Pelagic Insight Research

Sea Surface Temperature
What the Satellites Actually See

SST is the most important dataset in offshore fishing. Understanding where the data comes from, what it can and cannot show, and why “high-definition” often means “statistically fabricated” is the difference between finding fish and trusting a pretty picture.

Eric Whyne · Data Machines · April 5, 2026

📄 Download PDF
1978
first satellite SST
~70%
ocean obscured by clouds
750 m
best IR pixel resolution
Scroll to explore

A Brief History of Measuring the Ocean's Skin

Humans have been measuring sea surface temperature for over 200 years. The methods have changed dramatically; the fundamental challenge has not.

1780s
Bucket Thermometers
Sailors hauled seawater in canvas buckets and measured temperature by hand. Systematic records began with the British Royal Navy.
1870s
Engine Intake Readings
Steam-powered ships began measuring water temperature at engine cooling intakes. Warmer than bucket readings by 0.3-0.7°C, causing a known bias in the historical record.
1950s
Expendable Bathythermographs
XBTs launched from ships measured temperature profiles as they sank. The first systematic subsurface measurements, but surface coverage remained sparse and ship-track dependent.
1978
Satellite Era Begins
TIROS-N launched with the first space-based infrared SST sensor. For the first time, ocean temperature could be observed synoptically rather than along scattered ship tracks.
1981
AVHRR Operational SST
The Advanced Very High Resolution Radiometer on NOAA polar orbiters began producing operational SST at ~4 km resolution. This instrument class would dominate SST measurement for two decades.
1997
Microwave Breakthrough
TRMM Microwave Imager proved that passive microwave radiometry could measure SST through clouds, at the cost of much coarser resolution (~25 km vs 1 km).
2002
MODIS and the Modern Era
NASA's Aqua satellite carried MODIS, achieving ~1 km SST with improved atmospheric correction. AMSR-E provided complementary all-weather microwave SST on the same platform.
2011+
VIIRS and Multi-Sensor Fusion
Suomi NPP and NOAA-20 carry VIIRS with 750 m resolution, the sharpest operational IR SST today. GHRSST coordinates fusion of all sensors into gap-filled products.

Who Collects SST Data

No single agency owns SST. Data comes from a constellation of satellites operated by different nations, supplemented by buoys, drifters, and ships. The Group for High Resolution SST (GHRSST) coordinates it all.

NOAA / NESDIS
NOAA-20, NOAA-21 (VIIRS)
Infrared
750 m
resolution
Primary US operational SST
NASA
Aqua & Terra (MODIS)
Infrared
1 km
resolution
Science-grade, 20+ year record
ESA / EUMETSAT
Sentinel-3A/3B (SLSTR)
Infrared
1 km
resolution
Copernicus programme
JAXA
GCOM-W (AMSR2)
Microwave
25 km
resolution
Sees through clouds
ESA
SMOS
Microwave
~40 km
resolution
Primarily salinity, SST secondary
NOAA / NDBC
Moored buoys, drifters
In situ
Point
resolution
Ground truth for satellite cal/val
NOAA-20Sentinel-3☁ blockedOcean Surface

Polar-orbiting satellites scan the ocean in swaths. Clouds block infrared sensors entirely, creating data gaps that can persist for days over the same area.

How SST Is Measured

Two fundamentally different approaches, each with a critical tradeoff.

🔴

Infrared Radiometry

Measures thermal radiation emitted by the ocean surface in the 3.7-12 μm wavelength bands. The ocean radiates heat; the satellite detects it.

High resolution: 750 m to 1 km pixels
Detects fine thermal features and temp breaks
Measures true “skin” temperature (top 10-20 μm)
Completely blocked by clouds
Contaminated near coastlines (land radiance)
Aerosols and dust introduce bias
🔵

Microwave Radiometry

Measures natural microwave emission from the ocean surface at 6-10 GHz. Longer wavelengths pass through clouds.

Sees through clouds (non-precipitating)
Day and night measurement
Global all-weather coverage
Coarse resolution: 25-40 km pixels
Cannot resolve temperature breaks
Contaminated by rain and near land

Resolution Comparison

VIIRS Infrared750 m
MODIS Infrared1 km
AMSR2 Microwave25 km

Lower bar = finer resolution. A single microwave pixel covers the area of roughly 1,100 infrared pixels. Temperature breaks are invisible at microwave resolution.

The Practical Constraints

On any given day, approximately 70% of the ocean surface is obscured by clouds. This is not a solvable engineering problem. It is a physical reality.

☁️
~70%

Cloud Obstruction

Infrared sensors cannot see through clouds. Period. In tropical and high-latitude waters, cloud cover can persist for a week or more over the same area.

🌍
12-24 hr

Orbital Gaps

Polar-orbiting satellites take 14-16 orbits per day. At tropical latitudes, adjacent swaths do not overlap, leaving gaps between passes that are only filled on the next orbit.

🌡️
~0.5-1°C

Skin vs. Bulk

IR satellites measure the top 10-20 micrometers of water. Fishermen experience bulk temperature mixed by wind and waves. The difference can exceed 1°C, especially at night.

🏖️
5-10 km

Coastal Contamination

Land emits infrared radiation at different wavelengths than water. Pixels within 5-10 km of coastlines can be contaminated by land signal, making nearshore SST unreliable.

3-24 hr

Processing Lag

Raw satellite passes must be downloaded, calibrated, quality-controlled, and distributed. Near-real-time products arrive 3-6 hours after observation. Science-grade products take a full day.

🌊
up to 3°C

Diurnal Warming

In low-wind conditions, the top meter of ocean can warm 2-3°C during the day and cool at night. A satellite pass at 1 PM and another at 2 AM measure fundamentally different temperatures.

no datano datano data

Simulated single-pass SST image. Gray cells represent cloud-obscured pixels with no data. In the real ocean, these gaps are often concentrated over exactly the areas where weather is most active and fishing is most productive.

How Interpolation Fills the Gaps

To produce a “complete” SST map with no gaps, agencies use statistical methods to estimate what the temperature probably was in places the satellite could not see. This is interpolation, and the result is, by definition, not measured data.

1

Gather Observations

Collect all available satellite passes, buoy readings, ship reports, and Argo float data from the past 1-7 days.

2

Statistical Model

Use Optimal Interpolation (OI) or variational analysis to estimate temperature at every grid point, weighting observations by distance, age, and expected error.

3

Gap-Free Output

Produce a smooth, complete SST field with a value at every pixel. Where no observation existed, the value is a statistical best guess.

Major Interpolated (L4) SST Products

OSTIA (UK Met Office)
Foundation SST analysis
1/20° (~5 km)
grid
Optimal Interpolation
~1 day
latency
MUR (NASA JPL)
Highest resolution L4
1/100° (~1 km)
grid
Multi-Resolution Variational
~1 day
latency
OISST (NOAA)
Climate-grade, long record
1/4° (~25 km)
grid
Optimal Interpolation
~1 day
latency
RSS MW+IR
Microwave fills cloud gaps
9 km
grid
MW+IR fusion
~1 day
latency
Raw Satellite DataGaps visible. Temp break sharp.After InterpolationNo gaps. Temp break smoothed away.

Left: raw satellite data with cloud gaps. The sharp temperature break between warm and cool water is clearly visible. Right: after interpolation, gaps are filled but the temperature break has been smoothed into a gentle gradient. The feature fishermen need most is the feature interpolation destroys.

What This Means to Fishermen

Pelagic species, especially tuna, billfish, mahi, and wahoo, concentrate at temperature boundaries. Not in warm water. Not in cool water. At the edge between them.

Why Temperature Breaks Hold Fish

Temperature breaks are boundaries where different water masses meet. Where Gulf Stream water at 78°F meets shelf water at 72°F, the density difference traps nutrients and creates a convergence zone.

Phytoplankton bloom along these edges. Baitfish feed on phytoplankton. Predators follow the bait. The food chain concentrates along a line that may be only 100-500 meters wide.

A fisherman who can locate a 2°F temperature break in 60 miles of open ocean has transformed a random search into a targeted approach. The temp break is the single most actionable piece of information in offshore fishing.

What Interpolation Does to This

Optimal Interpolation is designed to produce smooth fields. By its mathematical construction, it spreads influence from each data point outward using a correlation function. Sharp gradients are softened. Edges become slopes.

A real temperature break of 4°F across 200 meters becomes a gentle 4°F gradient spread across 20 kilometers in an interpolated product. The break still “exists” in the data, but it is no longer actionable.

The higher the stated resolution of an interpolated product, the more false confidence it provides. A “1 km MUR” SST image looks sharp and detailed, but the underlying analysis smooths features over 10+ km. The pixels are small; the information content is not.

72°F74°F76°F73°F isotherm75°F isotherm🐟🐟🐟🐟🐟

Isotherm lines drawn from raw satellite data show exactly where temperature transitions occur. Fish concentrate along these boundaries. The sharper the line, the stronger the convergence.

100-500 m
Width of a productive temp break
Narrower than one interpolated pixel
2-4°F
Temperature change across the break
Enough to concentrate entire food chains
10-20 km
Smoothing scale of L4 products
The break disappears into gradient

Our Design Philosophy

Pelagic Insight is built on a simple principle: honest data is more useful than pretty data. We show you what the satellite measured, where the temperature breaks are, and where we do not have data.

🛰️

Raw Satellite Data

We display actual satellite measurements with cloud gaps visible. If the satellite did not measure a pixel, we do not fill it in. The absence of data is itself information: it tells you where to apply less confidence and where to rely on other sources.

📐

Temperature Break Detection

We apply gradient analysis to raw satellite passes to detect and highlight sharp temperature transitions. These are computed from actual observations, not interpolated fields. The result shows you where the satellite measured a real, abrupt change in water temperature.

🗺️

Isotherm Lines

Contour lines drawn at meaningful temperature intervals show the spatial structure of the ocean surface. Combined with raw data, isotherms let you see where temperature zones begin and end without relying on color perception alone. They are the cartographic tool that makes SST operationally useful.

Why We Do Not Default to “HD” Interpolated Imagery

It creates false confidence. A gap-free, smooth SST map looks authoritative. It suggests the satellite saw everything. It did not. The map is partially measured and partially invented.

It destroys actionable features. The smoothing inherent in interpolation blurs temperature breaks, the single most useful feature for offshore fishing, into gentle gradients that look normal.

Resolution claims are misleading. An L4 product gridded at 1 km does not contain 1 km information. The effective resolution of the analysis is 10-25 km. The small pixel size is a display choice, not a measurement.

Gaps are signal, not noise. Knowing that a region has been cloud-covered for three days tells you the SST in that area is uncertain. Filling it with a statistical estimate hides that uncertainty.

The Pelagic Insight Approach

Show raw satellite data with cloud gapsHonest
Detect temp breaks from actual measurementsActionable
Draw isotherm contours from observed dataPrecise
Use interpolation as supplementary context onlyTransparent
Label all data sources and processing levelsTraceable

We do not hide interpolation. It has value as background context. But we never present it as the primary analysis, and we always distinguish estimated values from measured observations.

Honest Data Catches Fish

Every other SST product on the market shows you a beautiful, smooth, HD image of the ocean. It looks professional. It looks precise. And it has systematically erased the features you went offshore to find.

Pelagic Insight takes a different approach. We show you what the satellite actually measured. We find the temperature breaks in that real data. We draw isotherm lines so you can see structure at a glance. And where we do not have data, we tell you, because knowing where uncertainty lives is part of making good decisions on the water.

Raw
Satellite observations
Not statistically generated
Breaks
Detected from real data
Not smoothed away
Honest
Gaps shown, not hidden
Uncertainty is information
← Back to Research