We Know the Seafloor Is Releasing Heat. Now I’m Testing Where It Goes.
I’ve reached a part of the ocean heat source project that I’ve been looking forward to for a while. We’ve calculated how much heat is coming through the seafloor, now I’m following that heat through the ocean.
So basically, I’m dropping modeled geothermal heat onto the ocean floor and letting the ocean circulation move it around month by month.
Where does it go? Does it stay near the bottom? Does it get carried up? Does it move between ocean basins? And, most importantly, does enough of it end up in the abyssal ocean, where the warming acceleration is actually being observed?
That’s what this test is for.
The three heat sources I’m testing
I’m testing three different versions of heat coming from below the ocean.
The first is conductive heat. This is the background heat slowly leaking out of the solid Earth and through the seafloor.
Globally, that source is about:
≈ 16.8 trillion watts
The second is lithospheric heat, using a model called GDH1. This map estimates how much heat is coming through different parts of the ocean floor. Younger seafloor is generally hotter, while older seafloor has had more time to cool. So, this map lets the heat vary depending on the age of the crust.
Its total heat is larger than the basic conductive estimate:
≈ 28.9 trillion watts
The third source is what I’m calling the hydrothermal remainder.
This is the extra heat in the GDH1 map after removing the basic conductive part. It gives me a way to test the part of the heat budget that could be associated with hot water moving through the ocean crust instead of heat simply leaking through it.
≈ 12.2 trillion watts
So, GDH1 is the larger total map. The conductive map is one part of it, and the hydrothermal remainder is the difference between them.
Why knowing the total heat isn't enough
This is the problem I kept running into. You can tell me that the Earth is putting tens of trillions of watts into the bottom of the ocean, but that still doesn't tell me whether that heat can explain the warming pattern we’re trying to understand.
The ocean is moving…heat entering at one place doesn't necessarily stay there. Water carries it sideways, upward, downward and between basins. So, I need to actually trace it.
At each month, the model starts with the heat already in the ocean, adds that month's geothermal heat, then lets the observed ocean circulation redistribute it.
=
Heat already there + Heat added from below + Heat moved in − Heat moved out
And ,the amount of new geothermal heat added during each month is just:
=
Source power × time
Nothing mysterious here. The hard part is following that heat through a real 3D ocean.
I’m running two versions of the ocean
This is where we get rigorous, and precise. For each heat source, I want to run the heat through the ocean twice. One run uses the actual monthly ocean circulation. The other uses a repeating average seasonal circulation.
So, the average version still has winter, spring, summer and fall. It just repeats the same typical seasonal pattern every year. The actual version contains the real changes that happened from year to year.
Then I can subtract them.
=
ACTUAL − CLIMATOLOGY
Both runs get the same geothermal heat, so if there’s a difference between them, that difference came from the circulation. This subtraction tells me whether changes in ocean circulation can make a steady geothermal heat source deliver more heat to the abyssal ocean over time. Then I can compare that pattern with the observed warming to see if they actually match.
We already did this with the conductive source
The first two runs used the basic conductive heat map, and that result surprised me a little.
The real circulation did make a meaningful difference in where the heat ended up.
About 12 percent of the total modeled conductive heat inventory was rearranged compared with the repeating seasonal ocean.
≈ 1.27 × 1021 J
≈ 12.0% of the modeled heat inventory
So, circulation really does change where the heat goes, although for the conductive heat source, it changed things in the wrong direction.
When I looked at the month by month difference between the real circulation and the repeating seasonal circulation, the real circulation delivered less conductive heat to the target in almost 90 percent of the months.
≈ 89.6%
I also measured whether that difference was changing over time.
The trend was negative:
df/dt ≈ −3.56 × 10−4 per year
And, I checked the uncertainty using a Newey West calculation, which accounts for the fact that one month of ocean circulation isn't completely independent from the month before it.
The uncertainty range stayed entirely negative:
−4.35 × 10−4 to −2.77 × 10−4 per year
Over the same 20 year window, that works out to a modeled change of:
Δf ≈ −0.0071
But from my earlier geothermal calculation, the conductive source would need to increase by about:
Δf ≈ +0.496
So, not only was the circulation effect way too small, it had the opposite sign. Compared with what I need to explain the warming, the modeled circulation effect was:
≈ −1.43%
The minus sign is really the important part here. The real circulation wasn't slowly making a steady conductive heat source look stronger. It was moving the modeled heat in the opposite direction.
And the deepest part of the ocean showed the same thing. In the 4 to 6 kbar layer, the real circulation left about:
≈ −5.52 × 1020 J
About 79% of the modeled cells in that deep layer had less conductive heat under the real circulation than under the repeating seasonal circulation. The major basins mostly moved in the same direction, with the Atlantic showing the largest circulation effect.
I also compared the geography with the actual abyssal warming pattern. The circulation difference had almost no useful spatial match with the observed warming, and using the real circulation actually made the overall match slightly worse than the repeating seasonal circulation. So, for the conductive source, this closes a pretty specific possibility:
A steady conductive heat source can't be made to look like the increasing abyssal heat source I need just by using the real changing ocean circulation.
But that still leaves the other two heat maps open. They put their heat into the ocean in different places, and that can change what the circulation does with it. That's why I'm now repeating the same experiment with the lithospheric heat source.
Tracing lithospheric heat
I’m currently running the lithospheric heat source through the actual month by month ocean circulation. This source puts more heat into younger parts of the seafloor and less into older parts. For this run, the total heat mapped onto the ocean model is:
≈ 2.889 × 1013 W
≈ 28.89 trillion watts
The experiment begins in April 2004 with zero tracer heat in the ocean. Then, the lithospheric heat starts entering through the seafloor and I follow it month by month through March 2024.
That gives me 240 months of modeled transport.
Animation on Twitter
This is probably the part I'm most excited about visually. For the conductive heat source, I only saved the final answer. Now, I'm saving a smaller 3D copy of the tracer every month.
Each monthly frame contains about:
×
240 monthly heat states
I also saved a true zero heat starting frame. So, that animation you might have seen on Twitter, that is you watching the geothermal heat enter through the seafloor and spread through the ocean over 20 years.
AND, the animation retains depth information meaning I can separate the shallow ocean from the deep ocean instead of just flattening everything onto a map.
The deepest layer I'm especially interested in is the abyssal target:
4 to 6 kbar
roughly the deepest several kilometers of the ocean
So, what you’re watching is whether the heat stays near the bottom, spreads sideways, gets pulled upward, or moves into different basins.
Lithospheric heat result only half done (run 3 of 6)
Run 3 of 6 is only the ACTUAL half of the experiment with lithospheric heat. It’s the full picture of how the lithospheric heat moves through the real month by month ocean circulation. But, we still don’t know how much of that movement is caused by the real changes in circulation from year to year. For that, I’ll need the second run.
I'll take the exact same lithospheric heat source and run it through the repeating seasonal ocean. Then, I can subtract the two and see what changed because of the real circulation. That comparison will tell me whether the real circulation is moving more geothermal heat into the abyssal ocean, pulling it away, or not changing much at all.
And, I’ve already decided how I’m measuring that before seeing the answer.
How I’m measuring
For the geothermal explanation to get stronger, I need more than a tracer animation which took me 3 days to make. The real circulation needs to move the heat in the right direction. I’d want to see more of it reaching the deep ocean over time, and enough of it to actually matter.
Then there’s the geography. The places gaining heat should also line up reasonably well with where the abyssal ocean is actually warming. If those things start lining up, then the geothermal idea gets a lot more interesting.
If the heat moves the wrong way, the effect is way too small, or it ends up in the wrong places, then that closes another piece of the explanation.
That’s really why I’m tracing the heat instead of just comparing total watts. A heat source can be huge, but if the ocean carries that heat somewhere else, it doesn’t help explain the warming I’m looking at.
