approach) | KML or GeoJSON trajectory layer per event, for overlay in client GIS environment |
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.