Sea-ice extent monitoring for Arctic and sub-Arctic fisheries access
Passive microwave and SAR satellites track sea-ice advance and retreat daily across the Barents, Bering and Okhotsk seas, giving fleet operators and quota managers objective evidence of when grounds open and close.
Sensors
- AMSR2 (GCOM-W1, JAXA): Passive microwave radiometer operating at 6.9 to 89 GHz. Delivers daily, all-weather ice-concentration maps at 3.125 km resolution (89 GHz channel) to 25 km (lower frequencies). Penetrates cloud and polar darkness; cannot resolve individual leads narrower than a few kilometres.
- SSMIS (DMSP constellation, NOAA/DoD): Long-running passive microwave series providing ice-concentration records back to 1979 at 12.5 to 25 km resolution. Continuity makes it the backbone of climatological baseline comparisons for anomaly detection.
- Sentinel-1 A/B SAR (ESA Copernicus): C-band SAR at 5.405 GHz. Extra Wide Swath mode covers 400 km at 20 m resolution, revisiting Arctic latitudes every one to three days. Resolves ice-edge position, lead structure and pressure ridges at scales relevant to vessel routing. Ice-type classification degrades significantly during melt-pond season (roughly June to August) when surface wetness reduces backscatter contrast.
- RADARSAT-2 (MDA): C-band SAR with ScanSAR Wide modes covering 500 km swaths at 50 m resolution and fine-beam modes down to 3 m. Commercially taskable, which allows priority imaging of specific fishing grounds on short notice. Useful for confirming ice-edge position before a fleet commits to a route.
Why the ice edge is a fishing season boundary
Sea ice does not merely obstruct navigation. Its seasonal retreat drives the biological productivity that makes Arctic and sub-Arctic fisheries worth entering at all. In the Barents Sea, ice-edge retreat in spring triggers phytoplankton blooms along the marginal ice zone, concentrating zooplankton and, in turn, capelin, cod and haddock. The Bering Sea pollock fishery is similarly structured around the position of the ice edge in late winter and early spring. Fleet operators who know where the edge is, and where it is heading, can stage vessels and plan quota uptake rather than waiting at port for conditions to become obvious from radio reports.
The challenge is that the edge moves on timescales of days, varies by hundreds of kilometres between years, and is not uniform: leads, polynyas and refrozen patches create a mosaic that looks open on a coarse map but may be impassable in practice. Satellite observation is the only practical way to characterise this at the spatial and temporal resolution that fleet planning requires.
What passive microwave gives you, and what it cannot
AMSR2 and SSMIS produce ice-concentration estimates by exploiting the large difference in microwave emissivity between open water and sea ice. Standard algorithms such as Bootstrap and NASA Team convert brightness temperatures at multiple frequencies into a percentage ice cover per grid cell, daily, regardless of cloud or solar illumination. That combination of all-weather availability and daily revisit is genuinely difficult to replicate with any other sensor type.
The cost is spatial resolution. At 25 km grid spacing, the ice edge is a zone rather than a line. A polynya two kilometres wide is invisible. Thin new ice and open water can be confused in certain conditions. For strategic planning, knowing that a region is 30 percent ice-covered is useful; for routing a vessel through it, you need something finer.
The SSMIS archive extends to 1979, making it possible to assess whether a given season's ice retreat is early, late or within normal range relative to four decades of observed variability. That climatological context is what turns a current observation into an actionable anomaly.
SAR resolves the edge. It does not always classify the ice.
Sentinel-1 in Extra Wide Swath mode and RADARSAT-2 in ScanSAR modes provide ice-edge positions accurate to roughly the pixel scale, 20 to 50 m depending on mode. At that resolution, leads wide enough for a vessel to transit are visible, pressure ridges that would damage a hull show up as bright linear features, and the boundary between consolidated pack and broken brash ice can be drawn with confidence.
Ice-type classification is a different matter. In winter, multi-year ice and first-year ice have distinct C-band backscatter signatures, and automated classifiers trained on SAR data perform reasonably well. From roughly June onward, melt ponds form on the ice surface, saturating it with liquid water and suppressing the backscatter contrast that distinguishes ice types. During this period, SAR reliably shows you where ice is but is less reliable about what kind. Operational users should treat ice-type maps produced in melt-pond season as indicative rather than definitive, and weight vessel-reported observations more heavily.
Combining passive microwave concentration with SAR edge position is standard practice in operational ice services. The coarse sensor sets the regional picture; the fine sensor pins the edge for the route that matters today.
Turning observations into a quota-season calendar
Fleet operators and fisheries managers need different things from the same underlying data. A vessel master wants a current ice-edge position and a short-range forecast of whether a lead will stay open long enough to complete a tow. A quota manager wants to know, weeks in advance, whether the season is tracking early or late relative to the historical distribution, so that vessel scheduling and processing capacity can be aligned.
Both needs can be served from the same satellite record. Daily AMSR2 composites, compared against the SSMIS climatology for the same calendar date, give an anomaly index: how many standard deviations early or late is this year's retreat? That index, updated daily, is a straightforward leading indicator for season-opening decisions. SAR-derived edge positions, delivered as GIS layers with timestamps, feed directly into voyage-planning software.
Ice-extent trend analysis across the Barents, Bering and Okhotsk seas also supports stock-assessment inputs, since ice-edge position in winter and spring is a documented covariate for recruitment in several commercially important species. Satellize builds these multi-sensor fusion pipelines on open constellations, adding commercial SAR tasking where a client needs priority coverage of a specific ground.
Honest limits and what fills the gaps
No satellite system provides a real-time, high-resolution, ice-type-certain view of Arctic waters simultaneously. Passive microwave is coarse. SAR has revisit gaps of one to three days even at high latitudes, longer in lower-priority tasking queues. Optical imagery is blocked by cloud for extended periods, and polar night removes it entirely for months at a time. Any operational system must be designed around these gaps, not in denial of them.
In practice, the combination of daily passive microwave for concentration and SAR for edge position covers the majority of operational needs. Where SAR revisit is insufficient, ice-drift models driven by satellite-derived wind fields can propagate the last observed edge position forward in time, with uncertainty bounds that widen honestly as the forecast extends. Ice services operated by NOAA, the Norwegian Meteorological Institute and the Canadian Ice Service publish operational products that provide a useful independent check on any analytical pipeline.
Typical figures
| Passive microwave spatial resolution | 3.125 km (AMSR2 89 GHz) to 25 km (SSMIS / AMSR2 low-frequency channels) |
| SAR spatial resolution | 20 m (Sentinel-1 EW mode) to 50 m (RADARSAT-2 ScanSAR Wide); fine-beam modes to 3 m but narrow swath |
| Passive microwave revisit | Daily global coverage; twice-daily at high latitudes due to orbital overlap |
| SAR revisit at Arctic latitudes | 1 to 3 days for Sentinel-1; commercial tasking on RADARSAT-2 can reduce to same-day for priority areas |
| Sensor frequency / bands | Passive microwave: 6.9, 10.65, 18.7, 23.8, 36.5, 89 GHz (AMSR2); SAR: C-band 5.405 GHz (Sentinel-1, RADARSAT-2) |
| Minimum detectable lead width (SAR) | Approximately 40 to 100 m in EW mode; narrower leads below detection threshold |
| Ice-concentration accuracy (passive microwave) | Typically ±5 to 10 percent concentration in consolidated pack; larger errors in marginal ice zone and melt-pond season |
| Archive depth | Passive microwave: SSMIS/SSMI back to 1979; AMSR2 from 2012. Sentinel-1: 2014 to present |
| Data latency | AMSR2 near-real-time products: 2 to 6 hours after overpass. Sentinel-1 GRD: typically available within 1 hour of downlink via Copernicus Dataspace |
| Delivery formats | GeoTIFF ice-concentration grids, GeoJSON / Shapefile ice-edge vectors, NetCDF anomaly time series, PDF daily briefing charts |
Analytics Satellize can run
| Daily ice-concentration anomaly map | Bootstrap or NASA Team passive microwave algorithm applied to AMSR2 brightness temperatures, differenced against SSMIS climatological mean and standard deviation for the same calendar date | GeoTIFF grid and one-page PDF briefing, updated daily, covering the Barents, Bering or Okhotsk sea as specified |
| SAR ice-edge vector | Automated edge detection on Sentinel-1 EW backscatter imagery, quality-checked against passive microwave concentration; manual review flagged during melt-pond season | GeoJSON polyline with timestamp and confidence flag, delivered within 3 hours of SAR scene availability |
| Ice-type classification (winter season) | Supervised classification of C-band SAR backscatter and texture features distinguishing multi-year from first-year ice; product issued October to May only, with explicit melt-season suspension notice | GeoTIFF classification layer with per-class confidence scores |
| Season-opening forecast index | Regression of current passive microwave anomaly against historical SSMIS record to estimate probable date of ice retreat below a user-specified concentration threshold for a defined fishing ground polygon | Weekly probabilistic forecast table (10th, 50th, 90th percentile opening date) in PDF and CSV |
| Lead and polynya mapping for vessel routing | High-contrast feature extraction from RADARSAT-2 fine or ScanSAR modes over a tasked area, identifying navigable water bodies above a minimum width threshold | GeoJSON polygon layer of open-water features, compatible with standard voyage-planning software |
| Multi-year ice-extent trend report | Time-series analysis of SSMIS and AMSR2 ice-extent data for a defined sea region, with trend decomposition and year-on-year comparison | Annual PDF report with figures, suitable for inclusion in stock-assessment environmental covariate documentation |
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.