Except near thermal vents, the deep ocean is generally considered to be quite cool. For instance, water temps in the Mariana Trench (~10 km deep) are 1-4 °C.

However, the geothermal gradient runs counter to that, where the lithosphere at the same 10km depth would be around 250-300 °C.

So, in a sense, water should be heated from beneath and only be able to release its energy into the atmosphere/space. Yet, we see the exact opposite of this, where the upper layers are the warmest.

Where did the thermal energy go? Is it the product of billions of years of ocean currents carrying the heat to the poles, or am I overlooking some other principle?


2 Answers 2


The simple answer to this question is that cold seawater is denser than warm seawater, so it sinks and fills up the abyssal ocean.

The water that fills up the abyssal ocean comes from the polar regions. Here's a 1990s plot of temperature in the Atlantic, retrieved from the WOCE Atlantic Ocean Atlas:

enter image description here

You can see that the coldest water is coming from the Antarctic region (dark blue stuff on the left hand side of the plot) and spreading north.

If you look at 1990s salinity instead (plotted below) you can see that the water from the Arctic isn't as cold, but is much saltier (yellow part of the plot spreading from the right hand side and sitting above the water coming north from Antarctica).

enter image description here

When thinking about the density of the ocean, it's important to remember that the ocean isn't filled with water, it is filled with seawater. Seawater contains approximately 35 g of salt per kilogram of seawater. The density of seawater therefore depends on both the temperature and the salinity. You could write this as

$$\rho = \alpha T + \beta S$$

which is correct enough for what we need here. (Small aside, the density also depends on pressure and the coefficients $\alpha$ and $\beta$ depend on pressure, salinity, and temperature - you can see this in the curvature of the lines of constant density in this figure. Truly understanding the density of seawater is a massive undertaking.)

The salt makes two other very important things happen:

  1. as seawater gets colder, it keeps getting denser (but $\alpha$ gets pretty small at low temperatures, so salinity really dominates the density)
  2. seawater freezes at ~-2°C

This is why we see water at -0.4°C spreading away from Antarctica in the bottom left corner of that figure. It was colder when it was near the surface, but as it sank it entrained ambient fluid, and warmed up.

This explains how we get cold water down into the abyss, but why doesn't the geothermal heat flux simply warm it up?

The reason that the geothermal heat flux doesn't heat up the abyssal ocean is that it is too small. The geothermal heat flux is up to ~0.5 W/m$^{2}$. By comparison, heat fluxes at the surface of the ocean range from -200 W/m$^{2}$ to 200 W/m$^{2}$ (where positive means into the ocean in both cases). The surface heat fluxes and ocean dynamics simply overwhelm the much much smaller geothermal heat flux. Lots of models include a heat flux through the ocean floor, but because it is so small the abyssal ocean remains cold. Here's an example of the heat budget from an ocean model that includes a flux of heat through the seafloor.

So, not much hard maths in there, but I'm not sure we need it.

  • $\begingroup$ It appears that you can post the WOCE images with acknowledgement. sam.ucsd.edu/whp_atlas/atlantic_index.html $\endgroup$ Aug 22, 2021 at 20:01
  • 1
    $\begingroup$ Your first plot shows the potential temperature, not the temperature. While those concepts are related, they give very different information. Is there a plot with the actual temperature as well? $\endgroup$ Aug 26, 2021 at 11:51
  • $\begingroup$ @AtmosphericPrisonEscape True, but for water the difference between potential temperature and actual temperature is small, as water is almost incompressible. This is very different from the atmosphere. So the actual temperature plot will look very similar. $\endgroup$ Sep 23, 2021 at 20:37
  • $\begingroup$ Not sure how this fits in here, but a rule-of-thumb principle in climatology is that the circulation in the atmosphere is driven by heating from the bottom, but the (deep) circulation in the ocean is driven by cooling from above. It is a simplified beginners' principle, but it does sum up what you explained in more detail. $\endgroup$ Sep 23, 2021 at 20:49

Cold water is denser than warm water. The very cold water in the Arctic and Antarctic regions sinks and circulates deep undersea.

  • 2
    $\begingroup$ Thanks for answering. You're on to something there, although it's not aways true that cold water is denser. It's denser only up to a certain point, and then it begins to expand. This leads to water column stability. The missing link is the salinity. When water freezes, it leaves most of the salt behind, increasing the density locally under the new ice. This denser water sinks through water of the same temperature but of reduced salinity. britannica.com/science/seawater/… has a great explanation, and discusses Antarctic Bottom Water in particular. $\endgroup$ Aug 1, 2021 at 16:51
  • 1
    $\begingroup$ I feel like this answer is not complete because it doesn't discuss the salinity, but I also feel you deserve the nod because it was through your answer that I quickly found the full explanation. If you want to update this one, I'll marked it as answered. Otherwise, if it's easier for you I can write up my findings and create a new answer here. $\endgroup$ Aug 1, 2021 at 22:43
  • 1
    $\begingroup$ Why isn't the ocean bottom heated by the crust from below? $\endgroup$
    – RonJohn
    Aug 2, 2021 at 1:28
  • $\begingroup$ And there's certainly none of the requested "hard science and math". $\endgroup$
    – RonJohn
    Aug 2, 2021 at 1:29
  • 2
    $\begingroup$ @RonJohn The ocean is heated by the crust of course, but that's mostly heat conduction, which happens at a far slower rate than the heat convection of cold water currents flowing in from the polar regions; hence water↔water convection dominating the equilibrium. $\endgroup$
    – Will
    Aug 2, 2021 at 2:47

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.