Southern Ocean chlorophyll and iron-fertilisation response mapping
The Southern Ocean is chronically iron-limited, yet icebergs and shelf sediments seed localised blooms detectable from orbit. Retrieving chlorophyll-a reliably demands sensors and algorithms built for optically complex, cloud-plagued sub-Antarctic waters.
Sensors
- Sentinel-3 OLCI: 21 spectral bands from 400 to 1020 nm at 300 m spatial resolution, global revisit roughly every two days per satellite (faster with Sentinel-3A and 3B combined). The 665 nm and 708 nm red-edge bands are critical for the fluorescence-based algorithms that outperform OC3 in high-latitude, sediment-influenced waters.
- MODIS-Aqua: 36 bands; ocean-colour bands at 250 m to 1 km resolution depending on channel. Daily global coverage at 1 km for chlorophyll products. The archive runs from 2002, making it the backbone of Southern Ocean climatology. Sensor degradation over two decades introduces calibration drift that must be corrected before trend analysis.
- PACE OCI: NASA's PACE satellite (launched February 2024) carries the Ocean Colour Instrument, which delivers hyperspectral coverage from 340 to 890 nm at roughly 1 km resolution with daily global revisit. Its ultraviolet and near-infrared range improves aerosol correction and phytoplankton community discrimination in Southern Ocean conditions.
- Landsat 8/9 OLI: 30 m spatial resolution across six reflective bands. Useful for resolving fine-scale bloom structure in coastal and shelf zones near sub-Antarctic islands, but 16-day revisit and narrow swath (185 km) make it impractical as a primary Southern Ocean sensor. Best used to validate bloom boundaries identified in coarser imagery.
Why the Southern Ocean defeats standard chlorophyll algorithms
The OC3 algorithm, which estimates chlorophyll-a from the ratio of blue to green water-leaving radiance, was calibrated largely on open-ocean Case-1 waters where phytoplankton dominate the optical signal. The Southern Ocean breaks that assumption in several ways. Coloured dissolved organic matter (CDOM) released from melting sea ice and iceberg meltwater absorbs blue light in a way that mimics high chlorophyll concentrations. Suspended sediments from shelf erosion and iceberg scour scatter light at wavelengths OC3 treats as chlorophyll signal. The result is systematic overestimation of chlorophyll-a in iron-fertilised shelf and marginal-ice-zone waters, sometimes by a factor of two to four according to published validation studies using in-situ HPLC measurements.
Algorithms that exploit the chlorophyll fluorescence peak near 683 nm, such as the fluorescence line height (FLH) product available from MODIS-Aqua, or the OLCI-based maximum chlorophyll index (MCI), are less susceptible to these interferences. They measure emitted light rather than reflected light, so CDOM and sediment contributions are largely decoupled. The trade-off is sensitivity: FLH saturates at high bloom concentrations and performs poorly at very low chlorophyll levels, roughly below 0.1 mg m⁻³, which matters in the oligotrophic Southern Ocean gyre.
Compositing through cloud: a necessary compromise
Cloud cover exceeds 80 percent of individual scenes over much of the Southern Ocean, particularly south of 50°S. A single overpass is almost never usable. Eight-day and monthly compositing, selecting the maximum or median valid pixel across all cloud-free observations in the window, is standard practice and is used in the official NASA MODIS Level-3 products. The problem is temporal smearing: an iceberg-fertilised bloom may peak and decline within ten to fourteen days, so a monthly composite can dilute the signal to near-background levels or merge it with adjacent water masses.
A more targeted approach uses the known drift trajectories of large tabular icebergs (tracked independently by SAR and altimetry) to define a downstream search window, then composites only the pixels within that window over a biologically plausible response time of seven to twenty-one days after the iceberg's passage. This reduces the compositing area dramatically and preserves temporal resolution. Cloud-gap-filling via spatial interpolation or machine-learning inpainting introduces artefacts and should be treated as a visualisation aid rather than a quantitative product.
Reading the iron signal: what a bloom's shape and timing reveal
Iron fertilisation from icebergs produces a characteristic downstream plume. Meltwater from a drifting iceberg releases dissolved iron and iron-bearing particles continuously; the bloom therefore appears as an elongated streak aligned with the prevailing current, not a symmetric patch. Studies of large Antarctic tabular icebergs, including published work on icebergs in the Scotia Sea, have documented chlorophyll anomalies extending 50 to 200 km downstream, with peak concentrations of 1 to 5 mg m⁻³ against a background of 0.1 to 0.3 mg m⁻³. The bloom typically lags the iceberg's position by several days as phytoplankton divide and accumulate.
Shelf sediment fertilisation, particularly around the Kerguelen Plateau and South Georgia, produces broader and more persistent blooms because the iron source is fixed and replenished by upwelling. The Kerguelen bloom is one of the largest natural iron-fertilisation signals on Earth and is detectable in every annual MODIS climatology. Distinguishing sediment-driven from iceberg-driven fertilisation requires combining the chlorophyll anomaly map with a current-field estimate and an iceberg position record. Neither source can be confirmed from ocean colour alone.
PACE OCI and the next generation of community discrimination
Chlorophyll-a concentration tells you biomass is present. It does not tell you which phytoplankton are growing. In iron-fertilised Southern Ocean waters this matters: diatoms, which dominate productive blooms and drive efficient carbon export, have a distinct spectral absorption signature compared with cryptophytes or haptophytes that colonise the bloom's edges. PACE OCI's continuous hyperspectral coverage allows phytoplankton absorption spectra to be decomposed into functional type contributions using established spectral unmixing methods, something that MODIS's discrete band set cannot support reliably.
PACE data have been publicly available only since mid-2024, so Southern Ocean climatologies built on it are thin. Comparison against the MODIS-Aqua archive requires careful inter-calibration; early published assessments suggest good agreement in open-ocean waters but residual differences in high-latitude scenes where atmospheric correction is uncertain. Treat PACE-derived community composition products as promising rather than operationally validated for the Southern Ocean until multi-year validation accumulates.
Honest limits and what ground truth can fix
Atmospheric correction is the single largest source of error in Southern Ocean ocean-colour retrievals. The standard MODIS and OLCI correction algorithms assume the ocean is dark in the near-infrared, which fails in turbid shelf waters. The POLYMER algorithm and its successors improve performance in these conditions and also partially recover data in sun-glint-affected pixels, but they require careful parameterisation and do not eliminate uncertainty. Validation against in-situ fluorometric or HPLC chlorophyll measurements from Argo floats with optical sensors, or from research vessel transects, remains essential for any quantitative product delivered to a client.
Satellize runs chlorophyll anomaly and bloom-tracking products on Sentinel-3 OLCI and MODIS-Aqua open data, applying fluorescence-based retrievals and iceberg-referenced compositing windows. The workflow is similar in structure to the approach used in our Tonga crop-estimation programme: define a biologically meaningful spatial and temporal window, select the appropriate algorithm for local optical conditions, and report uncertainty explicitly alongside the estimate. If your programme requires absolute chlorophyll calibration rather than anomaly detection, budget for at least one in-situ validation campaign.
Typical figures
| Primary sensor spatial resolution | 300 m (Sentinel-3 OLCI); 1 km (MODIS-Aqua ocean-colour bands); ~1 km (PACE OCI); 30 m (Landsat 8/9 OLI, coastal only) |
| Revisit frequency | ~1.4 days combined (Sentinel-3A+3B); daily (MODIS-Aqua); daily (PACE OCI); 16 days (Landsat 8/9) |
| Effective cloud-free revisit (Southern Ocean) | 8 to 30 days for a usable composite; individual cloud-free scenes rare south of 50°S |
| Key spectral bands | 443, 490, 560, 665, 674, 681, 708 nm (OLCI); 443, 488, 547, 667, 678 nm (MODIS); 340–890 nm continuous (PACE OCI) |
| Chlorophyll-a detection range (algorithm-dependent) | ~0.05 to >50 mg m⁻³; FLH/MCI methods less reliable below 0.1 mg m⁻³ |
| Minimum detectable bloom anomaly | Approximately 0.2 mg m⁻³ above local background in 8-day composite at 1 km (indicative; validation-dependent) |
| Archive depth | MODIS-Aqua from 2002; Sentinel-3A OLCI from 2016; Landsat 8 from 2013; PACE OCI from February 2024 |
| Standard latency (Level-3 products) | MODIS NRT: ~3 hours; Sentinel-3 NRT: ~3 hours; standard science products: 1–2 days |
| Delivery formats | NetCDF-4, GeoTIFF, Cloud-Optimised GeoTIFF; time-series CSV for point or polygon extracts |
Analytics Satellize can run
| Chlorophyll-a concentration map (fluorescence-based) | OLCI Maximum Chlorophyll Index or MODIS Fluorescence Line Height retrieval, replacing OC3 for optically complex waters | 8-day or monthly GeoTIFF with per-pixel uncertainty layer; anomaly relative to MODIS climatological baseline |
| Iceberg-referenced downstream bloom detection | Chlorophyll anomaly computed within a current-advected search polygon centred on tracked iceberg positions; composite window matched to 7–21 day biological response time | Bloom presence/absence flag and peak concentration per iceberg event; GIS polygon layer with confidence score |
| Kerguelen and South Georgia shelf bloom seasonality report | Annual phenology extraction from MODIS-Aqua Level-3 monthly climatology; bloom onset, peak and decline dates derived from time-series inflection points | Annual PDF report with time-series plots and comparison against prior-year and long-term mean |
| Phytoplankton functional type fraction (PACE) | Spectral absorption decomposition using published matrix-factorisation or Gaussian-decomposition methods applied to PACE OCI hyperspectral Rrs | Gridded NetCDF with diatom, haptophyte and mixed-community fraction per pixel; flagged as experimental pending extended validation |
| Cloud-gap composite with quality mask | Maximum valid-pixel compositing over user-defined window; POLYMER atmospheric correction for turbid-water and sun-glint scenes | Cloud-optimised GeoTIFF with pixel-count and quality-flag layers; unsuitable for quantitative use where fewer than three valid observations contribute |
| Iron-source attribution assessment | Overlay of chlorophyll anomaly map with ocean-current climatology (AVISO or CMEMS) and iceberg drift record to assign probable fertilisation source | Analyst report with annotated maps; explicitly notes cases where attribution is ambiguous |
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.