Coastal algal bloom alerts for beach management

Ocean-colour satellites detect chlorophyll-a and phycocyanin proxies at coastal scales, giving beach managers days of warning before harmful algal blooms reach swimmers. This page explains the sensors, methods, honest limits and analytic products.

Sensors

Sentinel-3 OLCI (Ocean and Land Colour Instrument): 21 spectral bands from 400 to 1020 nm at 300 m spatial resolution, with a swath of 1270 km and a revisit of roughly 1.4 days at the equator (better at higher latitudes with two satellites). The 665 nm and 708 nm band pair is the standard basis for the Maximum Chlorophyll Index (MCI) algorithm, which isolates the red-edge fluorescence peak used to flag dense surface blooms.
Sentinel-2 MSI (MultiSpectral Instrument): 10 m and 20 m bands covering the visible and near-infrared, with a revisit of 5 days at the equator using both Sentinel-2A and 2B. The 705 nm red-edge band enables a chlorophyll-a index in optically complex coastal water, though the sensor was designed for land and its ocean-colour atmospheric correction is less mature than OLCI's. Useful for resolving bloom boundaries in small embayments that OLCI cannot resolve.
MODIS Aqua ocean colour: 36 bands, 250 m to 1 km depending on band, daily global coverage. The standard NASA OC3M chlorophyll-a algorithm operates in the blue-green ratio (443/547 nm). Aqua has been operational since 2002, giving a 20-plus-year coastal time series, though the sensor is well past its design life and data quality has degraded in some bands.
VIIRS SNPP and NOAA-20 ocean colour: VIIRS carries a dedicated 750 m ocean-colour band suite (M-bands) and a 375 m day-night band. It is the operational successor to MODIS for global chlorophyll-a and coloured dissolved organic matter (CDOM) retrieval, with daily coverage. NOAA's CoastWatch programme distributes near-real-time VIIRS ocean-colour composites.
Copernicus Marine Service (CMEMS) biogeochemical model: CMEMS delivers daily 4 km global biogeochemical reanalysis and forecast fields, including chlorophyll-a, nitrate and surface current, by assimilating satellite ocean-colour observations into coupled physical-biogeochemical models. Forecast horizon is typically 5 days. Not a sensor, but the operational synthesis layer that converts satellite observations into predictive bloom guidance.

What ocean colour actually measures, and what it does not

Phytoplankton absorb blue light and reflect green. Satellites infer chlorophyll-a concentration by comparing water-leaving radiance in the blue and green parts of the spectrum, after removing the much larger atmospheric signal, which accounts for roughly 90 percent of the top-of-atmosphere radiance over open ocean. In clear offshore water this works well. In coastal water, CDOM from river runoff and suspended sediment from wave action both absorb and scatter light in ways that confound the standard blue-green ratio algorithms, pushing retrieved chlorophyll-a high or low by a factor of two or more.

Phycocyanin, the pigment that distinguishes cyanobacteria (the toxin-producing group of most concern to public health) from harmless green algae, has an absorption peak near 620 nm. OLCI's band at 620 nm, used alongside the 665 nm and 708 nm bands, provides a phycocyanin proxy index. This is not a direct toxin measurement. Microcystin, saxitoxin and other cyanotoxins require water sampling to quantify. Satellite data narrows where to sample; it does not replace the laboratory.

From spectral signal to beach advisory: the operational chain

A practical beach-alert system runs in three steps. First, daily or near-daily OLCI or VIIRS scenes are ingested, atmospherically corrected (the Polymer or C2RCC processors are the standard open-source options for coastal water), and converted to chlorophyll-a and phycocyanin index maps. Second, those maps are compared against authority-defined thresholds. The WHO recreational water guideline, for example, uses 10 µg/L chlorophyll-a as a low-probability-of-adverse-health-effects threshold and 50 µg/L as a high-probability threshold. Third, threshold exceedances trigger a tiered advisory: advisory notice, beach closure, or all-clear.

CMEMS biogeochemical forecasts extend the horizon. By combining the current satellite-observed bloom location with modelled surface currents, a manager can estimate where a bloom will be in 48 to 72 hours. This is particularly useful for tidal systems where a bloom offshore on Monday may reach a bathing beach by Wednesday. The forecast skill degrades beyond about five days and is sensitive to wind-driven mixing events that can disperse or concentrate a bloom rapidly.

Aerosol exposure is a secondary concern. Dense cyanobacterial surface scums can release aerosolised toxins and odorous compounds when wave action breaks up the mat. Satellite data does not directly measure airborne toxin concentrations, but bloom proximity to shore combined with onshore wind forecasts gives a reasonable qualitative basis for issuing aerosol-exposure advisories to people with respiratory conditions.

Resolution floors and the small-bay problem

OLCI's 300 m pixel is adequate for open embayments and coastlines with fetch of at least a kilometre, but it will miss a bloom confined to a small harbour or lagoon. Sentinel-2 MSI at 10 to 20 m resolves those features, but its atmospheric correction over water is less reliable, its revisit is 5 days rather than daily, and it has no dedicated ocean-colour bands below 443 nm. The practical answer for small enclosed water bodies is to use OLCI for daily surveillance and task Sentinel-2 for confirmation when OLCI signals an anomaly near a small bay.

Cloud cover is the other hard limit. Coastal regions with persistent marine stratus, common along upwelling coasts in summer, can go days without a usable optical scene. Synthetic aperture radar can detect surface films associated with very dense blooms through cloud, but the signal is ambiguous and not suitable as a primary chlorophyll indicator. Managers relying on satellite alerts in persistently cloudy climates need a contingency sampling protocol for cloud-blocked periods.

NOAA's harmful algal bloom programme as a public benchmark

NOAA's National Centers for Coastal Ocean Science (NCCOS) has operated a Harmful Algal Bloom Monitoring System for US coastal waters for over a decade, using MODIS and VIIRS chlorophyll-a and phycocyanin products to issue weekly bulletins for regions including the Gulf of Mexico, Great Lakes and Pacific Northwest. These bulletins are publicly available and represent the operational standard against which any commercial or government bloom-alert system should be benchmarked.

The NOAA system is explicit about what it cannot do: it cannot speciate the bloom from orbit, it cannot measure toxin concentration, and it is subject to cloud gaps. It issues probability-of-presence statements, not certainties. That epistemic honesty is the correct framing for any satellite-based bloom advisory. A satellite flag means "sample here now", not "close the beach".

Building a sovereign coastal monitoring capability

For a national authority that wants to own its bloom-alert capability rather than depend on a third-party bulletin, the architecture is straightforward in principle: ingest Copernicus open-access OLCI data, run a validated atmospheric correction and chlorophyll-a retrieval, compare outputs to national regulatory thresholds, and push alerts to beach managers via an API or GIS dashboard. The complexity is in the validation. A retrieval algorithm tuned for the North Sea will perform differently in turbid tropical coastal water. Local in-situ matchup datasets, even small ones of 50 to 100 paired satellite-sample observations, are necessary to characterise bias and uncertainty for the specific water optical type.

Satellize runs this kind of analytics stack on open constellations, adapting retrieval parameters to local water optical conditions. The Kingdom of Tonga crop-estimation programme is a different domain, but the underlying workflow, open-data ingest, local validation, threshold-triggered alert, is the same pattern applied here. A coastal authority wanting to scope a similar system should start with a 12-month historical validation study against any existing water-quality monitoring records before committing to an operational deployment.

Typical figures

Primary spatial resolution (OLCI)	300 m per pixel
High-resolution confirmation (Sentinel-2 MSI)	10–20 m, with caveats on ocean-colour atmospheric correction
Revisit frequency	~1.4 days (OLCI, two-satellite); 5 days (Sentinel-2); daily (MODIS/VIIRS)
Key spectral bands for bloom detection	620 nm (phycocyanin proxy), 665 nm, 708 nm (MCI red-edge), 443/547 nm (OC3M blue-green ratio)
Minimum detectable chlorophyll-a concentration	Approximately 1–2 µg/L in clear coastal water; degrades to 5–10 µg/L or higher in turbid or CDOM-rich water
CMEMS forecast horizon	Up to 5 days; skill degrades after day 3 in wind-event conditions
Archive depth	MODIS Aqua: 2002–present; Sentinel-3 OLCI: 2016–present; VIIRS SNPP: 2012–present
Data latency (near-real-time products)	OLCI NRT: typically 3 hours after overpass; CMEMS daily analysis: same-day
Cloud limitation	All optical sensors blind under cloud; SAR not a reliable substitute for chlorophyll retrieval
Delivery formats	NetCDF (standard ocean colour), GeoTIFF, OGC WMS/WFS, JSON alert feed

Analytics Satellize can run

Daily chlorophyll-a concentration map	Atmospheric correction (Polymer or C2RCC) plus blue-green ratio or red-edge band algorithm applied to OLCI or VIIRS Level-1 data	GeoTIFF raster with µg/L values, clipped to national coastal zone, updated each overpass
Phycocyanin index layer	Three-band phycocyanin proxy index using OLCI 620, 665 and 708 nm bands, flagging probable cyanobacterial dominance	GIS layer with categorical risk classes (low/moderate/high), delivered alongside chlorophyll-a map
Beach-specific threshold alert	Zonal statistics over defined beach polygons, compared against WHO or national authority chlorophyll-a thresholds (10 µg/L and 50 µg/L reference levels)	JSON alert feed to beach-management dashboard, with tiered advisory status per named beach
48–72 hour bloom trajectory forecast	CMEMS surface current and biogeochemical forecast fields combined with observed bloom centroid and extent to project probable position	Daily PDF bulletin with map and confidence statement, issued to public health authority
Seasonal bloom climatology and anomaly score	Monthly mean chlorophyll-a from multi-year archive (MODIS/VIIRS/OLCI), with current-year anomaly expressed as standard deviations from the historical mean	Annual trend report with per-beach anomaly time series, suitable for regulatory reporting
Satellite-to-sample matchup validation report	Comparison of satellite-retrieved chlorophyll-a against in-situ fluorometry or HPLC records provided by the authority, with bias and RMSE statistics by season and water type	One-off technical report establishing retrieval accuracy for the specific coastal domain before operational deployment

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.