How sea temperature is measured, and how to read it
What satellites, buoys and pier sensors each measure, why their numbers differ, and how to read a sea temperature so it tells you what the water will feel like.
A sea temperature looks like a simple fact: 18.4°C65°F at Bournemouth today. Behind it sits a chain of instruments, each measuring a slightly different thing at a different place and depth. Knowing what each one sees makes the number far more useful, and explains why a swimmer can step into water that feels nothing like the figure on the screen. This guide walks through the three kinds of measurement that appear on this site, in order of how far they are from your toes.
Satellites: a picture of the whole ocean every day
Most of the numbers on this site, including the map and every daily value, come from satellites. Two kinds of instrument are involved. Infrared radiometers on weather satellites measure the heat glowing from the sea surface in fine detail, about one to four kilometres per pixel, but cannot see through cloud. Microwave radiometers see through cloud but at a coarser resolution, around 25 to 50 km (16 to 31 mi). Neither on its own gives a complete daily picture, so agencies combine both, add measurements from ships and buoys to correct the satellites, and produce what is called an analysis: a best estimate of the temperature in every cell of a global grid for one day, with the gaps filled from neighbouring cells and the previous day.
The analysis used here, NOAA Coral Reef Watch’s CoralTemp, has cells of about five kilometres and uses night-time observations only. The night-time choice matters. On a calm sunny afternoon the top few millimetres of the sea, the skin that the satellite sees, can be a degree or two warmer than the water a swimmer is in; at night that skin layer mixes away and the satellite reading matches the top metre or so of water far better. A night-time value is therefore a conservative, stable figure for the day, but it will not show the afternoon warm-up in a shallow bay.
Satellites measure the open sea. A five-kilometre cell is a large area, and for a coastal location the cell used is the nearest one that is fully over water, which is usually a few kilometres offshore. The location page states that distance. Where the coast is complicated, with estuaries, lagoons or long shallow flats, the nearest open-water cell may not represent the water at the beach at all; those locations carry a “sheltered water” note and are left out of the rankings.
Buoys: a thermometer in the water
Moored weather buoys measure the water directly, usually at a depth of about one metre below the surface, every hour or less. In the United States the National Data Buoy Center runs an offshore network, and in Canada the Department of Fisheries and Oceans does the same. A buoy reading is a real measurement at a real time, so it catches the afternoon warming, the cold snap after a storm and the arrival of upwelled water that a daily satellite analysis smooths over. The limitation is coverage: buoys are hundreds of kilometres apart and mostly offshore, and a reading from a buoy 30 km19 mi out says little about a beach inside a bay.
Pier and harbour sensors: closest to the swimmer
Tide gauges and harbour sensors, such as those run by NOAA’s Center for Operational Oceanographic Products and Services, sit on piers and quays in exactly the kind of place people swim. They record water temperature every six minutes, a metre or two below the surface, in the shallow water the public uses. Where such a sensor exists within about 15 km9.3 mi of a location, this site shows its latest reading as a separate line, with the station name, its distance and the time. It is nearly always the better guide to what the water will feel like. Its weakness is the opposite of the satellite’s: a harbour sensor measures one spot that may be affected by a river, a power station outfall or a sheltered basin, and says nothing about the open beach a few kilometres away.
Why the beach disagrees with all of them
Even with a sensor on the pier, the water at the shoreline can be different, for reasons any regular swimmer will recognise:
- Shallow water over sand warms and cools fast. On a sunny afternoon at low tide the water on a wide beach can be two or three degrees above the open sea; after a clear cold night it can be below it.
- Tides move water in and out. A rising tide brings offshore water onto the beach; a falling tide drains sun-warmed water off the flats.
- Wind drives upwelling and mixing. A few days of offshore wind can drop the nearshore temperature by several degrees on an exposed coast, especially on the Pacific side of North America.
- Rivers and rain. Fresh water from a river mouth floats on the sea and can be much warmer in summer and colder in winter than the salt water beneath.
- Depth of the swimmer. The top metre is what all these instruments report. Diving down a few metres in summer often means dropping through a sharp temperature step into much colder water.
None of these is an error in the measurement. They are the reason this site labels every number with its kind and never averages a satellite value and a sensor value into one figure.
Reading a location page
Each page follows the same order. The large figure is the latest satellite analysis for the nearest open-water cell, with its date; if no newer analysis has been published it says so. Beside it, a chip compares today with the long-term average for the month: “slightly warmer than typical” means half a degree to a degree and a half above, “noticeably” means up to three degrees, “much” means more. A second chip gives the swim-feel band, explained in the swimming comfort guide.
Where a station exists, the measured line appears under the satellite value and is marked as a measurement. If the two disagree by a degree or two, that is normal and the measurement is usually the one to trust for swimming.
The monthly table shows the average for each month over the years since 2003, with the “usual range” being the coldest and warmest that month has averaged in any single year. A month whose average rests on fewer years than usual, because sea ice interrupted the satellite record, is marked; where there are too few years to average, the table says so rather than guessing.
Reading the map
The map shows the same daily analysis as a colour field, with the colour scale running from the near-freezing water of polar seas through blue, green and yellow to the hottest tropical water in red and magenta. Smooth bands of colour are the open ocean; sharp boundaries are fronts where currents meet, such as the edge of the Gulf Stream off the eastern United States. Tap anywhere on the sea and the value for that cell appears, with the nearest location in our inventory. Grey near the poles is sea ice, where no sea temperature is defined.
The map is a daily picture, so it shows weather as well as climate: a cool patch off California after a windy week, or a warm plume off a river mouth after rain. For the long-term pattern, the ocean and sea pages and the monthly tables are the better source.
Precision and honesty
A satellite analysis is accurate to within about half a degree of a well-placed buoy on average over open water, and less certain close to complicated coasts. We show values to one decimal place because that is how the data arrives, not because the tenth of a degree is meaningful at your beach. Read 18.4°C65°F as “about 18, in the cool band, warmer than usual for the time of year”, and let a thermometer in the water settle the rest. The methodology page gives the exact rules we follow, and the data sources page credits each provider.