Cabbeling

When two volumes of water at the same temperature are mixed, the result is intuitive. But what happens when two parcels of seawater with the exact same density are mixed? The ocean often defies simple intuition: the mixture can become denser and sink. This surprising phenomenon, known as cabbeling, is not a minor curiosity but a fundamental engine that helps drive the planet's vast ocean circulation system, influencing global climate. This article delves into the physics behind this counter-intuitive process, addressing the knowledge gap of how simple mixing can lead to densification. Across the following sections, you will discover the science behind this effect and its profound consequences.

The first section, "Principles and Mechanisms," will unpack the core concept of cabbeling by exploring the nonlinear equation of state for seawater, illustrating how curved lines on a Temperature-Salinity diagram lead to this density increase. We will then transition in "Applications and Interdisciplinary Connections" to explore how this microscopic effect scales up to influence large-scale ocean dynamics, driving the formation of deep water and presenting significant challenges and elegant solutions in the world of computational climate modeling.

Imagine you have two glasses of water, both at exactly the same temperature. If you mix them, what’s the temperature of the mixture? The answer is, of course, the same temperature. Now, imagine you have two parcels of seawater, sitting at the same depth, that have exactly the same density. If you mix them, what’s the density of the mixture? The intuitive answer would be "the same density," but the ocean, in its beautiful complexity, often defies our simple intuitions. In many crucial situations, the mixture is actually denser than the two parcels you started with. This surprising phenomenon, a kind of chemical alchemy performed by physics, is called ​​cabbeling​​. It is not a minor curiosity; it is a fundamental engine that helps drive the ocean's vast global circulation.

To understand this apparent magic, we must look at what determines the density of seawater. Density is not a single, simple property but a function of three variables: its temperature ($T$), its salinity ($S$), and the pressure ($p$) it's under. This relationship is known as the ​​equation of state (EOS)​​ for seawater, written as $\rho(T, S, p)$.

Let's simplify for a moment and forget about pressure, focusing on what happens at a constant depth. We can map out the properties of seawater on a Temperature-Salinity (T-S) diagram. On this map, we can draw lines connecting all the points that have the same density. These lines are called ​​isopycnals​​ (from the Greek iso- for "equal" and -pyknos for "dense").

If the equation of state were a simple, linear relationship—if density changed by a fixed amount for every degree of temperature change and every gram of salt added—then these isopycnals would be straight, parallel lines. Mixing two water parcels, say parcel A $(T_A, S_A)$ and parcel B $(T_B, S_B)$, would produce a mixture with the average properties, $(T_A+T_B)/2$ and $(S_A+S_B)/2$. On the T-S diagram, this mixture would lie at the midpoint of the straight line connecting A and B. If A and B were on the same straight isopycnal, their mixture would land on that exact same line. No density change.

But nature is more subtle. The true equation of state for seawater is ​​nonlinear​​. This means the isopycnals are not straight; they are curved. Specifically, on a T-S diagram, they are slightly concave, sagging "downward" toward the region of higher density.

Now, picture our two parcels, A and B, lying on the same curved isopycnal. When we mix them, the resulting parcel still lands at the midpoint of the straight line connecting A and B. But because the isopycnal sags like a hanging rope, this midpoint no longer lies on the original line. It falls below it, into the region of higher density. The mixture is denser! This is the entire secret of cabbeling: it is a direct consequence of the curved, nonlinear nature of the seawater equation of state. Mathematically, this curvature is captured by the second-order terms in a Taylor expansion of the density function. While the linear terms cancel out when you average the two initial parcels, the second-order terms—representing the curvature—do not. They leave behind a small but significant increase in density.

This might sound abstract, so let's consider a concrete example. Imagine we are in a part of the ocean where a parcel of warm, salty water meets a parcel of cold, fresher water.

• ​​Parcel A​​: Warm and salty, with a temperature of $T_{\mathcal{A}} = 22^\circ\mathrm{C}$ and salinity of $S_{\mathcal{A}} = 36.8 \text{ g/kg}$.
• ​​Parcel B​​: Cold and fresh, with a temperature of $T_{\mathcal{B}} = 2^\circ\mathrm{C}$ and salinity of $S_{\mathcal{B}} = 34.2 \text{ g/kg}$.

Let's say that due to the balance of their properties, they both happen to have the exact same density at a given pressure. Now, we mix them together in equal parts. The resulting mixture will have the average temperature and salinity: $T_{\mathrm{mix}} = 12^\circ\mathrm{C}$ and $S_{\mathrm{mix}} = 35.5 \text{ g/kg}$. If we now use a realistic, nonlinear equation of state for seawater to calculate the densities, we find a remarkable result. The density of the mixture is about $0.026 \text{ kg/m}^3$ greater than the density of the original parcels.

This number may seem tiny—a mere fraction of a percent of the total density. But in the vast expanse of the ocean, where colossal masses of water are delicately balanced, such a small difference is the equivalent of a giant weight being placed on one side of a scale. The water parcel has become heavier, and there is only one way for it to go: down.

This sinking is the critical link between the microscopic process of mixing and the planet-spanning motion of the oceans. In the ocean, mixing is not a gentle stirring in a beaker; it is a chaotic, turbulent process driven by swirling eddies. Much of this mixing happens "laterally" along surfaces of constant density (or, more accurately, along ​​neutral surfaces​​). This is called ​​isoneutral mixing​​.

But as we've just seen, when this sideways mixing occurs between waters of different temperatures and salinities, cabbeling generates denser water. This is a form of ​​diapycnal transformation​​—a process that moves water across isopycnal surfaces. The result of this "sideways" mixing is a net "downward" velocity.

This mechanism is particularly potent in the high-latitude oceans, like the North Atlantic and the Southern Ocean around Antarctica. Here, cold, fresh polar waters meet warmer, saltier waters from the mid-latitudes. Their mixing, driven by cabbeling, helps create the dense deep waters that sink to the abyss and begin a slow journey along the seafloor. This process is a crucial contributor to the ​​thermohaline circulation​​, often called the Global Ocean Conveyor Belt, which transports heat, salt, carbon, and nutrients around the planet, fundamentally shaping Earth's climate. Cabbeling is one of the quiet, persistent engines that drives this global system.

So far, we've mostly ignored pressure. But the ocean is deep, and pressure changes dramatically with depth. This introduces another, related nonlinear effect called ​​thermobaricity​​. The name itself gives a clue: it's about the coupling of temperature (thermo) and pressure (baric).

Thermobaricity arises because the properties of water, such as its coefficient of thermal expansion (how much it expands when heated), change with pressure. In the cold waters of the polar regions, surface water near freezing is surprisingly insensitive to temperature changes. But as a parcel of this water sinks, the immense pressure alters its molecular structure, making it more sensitive to temperature. The thermal expansion coefficient, $\alpha_T$, increases with pressure.

This creates a powerful positive feedback loop for convection. Imagine a parcel of water at the surface that is slightly colder, and thus denser, than its surroundings. It begins to sink. As it descends, the pressure on it increases. Due to thermobaricity, its thermal expansion coefficient increases. This amplifies the effect of its coldness, making it even denser relative to the surrounding water at that new depth. This makes it sink faster, which increases the pressure further, and so on. A small initial push can cascade into an unstoppable plunge into the deep ocean.

Cabbeling and thermobaricity are two sides of the same coin: the nonlinearity of the seawater equation of state $\rho(T,S,p)$. Cabbeling is about the curvature in the T-S plane, while thermobaricity is about the "twist" or cross-dependence between temperature and pressure (the $\rho_{Tp}$ term in a Taylor expansion). Together, they make the concept of a "surface of constant density" wonderfully slippery. If you were to trace a path in the ocean that should, in theory, keep you on a single neutral surface, these effects could cause you to drift onto a different density surface. In formal terms, the field of neutral surfaces is ​​non-integrable​​—a beautiful mathematical quirk of nature that allows the ocean to transform water masses in ways a simpler, linear world never could.

It's one thing to understand these principles, but another to build them into the complex computer simulations—the Ocean General Circulation Models—that we use to predict weather and climate. One might wonder: how can a model handle this? In particular, many ocean models use the ​​Boussinesq approximation​​, which assumes that the flow is non-divergent ($\nabla \cdot \mathbf{u}=0$). This means volume is conserved. But if cabbeling makes water denser, shouldn't its volume shrink?

This apparent paradox is resolved with stunning elegance. The Boussinesq approximation cleverly decouples the conservation of volume from the effect of density on buoyancy. The model enforces $\nabla \cdot \mathbf{u}=0$ as a kinematic constraint on the flow field, filtering out sound waves. Meanwhile, it separately tracks temperature and salinity. After mixing occurs in the model, it uses the full nonlinear equation of state to calculate the new density. If cabbeling has occurred, this new, higher density creates a larger downward buoyancy force in the momentum equations, causing the water to sink. The densification doesn't violate volume conservation; it drives a volume-conserving flow.

Therefore, a model that includes a realistic, nonlinear equation of state and a scheme for mixing tracers like temperature and salinity will automatically produce cabbeling and its consequences. It is not an extra ingredient one has to add, but an emergent property that arises from representing the physics faithfully. If, on the other hand, one were to use a simplified, linear equation of state, the buoyancy tendency from isoneutral mixing would be identically zero. The engine of cabbeling would be turned off. It is the nonlinearity—the subtle curves and twists in the fundamental laws of seawater—that gives the ocean some of its most profound and important behaviors.

Now that we have unwrapped the basic mechanism of cabbeling, we can begin a truly fascinating journey. We will see how this subtle quirk of water, this refusal to behave in a perfectly linear fashion, is not merely a curiosity for the laboratory but a potent force that shapes our planet's climate. Its influence cascades from the physical sinking of water masses in the deep ocean to the most abstract challenges in computational climate modeling. We are about to discover that in the nonlinear world of the ocean, the whole is truly different from the sum of its parts.

Imagine two different masses of water meeting in the open ocean. This happens constantly at oceanic fronts, the vast, slow-motion boundaries where, for instance, warmer, saltier water from the subtropics encounters colder, fresher water from the polar regions. Let’s say these two water masses, through a conspiracy of their differing temperatures and salinities, happen to have the exact same density. If the ocean were a simple, linear world, mixing them would produce a new water mass with that same density. Nothing much would happen.

But the ocean is not a linear world. As we have seen, the equation of state for seawater is curved. When these two equal-density parcels mix, the resulting water is invariably denser than either of its parents. This phenomenon is cabbeling. What is the consequence? This newly formed, denser water is now heavier than its surroundings, and so it begins to sink.

This is not just a theoretical density increase; it translates into a real, physical motion—a downward velocity that can be calculated from first principles. Mixing that occurs horizontally across a front can induce a vertical flow. Cabbeling thus acts as a quiet but persistent engine, helping to drive the vertical circulation of the ocean. It contributes to the formation of new, dense water types that ventilate the deep sea, supplying it with oxygen and forming the deep, cold limbs of the global thermohaline circulation—the great ocean conveyor belt that regulates our planet’s climate.

This same principle applies with even greater force to dense water overflows, which are like massive, cold, salty rivers cascading down the undersea continental slopes. As these plumes descend, they turbulently mix with the surrounding ambient water. Cabbeling provides an additional source of density to the mixture, making the plume heavier than it would otherwise be and strengthening its downward plunge. When oceanographers build models of these critical processes, ignoring this nonlinearity means underestimating the driving force of the cascade. The difference in buoyancy between a simple linear model and one that correctly includes the quadratic terms responsible for cabbeling can be significant, highlighting a real, physical force born from the water's subtle nonlinearity.

The quiet effectiveness of cabbeling in the real world becomes a cacophonous headache for scientists trying to build digital replicas of the Earth’s climate. In modern ocean models, which are necessarily coarse and cannot resolve every small swirl and eddy, one of the biggest challenges is to represent the effect of unresolved mixing.

Oceanographers know that eddies in the ocean preferentially mix water along surfaces of constant density, not across them. Mixing across density surfaces, or "diapycnal" mixing, requires energy and is relatively weak in the ocean interior. Therefore, a primary goal in parameterizing eddy mixing is to ensure it is "isoneutral"—that is, it only mixes properties along paths of neutral buoyancy. This seems simple enough: just tell the model to mix along density surfaces. But here, the nonlinearities of the equation of state create a profound and beautiful problem.

First, as we’ve seen, cabbeling ensures that even if you mix two parcels perfectly along a neutral surface, the resulting mixture is denser and will sink off that surface. So, the very act of isoneutral mixing is, paradoxically, a source of diapycnal motion.

Second, a far more difficult demon enters the stage: ​​thermobaricity​​. This term describes the fact that water's properties, particularly its thermal expansion coefficient ($\alpha$), change with pressure. Intuitively, you can think of it this way: warm water is more compressible—more "squishy"—than cold water. This means that as a water parcel moves up or down, the rate at which its density changes with temperature is altered. The consequence is that the very definition of a "neutral" direction is a function of depth.

These two effects, cabbeling and thermobaricity, conspire to create a "neutral surface mirage." There is no single, globally consistent set of density surfaces that an eddy can follow. The local direction of neutral buoyancy changes from point to point, and the path you must follow depends on where you are in $(T, S, p)$ space. Mathematically, we say the field of neutral vectors has a non-zero curl, which means it is non-integrable. You cannot "connect the dots" to form a globally coherent surface that is everywhere neutral.

This is a disaster for a simple-minded ocean model. If a model uses a simplified density surface to guide its mixing—for example, potential density referenced to a fixed pressure level like the surface ($\sigma_0$) or 2000 decibars ($\sigma_2$)—it is guaranteed to get the direction wrong almost everywhere. The model will think it is mixing along a neutral surface, but it will actually be mixing at a slight angle to the true local neutral direction. This creates a spurious diapycnal flux. Because the parameterized "along-surface" mixing is so vigorous compared to the true diapycnal mixing, even a tiny misalignment angle can create an artificial vertical mixing that is orders of magnitude larger than the real thing, completely corrupting the model's heat and salt budgets and its simulation of climate.

The solution? Immense sophistication. Modern ocean models must abandon the idea of simple density surfaces. Instead, they must compute the neutral direction locally, at every point and every time step, using the full nonlinear equation of state. Parameterizations like the Gent-McWilliams (GM) and Redi schemes are designed to do just this, navigating the neutral surface mirage by constantly recalculating their bearings. These nonlinearities also create notorious problems in calculating the pressure gradient force in models that use terrain-following coordinates, a source of error that has plagued modelers for decades.

The story does not end with modeling the flow of water. The tendrils of these nonlinearities reach into the even more abstract world of climate sensitivity and data assimilation. Scientists often want to ask questions like: "If the sea surface temperature was warmer in the North Atlantic last year, how much did that contribute to the melting of a specific glacier today?" Answering such questions requires "adjoint models," which can be thought of as a mathematical time machine for tracing the sensitivity of an effect back to its causes.

When we build an adjoint model of the ocean, the nonlinearities of the equation of state once again reveal their character in a new light. Here, an interesting distinction appears. The thermobaric effect, which enters the dynamics through the first derivative of density with respect to pressure ($\rho_p$), creates a non-local coupling in the adjoint equations. The sensitivity of a quantity at one depth becomes explicitly linked to properties throughout the entire water column via vertically integrated kernels.

Cabbeling, on the other hand, does not appear at all in this first-order sensitivity analysis. Because it arises from the second derivatives of density ($\rho_{TT}$, etc.), it is a fundamentally second-order effect. It tells you about the curvature of the system's response, but not about the initial slope. To capture the influence of cabbeling on climate sensitivities, one must venture into the even more complex territory of second-order adjoints or Hessian analysis. This elegant mathematical separation reveals a deep truth about the structure of the ocean's dynamics: thermobaricity affects the system's linear response to small nudges, while cabbeling governs the next level of complexity.

From a source of sinking water to a phantom that haunts our most powerful supercomputers, cabbeling is a testament to the power of nonlinearity. A simple-looking curve in the relationship between temperature, salt, and the density of water gives rise to a cascade of complex, beautiful, and challenging phenomena. It reminds us that in nature, the rules are often subtle, and understanding our world requires not just an appreciation for its grand, sweeping laws, but also for its elegant, intricate, and wonderfully nonlinear details.