The Abyss Is Warming and It’s not Adiabatic Compression

The deepest ocean appears to be warming at an accelerating rate.

Before looking for a new source of heat, I wanted to test a simpler possibility:

Could the water only look warmer because it moved deeper and was compressed by increasing pressure?

That process is called adiabatic compression. It is real, measurable, and already built into ocean thermodynamics. After tracing the Johnson abyssal-warming calculation back to its source variable, the answer became clear:

Adiabatic compression cannot create the warming acceleration Johnson mapped.

This is the first mechanism I can cross off.

What is adiabatic compression?

Pressure increases as water moves deeper into the ocean.

When a water parcel sinks, that increasing pressure raises its in-situ temperature slightly. “In-situ temperature” means the temperature a thermometer would measure at the parcel’s current depth.

The process is reversible. If the same parcel rises again without exchanging heat, the pressure drops and the temperature decreases.

This pressure effect does not require new heat to enter the water.

It’s similar to a bicycle pump warming when air is compressed. The temperature rises because of compression. The air has not necessarily received heat from an outside source.

That gave me a specific quantitative test:

Does the observed abyssal warming disappear after the reversible pressure-temperature effect is removed?

The data I used

The analysis used the observational products from the Johnson[1] abyssal-warming workflow, including:

• shipboard CTD profiles from the World Ocean Database;

• Deep Argo profiles;

• the combined profile cache containing Conservative Temperature and Absolute Salinity;

• Johnson’s fixed-pressure acceleration field;

• Johnson’s pressure-layer volume estimates;

• direct-observation and spatially interpolated products;

• bathymetry products and ETOPO seafloor data for geographic checks.

The validated profile cache contains 1,124,918 profiles on a pressure grid extending to 6,000 dbar. The main mechanism test focused on 4,000 to 6,000 dbar, with additional checks over 2,000-4,000, 4,000-5,500, and 2,000-6,000 dbar.

The first important step was identifying exactly what temperature variable Johnson had fitted.

Step 1: Find the temperature variable behind the map

Oceanographers use several different definitions of temperature, and pressure directly affects some temperature variables and is intentionally removed from others.

I traced the Johnson workflow from its profile cache through the trend calculation. The source variable was:

gct = Conservative Temperature

The mapped acceleration variable was:

dct_accel = Conservative Temperature acceleration

This was unambiguous in the workflow code and output metadata. The profiles were interpolated without extrapolation to Johnson’s 10-dbar pressure grid, and the acceleration was calculated independently at each fixed pressure level.

That variable identity is the decisive part of the test.

Step 2: Understand what Conservative Temperature measures

Conservative Temperature, usually written as CT, tracks the thermal state of seawater while removing the reversible temperature change caused by pressure.

In simple terms:

• in-situ temperature includes pressure warming;

• Conservative Temperature removes that reversible pressure effect.

Conservative Temperature is directly related to potential enthalpy:

h⁰ = c_p⁰ CT

where the TEOS-10 reference heat capacity used in the calculation was:

c_p⁰ = 3991.867957 J kg⁻¹ K⁻¹

A cluster can sink, become compressed, and register a higher in-situ temperature while its Conservative Temperature and potential enthalpy remain unchanged.

Johnson’s map therefore does not show the simple thermometer warming caused by increased pressure. It shows acceleration in the pressure-corrected thermal variable.

Step 3: Reproduce the Johnson time calculation

Johnson’s temperature history was represented with a quadratic model.

For each geographic bin and pressure level, time was centered on 2005:

τ = year − 2005

Conservative Temperature was then fitted as:

CT(τ) = β₀ + β₁τ + β₂τ²

The acceleration was:

a_CT = 2β₂

The workflow used individual profiles, ordinary least-squares fitting, a minimum 10-year time span, and more than two occupied time intervals. The outlier procedure applied six-interquartile-range limits to the provisional trend and acceleration fields.

Later reproduction work matched the released Johnson acceleration field on the exact final support, confirming that the mapped field really was fixed-pressure Conservative Temperature acceleration rather than an in-situ temperature artifact.

Step 4: Reconstruct the pressure only temperature effect

I also calculated what pressure movement would do to an ordinary in-situ thermometer reading.

The TEOS-10 function reconstructs in-situ temperature from Absolute Salinity, Conservative Temperature, and pressure:

t = t(SA, CT, p)

For the pressure-only calculation, I held salinity and Conservative Temperature fixed and allowed only pressure to change:

t_pressure(t) = t( SA_ref, CT_ref, p(t) )

The predicted pressure-only temperature change was:

Δt_pressure = t_pressure(t₂) − t_pressure(t₁)

I then reconstructed the full in-situ temperature change while allowing pressure, salinity, and Conservative Temperature to vary:

Δt_total = t( SA(t₂), CT(t₂), p(t₂) ) − t( SA(t₁), CT(t₁), p(t₁) )

The remaining change was:

Δt_residual = Δt_total − Δt_pressure

Finally, the exploratory pressure contribution was calculated as:

f_adiabatic = Δt_pressure Δt_total

This decomposition used the TEOS-10 seawater equations rather than a constant temperature-per-meter approximation.

What the exploratory calculation found

The first-pass decomposition produced a median raw pressure fraction of approximately:

f_adiabatic ≈ 0.087

That corresponds to about 8.7% across the limited set of supported density surfaces in that exploratory calculation. The pressure contribution varied substantially among regions and individual density surfaces.

I do not use that 8.7% value as the final basis for crossing off the mechanism.

A later methodological audit found that the original regional calculation needed stronger basin definitions, exact fixed-pressure fitting, improved sampling blocks, and stricter density-surface interpolation. I withdrew the provisional regional classification rather than presenting it as a completed result.

Fortunately, the central conclusion does not depend on that regional fraction.

The decisive result

Johnson’s observed warming field is already an acceleration in Conservative Temperature.

Reversible pressure compression can change in-situ temperature:

Δt_pressure ≠ 0

For a cluster undergoing only reversible adiabatic compression, its Conservative Temperature remains unchanged:

ΔCT_adiabatic = 0

Therefore:

a_{CT, adiabatic} = 0

Johnson measured:

a_{CT, observed} ≠ 0

Pressure alone cannot generate the observed Conservative Temperature acceleration.

That is why adiabatic compression fails this mechanism test.

What this conclusion means

The result is specific and strong:

The Johnson abyssal warming acceleration is not a false warming signal created by water being pushed deeper and squeezed.

Adiabatic compression can still alter the temperature measured by an in-situ thermometer. It can accompany sinking water. It can contribute to the difference between in-situ temperature and pressure-corrected temperature.

It cannot create an acceleration in Conservative Temperature.

What this result does not tell us

Crossing off adiabatic compression does not identify the ultimate source of the warming.

Several possibilities remain:

• vertical movement of density surfaces;

• redistribution of existing ocean heat;

• changes in Antarctic Bottom Water and overturning circulation;

• geothermal or hydrothermal input;

• electromagnetic induction and ohmic heating.

The next question is whether moving density layers can shift enough existing heat through fixed pressure levels to reproduce the observed pattern.

For now, one mechanism is crossed off:

The abyssal warming acceleration is present after reversible pressure warming has already been removed.

[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