№ 171 · earth science

The Gulf Stream is a heat pipe, squeezed west by the spinning Earth

The Gulf Stream is a fast, narrow river of warm water running north along the American coast. The wind and the spinning Earth push it into the western edge of the Atlantic. Like a pipe, it carries heat only if the water comes back colder than it went. About 1.2 petawatts of heat flows north across the subtropical Atlantic, most of it because water that went north warm comes back deep and cold.

A river in the sea

Between Florida and the Bahamas the stream is called the Florida Current, and it has been measured nearly continuously since 1982. Seawater conducts electricity, so as it flows through Earth's magnetic field it makes a voltage, which a cable on the sea floor picks up. Denis Volkov and colleagues at NOAA, the University of Miami and the UK's National Oceanography Centre report that the current carries about 32 Sv. One sverdrup (Sv) is a million cubic metres of water a second. William Johns and colleagues give its flow-weighted temperature as 19.6 °C on average.

Downstream it grows. Robert Stewart's open textbook Introduction to Physical Oceanography notes that measurements show the flow rising from about 30 Sv in the Florida Strait to about 150 Sv near 40° N. Most of the extra is not new water. A tight loop of recirculating flow just south of the stream feeds water back into it. Averaged over the tracks of 110 drifting buoys, the mean current runs at up to about 0.6 m/s near 37° N, 71° W. The same tracks are a tangle: in the North Atlantic the eddies carry 8 to 37 times the energy of motion of the mean flow.

Why the fast side is the west

Across the North Atlantic, westerly winds blow in the north and trade winds blow from the east in the south. Between them the wind turns the subtropical ocean clockwise into a great round current called a gyre. You might expect it to circle evenly, like water stirred in a bowl. It does not. Over most of the ocean the water drifts slowly south, and it all comes back north in one narrow, fast current hugging the western coast.

Stewart walks through how Henry Stommel showed why, in 1948. Stommel solved for a wind-driven gyre in a rectangular ocean with simple friction on the bottom. On a planet that does not spin, the gyre came out symmetric. With the same spin at every latitude, it was still symmetric. Only when he let the effect of Earth's spin grow from the equator toward the pole did the streamlines crowd together against the western side. Stewart concludes that this change of spin with latitude is what a western boundary current needs in order to exist.

Stewart also gives the reason in terms of spin. Every patch of water carries spin of its own plus a share of Earth's spin, and that share is zero at the equator and largest at the pole. The wind keeps adding clockwise spin to the gyre. In the broad interior, water drifting south loses some of Earth's share. On its own it would make that up by spinning counterclockwise, but the wind's clockwise push cancels it, so the books balance. Water going back north gains Earth's share, so on its own it would have to spin more clockwise, and the wind is pushing it clockwise as well. Something has to supply counterclockwise spin, and the open ocean cannot. A narrow, fast current along the coast can, through the shear between the fast water and the slower water at its edge. In Walter Munk's related model of 1950, which uses sideways friction instead, one plausible friction value gave a western current about 225 km wide, with a shape like the real Gulf Stream and the Kuroshio off Japan.

Interactive Top: Stommel's wind-driven ocean, seen from above, with north at the top. The wind on the left blows east in the north and west in the south. The lines are streamlines, and the water runs along them. Set how much Earth's spin changes from south to north, compared with friction, and watch where the return flow goes. The strip under the map shows the north–south speed across the middle of the basin. Bottom: a heat-pipe calculator. Set the flow, the temperature going north and the temperature coming back. Every control works with Tab, Enter and the arrow keys.

Where the return flow goes

spin change ÷ friction40
wind (trade winds) (westerlies) N S west coast east coast north south north–south speed along the dashed line
Spin change 40.

The heat pipe

flow north and back14 Sv
going north at18 °C
coming back at2 °C
measured at 26.5° N: 1.2 PW −1 0 1 2 3 PW heat carried north, petawatts
14 Sv, 18 °C north, 2 °C back.

What a heat pipe carries

Water does not carry "heat" by temperature alone. If a stream of water goes north at 18 °C and the same amount comes back at 18 °C, nothing has been delivered. Heat is delivered only when the water comes back colder than it went. What reaches the north each second is the mass of water flowing per second, times the temperature drop, times the heat needed to warm a kilogram of seawater by a degree (about 4,000 joules, in Stewart's figure).

Stewart does the sum with figures from oceanographer Bill Schmitz. The Gulf Stream carries 40 Sv of 18 °C water north, and 14 Sv of it returns south deep down at about 2 °C. To make that trip the water has to give up about 0.9 petawatts (0.9 PW, or 900 million million watts) north of 24° N. Stewart notes how close that crude sum is to the 1.2 ± 0.2 PW that Stephen Rintoul and Carl Wunsch estimated much more carefully in 1991. Had the water come back at the surface, it would have come back much warmer and carried far less heat.

Today that heat transport is measured directly. Since April 2004 the RAPID array has spanned the Atlantic at 26.5° N, with moorings across the basin and the Florida cable on the western edge. Johns and colleagues report, from April 2004 to December 2020, an overturning of 16.9 ± 1.2 Sv and a northward heat transport of 1.20 ± 0.12 PW. That is about 60% of the heat that all the oceans carry poleward at that latitude, and about 30% of what ocean and atmosphere carry together. In the simplest way of splitting it, about 90% comes from the overturning: warm water north near the surface, cold water south at depth. The rest comes from the horizontal circulation of the gyre. Johns and colleagues note that this split depends on how it is done, because water that circles the gyre does not all come back at the depth it left. The pipe also leaks as it goes. The northward heat transport falls to about 0.5 PW at 60° N and about 0.3 PW where the warm water enters the Nordic Seas.

Predict first

In January, London's latitude is 15–20 °C warmer on the European side of the Atlantic than on the American side. Climate scientists ran climate models two ways: as normal, and with the ocean forbidden to carry heat sideways. What happened to that 15–20 °C difference?

Check your answer

It stayed almost the same. Richard Seager and David Battisti found the winter difference across the North Atlantic unchanged, south of northern Norway, whether or not their model's ocean could move heat. Taking away the ocean's heat transport cooled the far north more, but it did not remove the contrast.

Does it keep Europe warm?

The popular story says yes, and Stewart's textbook says the Gulf Stream's heat "helps warm Europe". Richard Seager of Columbia's Lamont-Doherty Earth Observatory tested how much. In the model runs without ocean heat transport, the subpolar North Atlantic was up to 8 °C colder in places, but land mostly cooled by less than 3 °C. Seager writes that the heat the North Atlantic currents release warms both sides of the ocean by roughly the same few degrees. That leaves the much larger 15–20 °C contrast to other causes.

He names two. First, the ocean stores summer heat and lets it out in winter, and in the middle latitudes the winds blow from west to east. Western Europe gets air that has crossed the ocean, while the American east coast gets air that has crossed a cold continent. Second, the Rocky Mountains bend the westerly winds into a vast wave that brings Arctic air south over eastern North America and mild air north over western Europe. Seager also notes that at middle latitudes the atmosphere carries several times more heat poleward than the ocean does. The heat the ocean brings north still counts. Seager places it mostly in two spots: off the American east coast in winter, where cold dry air from the continent pulls a few hundred watts per square metre out of the stream, and in the Nordic Seas.

Is it slowing down?

Climate models suggest that warming could weaken the Atlantic overturning, so the record is watched closely. The Florida cable had shown a small decline. In 2024 Volkov and colleagues found that the cable's voltage had to be corrected for the slow change in Earth's own magnetic field, the same field that turns a compass. With that correction, the Florida Current's trend from 1982 to 2023 is −0.1 ± 0.2 Sv per decade, which is no significant change in four decades. The correction also shrinks the measured weakening of the whole overturning at 26.5° N from 2004 to 2022 by about 40%, which leaves it only marginally significant. In 2025 Jens Terhaar, Linus Vogt and Nicholas Foukal rebuilt the overturning from the heat that the North Atlantic gives off to the air. They found that its ten-year average at 26.5° N did not weaken from 1963 to 2017. Some reconstructions built from other evidence do find a decline, and Volkov's team writes that the direct records may still be too short to tell a long slowdown from swings that last years to decades. Newer direct measurements of the whole overturning lean toward a decline. In 2026 Qianjiang Xing, Shane Elipot and colleagues reported that the deep southward flow along the western edge of the North Atlantic weakened over the past two decades at four mooring arrays from 16.5° N to 42.5° N, partly offset by a strengthening on the eastern side. In October 2026 the UK's National Oceanography Centre said the RAPID record at 26.5° N still shows a weakening trend of −0.7 ± 0.6 Sv per decade, and that it needs watching to at least 2034 to tell a natural swing from a change caused by warming. So the Florida Current itself has held steady for four decades, the whole overturning appears to have weakened modestly over the last twenty years, and whether that is a lasting slowdown is still argued.

In short

The wind turns the subtropical Atlantic into a clockwise gyre. Because Earth's spin grows from equator to pole, the gyre's return flow is forced into one narrow, fast current against the western coast: the Gulf Stream. It acts as a heat pipe because water that went north warm comes back cold, much of it deep down. The result is about 1.2 PW of heat crossing 26.5° N. That heat warms the far North Atlantic, but in Seager's experiments most of the winter contrast between Europe and eastern North America comes from the winds and the ocean's seasonal store of heat. The Florida Current has held steady through four decades of measurement, while the direct records of the whole overturning point to a modest weakening over the last twenty years whose cause is still argued.

Where this comes from

  1. Introduction to Physical Oceanography, ch. 5, 11, 12 and 13 linked only, not reproduced
    Robert H. Stewart (Texas A&M University) · 2006
    earthweb.ess.washington.edu/booker/ESS514/stewart/stewart_ocean_book.pdf
  2. The westward intensification of wind-driven ocean currents (Transactions, American Geophysical Union) linked only, not reproduced
    Henry Stommel · 1948
    doi.org/10.1029/TR029i002p00202
  3. Towards two decades of Atlantic Ocean mass and heat transports at 26.5° N (Phil. Trans. R. Soc. A) linked only, not reproduced
    William E. Johns, Shane Elipot, David A. Smeed, Ben Moat, Brian King, Denis L. Volkov and Ryan H. Smith · 2023
    pmc.ncbi.nlm.nih.gov/articles/PMC10590663/
  4. Florida Current transport observations reveal four decades of steady state (Nature Communications) linked only, not reproduced
    Denis L. Volkov, Ryan H. Smith, Rigoberto F. Garcia, David A. Smeed, Ben I. Moat, William E. Johns and Molly O. Baringer · 2024
    www.nature.com/articles/s41467-024-51879-5
  5. Atlantic overturning inferred from air-sea heat fluxes indicates no decline since the 1960s (Nature Communications) linked only, not reproduced
    Jens Terhaar, Linus Vogt and Nicholas P. Foukal · 2025
    pmc.ncbi.nlm.nih.gov/articles/PMC11736101/
  6. Meridionally consistent decline in the observed western boundary contribution to the Atlantic Meridional Overturning Circulation (Science Advances 12, eadz7738) linked only, not reproduced
    Qianjiang Xing, Shane Elipot, William E. Johns, David A. Smeed, Ben I. Moat and John W. Loder · 2026
    doi.org/10.1126/sciadv.adz7738
  7. Latest AMOC data continues to show a weakening trend linked only, not reproduced
    National Oceanography Centre (quoting Ben Moat) · 2026
    www.noc.ac.uk/news/latest-amoc-data-continues-show-weakening-trend
  8. The Source of Europe's Mild Climate (American Scientist 94) linked only, not reproduced
    Richard Seager (Lamont-Doherty Earth Observatory) · 2006
    ocp.ldeo.columbia.edu/res/div/ocp/gs/pubs/Seager_AmSci_2006.pdf