If the Heat Is Coming From Above, Why Is the Abyss Most Affected?
One of the explanations I keep hearing for deep-ocean warming is that the atmosphere is warming the ocean, and that heat eventually works its way downward.
I wanted to see what the vertical structure actually looks like to see if it supports this idea.
It doesn’t.
I extended the exact same Johnson [1] analysis through the full water column to compare the abyssal ocean at 4,000-6,000 dbar (4 - 6 kdbar) with the four layers directly above it at:
0-1 kdbar
1-2 kdbar
2-3 kdbar
3-4 kdbar
4-6 kdbar
The abyssal signal was more extensive.
Same Conservative Temperature field. Same quadratic time model. Same fitting rules. Same definitions of warming and acceleration.
And, the result is pretty important.
The 4-6 kdbar abyssal layer is the maximum of all five layers for all three measures of how widespread the signal is.
For the volume where the signal can actually be estimated:
Warming
0-1 kdbar: 62.7%
1-2 kdbar: 68.3%
2-3 kdbar: 60.1%
3-4 kdbar: 59.9%
4-6 kdbar: 72.6%
Positive acceleration
0-1 kdbar: 51.4%
1-2 kdbar: 54.7%
2-3 kdbar: 53.2%
3-4 kdbar: 53.3%
4-6 kdbar: 58.1%
Warming AND accelerating
0-1 kdbar: 38.0%
1-2 kdbar: 41.8%
2-3 kdbar: 34.8%
3-4 kdbar: 33.6%
4-6 kdbar: 43.0%
So, the layer farthest from the atmosphere is the layer where the observed warming signal is the most geographically AND volumetrically extensive.
That needs explaining.
And, before blaming this on different observational coverage at different depths, I checked that too.
I restricted the comparison to the same 1,392 horizontal columns with valid observations in all five layers. That strict subset still represents about 96.7% of the observed positive 4-6 kdbar acceleration signal.
And, the result persists.
The abyss still has the largest:
warming fraction;
positive-acceleration fraction;
warming + accelerating fraction.
So this isn't just an artifact of comparing different sampled parts of the ocean.
Important caveat
The acceleration is not actually strongest in the abyss.
The mean acceleration is larger higher in the water column, especially around 1-2 kdbar. And, the positive acceleration signal per unit observed volume decreases strongly with depth. So, I’m not finding that the thermal signal simply grows stronger and stronger as you move toward the seafloor.
The actual structure is more interesting than that.
The magnitude generally weakens with depth, while the spatial extent of warming and acceleration becomes greatest in the abyss.
The abyss is not the location of the largest average acceleration.
It’s the place where the warming-acceleration pattern has spread through the largest fraction of the observed water.
I made a note in my notebook to return to this…. seems important.
Not explained by a “top-down” mechanism
A common explanation for deep-ocean warming is ..the atmosphere warms the surface ocean, and the heat eventually makes its way downward.
At a broad level, sure. Heat can enter through the surface and ocean circulation can move it around. But, that statement alone doesn't explain the vertical pattern.
If the proposed mechanism were a simple local top-down pathway, with heat progressively diffusing, mixing, or moving through the water directly above the abyss, I’d expect the signal to become progressively less widespread as it gets farther from the source.
Instead, that isn't what the observations show. The fraction of the observed ocean participating in the warming signal increases with depth, and reaches its maximum at 4-6 kdbar.
And, the vertical structure isn't even smoothly monotonic! From one layer to another the acceleration strengthens, weakens, weakens again, then strengthens into the abyss.
So, this doesn’t look like a simple thermal front gradually fading as it moves downward through a stack of water.
Matched columns make this even more interesting
The deeper signal isn't disconnected from the water above. About 81.4% of the positive abyssal acceleration signal on matched support occurred beneath a positively accelerating 3-4 kdbar column. So, the signal is vertically coherent.
But…the abyssal layer is still more extensively warming and accelerating.
Vertically connected, but abyssally more extensive.
That’s much harder to explain with a simple local downward transfer mechanism.
Does this prove the heat isn't atmospheric?
Not yet.
There’s another pathway we need to test…
Antarctic Bottom Water forms near Antarctica, sinks into very dense water classes, and spreads laterally through the abyss. So, surface-origin heat does not necessarily have to pass vertically through the 0-1, 1-2, 2-3, and 3-4 kdbar water directly above every downstream abyssal location.
A possible pathway is: surface ocean near Antarctica → changing bottom water properties → sinking → lateral abyssal transport.
That could potentially create an abyssal signal with a different spatial structure than the water sitting immediately above it. Which means the next question is:
Can the actual AABW and overturning pathway produce the vertical pattern we observe?
Can it explain why the normalized magnitude weakens with depth while the warming and acceleration become most spatially extensive in the abyss?
Can we follow those changes on density surfaces?
Do the CT and salinity changes track Antarctic Bottom Water?
Do they align with the expected water-mass pathways?
And, is the transport actually sufficient?
Those are testable questions. For now, this part is clear:
The observed warming-acceleration signal does not fade away with distance from the surface.
At 4-6 kdbar, the abyss has the largest observed warming fraction at 72.6%.
It has the largest positive-acceleration fraction at 58.1%.
And, it has the largest fraction simultaneously warming and accelerating at 43.0%.
Of all five layers I tested.
So pointing at atmospheric warming and saying “that's where the abyssal heat came from” isn't enough.
The transport mechanism has to explain the structure of the signal too.
And the structure is telling us something interesting...
References
[1] Johnson, G. C. (2026). Observed multi-decadal acceleration of globally averaged abyssal ocean warming. Geophysical Research Letters, 53(14), e2026GL124104. https://doi.org/10.1029/2026GL124104
