Diapycnal Mixing

The vast expanse of the ocean is not a uniform body of water but a complex, layered system where density differences create stable strata. While movement along these layers is easy, mixing across them—a process known as diapycnal mixing—is a slow, energetically demanding feat. This vertical stirring, however, is fundamentally important, acting as a crucial regulator for the Earth's climate system and the global distribution of heat and nutrients. This article addresses how such a subtle process can exert control over the planetary-scale ocean conveyor belt. We will explore the physics governing this deep-ocean mixing and its far-reaching consequences. The first chapter, "Principles and Mechanisms," will unpack the fundamental physics of stratified fluids, the energy sources that power mixing, and its essential role in sustaining global ocean circulation. Subsequently, "Applications and Interdisciplinary Connections" will reveal how diapycnal mixing connects to climate change, the carbon cycle, marine life, and the challenges of simulating our planet's future.

Imagine the ocean not as a uniform tub of water, but as a giant, intricate layer cake. For the most part, the ocean is ​​stably stratified​​: lighter, warmer, and less salty water sits on top of denser, colder, and saltier water. These layers of constant density, known as ​​isopycnals​​, create a kind of grain or fabric running through the ocean. Now, picture yourself as a tiny, neutrally buoyant submarine. Moving horizontally along one of these layers is relatively easy; you're just swapping places with water of the same density. This process of stirring along isopycnals is called ​​isopycnal mixing​​. But what if you try to move vertically, across the layers? You would immediately feel a powerful force pushing you back to your original layer. This is the restoring force of buoyancy, the same force that makes a beach ball pop back to the surface when you try to push it underwater.

To move a parcel of water across these density layers requires doing work against this buoyancy force. This is the essence of ​​diapycnal mixing​​: mixing across the layers of constant density. The "stiffness" of this stratification is quantified by a value called the ​​buoyancy frequency​​, denoted $N$. Where the density changes rapidly with depth, the stratification is stiff, $N$ is large, and it takes a great deal of energy to mix vertically. The work per unit mass required to lift a fluid parcel a vertical distance $\xi$ against gravity in a stratified fluid is a beautiful and simple result: it scales as $\frac{1}{2}N^2\xi^2$.

This energy cost has a profound consequence: turbulent eddies in the ocean find it far easier to stir things along isopycnals than across them. The result is a dramatic anisotropy in mixing. The diffusivity along isopycnals, which we can call $K_{\parallel}$, is enormously larger than the diffusivity across them, $K_{\perp}$. In fact, observations show that $K_{\parallel}$ can be ten million times greater than $K_{\perp}$! The ocean, therefore, prefers to mix along its density surfaces, a fundamental principle that has far-reaching consequences for how heat, salt, carbon, and nutrients are transported. At oceanic fronts, where sharp horizontal temperature and salinity gradients create steeply tilted isopycnals, this distinction becomes especially vivid. Isopycnal mixing acts to smooth out gradients along these tilted surfaces, while diapycnal mixing, often intensified by strong shear and turbulence in these dynamic regions, is the only process that can irreversibly alter the stratification itself.

This brings us to a critical question. If diapycnal mixing is so energetically "expensive," requiring work to lift the ocean's center of mass, what pays the bill? The ocean isn't a perpetual motion machine; this energy must come from somewhere. The answer lies in two colossal sources of mechanical power: the ​​wind​​ blowing across the ocean surface and the gravitational pull of the ​​Moon and Sun​​ that drives the tides.

First, consider the tides. As the great tidal waves slosh back and forth across the ocean basins, they run into the rugged topography of the seafloor—underwater mountain ranges, ridges, and canyons. This interaction is not smooth; it's a violent, churning process that generates waves not on the surface, but within the stratified interior of the ocean. These "internal waves" can travel thousands of kilometers before they become unstable and break, much like waves on a beach, releasing their energy into small-scale turbulence.

Second, the wind does work on the ocean surface, driving currents and also generating a different spectrum of internal waves near the surface. These waves, too, propagate into the abyss, carrying energy far from where they were created. When they break, they contribute to the turbulent cocktail that stirs the deep sea.

This grand connection between celestial mechanics, atmospheric winds, and deep-ocean turbulence can be captured in a surprisingly elegant global energy budget. The total mechanical power put into the ocean by winds ($P_w$) and tides ($P_t$) must, in a steady state, be balanced by the rate at which that energy is dissipated. This dissipation happens in two ways: some of the turbulent energy is simply turned into heat through viscous friction, but a crucial fraction is used to do the work of mixing the stratified fluid. The ratio of the energy used for mixing (the increase in background potential energy, $B$) to the total energy dissipated by turbulence is called the ​​mixing efficiency​​, $\Gamma$. A wonderfully compact relationship emerges: the globally integrated rate of potential energy increase due to mixing is given by the total power input, modulated by this efficiency. $\int_{\mathcal{V}} \rho_0 \kappa N^2 \, \mathrm{d}V = \Gamma (P_w + P_t)$ Here, $\kappa$ is the local diapycnal diffusivity. This single equation tells an incredible story: the amount of small-scale mixing happening throughout the entire world's oceans is constrained by the power supplied by winds and tides. It is a testament to the profound unity of physics, linking the vast scales of the cosmos to the microscopic dance of molecules in the deep. This energy perspective, often framed within the theory of ​​Available Potential Energy (APE)​​, clarifies that diapycnal mixing acts to increase the ocean's background potential energy (the energy of its most stable state), and the energy to do so is ultimately drawn from the APE generated by surface heating and cooling.

Now we arrive at the heart of the matter: why is this seemingly subtle process of diapycnal mixing so critical for the Earth's climate? The answer lies in its role as the gatekeeper of the ​​Thermohaline Circulation​​, the great global "conveyor belt" of ocean currents.

This circulation begins in the polar regions, where cold, salty water becomes very dense, sinks to the abyss, and begins a slow, thousand-year journey along the ocean floor towards the equator. For the circulation to be a closed loop, this deep, dense water must eventually return to the surface. But how? To rise, it must become lighter. It cannot simply decide to defy gravity.

The only way for deep water to become light enough to upwell is through diapycnal mixing. Over vast areas of the ocean interior, slow but persistent turbulent mixing gradually blends the cold, dense abyssal waters with the warmer, lighter waters above. This process, happening molecule by molecule, lightens the deep water, allowing it to slowly rise back towards the surface, completing the circuit.

This relationship can be captured by a simple but profound scaling law derived from a balance between upward advection and downward diffusion. In a thermocline (the region of strong vertical temperature gradient) of thickness $h$, the average upwelling velocity $w$ is directly proportional to the diapycnal diffusivity $K_v$: $w \sim \frac{K_v}{h}$ This result is stunning in its simplicity. It tells us that if there were no diapycnal mixing ($K_v = 0$), there would be no large-scale upwelling ($w = 0$), and the global overturning circulation would grind to a halt. Without diapycnal mixing, the deep ocean would fill up with cold, stagnant water, and the climate system as we know it would be completely different.

Given its importance, how do scientists actually measure diapycnal mixing? We can't just release dye in the deep ocean and wait a thousand years. The key is to measure the turbulence itself. Oceanographers deploy remarkable instruments called microstructure profilers, which free-fall through the water column, measuring tiny, centimeter-scale fluctuations in velocity and temperature. From these measurements, they can calculate the rate at which turbulent kinetic energy is dissipated into heat, a quantity known as ​​epsilon​​ ($\epsilon$).

But how does dissipation relate to mixing? The connection is made through the ​​Osborn-Cox model​​, a cornerstone of physical oceanography. The logic is as follows: the downward diffusive flux of buoyancy is given by $-K_\rho N^2$, where $K_\rho$ is the diapycnal diffusivity. The rate at which this flux does work against gravity increases the fluid's potential energy at a rate of $B = K_\rho N^2$. We already defined the mixing efficiency, $\Gamma$, as the ratio of this potential energy gain to the turbulent dissipation rate, $\Gamma = B/\epsilon$. By simply rearranging these two equations, we can solve for the diffusivity: $K_\rho = \frac{\Gamma \epsilon}{N^2}$ This is a powerful tool. By measuring the local dissipation rate $\epsilon$ and the stratification $N^2$, and by making a reasonable assumption about the mixing efficiency (experiments suggest $\Gamma$ is about $0.2$), we can estimate the local diapycnal diffusivity. This is how we obtain the crucial values of $K_\rho$ needed to understand and model the ocean's role in the climate system.

Representing this complex process in global climate models is one of the greatest challenges in computational oceanography. Models divide the ocean into grid cells that are kilometers wide, far too coarse to resolve the centimeter-scale eddies that cause mixing. Therefore, mixing must be parameterized—represented by a simplified rule.

A naive approach, such as applying a simple diffusion operator in the model's regular grid coordinates (e.g., longitude, latitude, and depth), leads to a disaster. In the real ocean, isopycnal surfaces are often tilted. A diffusion term that acts purely horizontally on a flat depth level in a model will inevitably transport water across these tilted density surfaces. This creates a massive amount of artificial, non-physical mixing known as ​​spurious diapycnal mixing​​. This numerical artifact can be so large that it completely swamps the real, physical mixing, rendering the model's simulation of ocean circulation useless. Other numerical errors, such as inaccuracies in calculating the pressure gradient force in models that use coordinates that follow the seafloor topography, can also generate spurious currents that drive artificial mixing.

To overcome this, ocean modelers have developed sophisticated parameterizations. One of the most famous is the ​​Gent-McWilliams (GM) scheme​​. Instead of diffusing tracers horizontally, the GM scheme cleverly rotates the diffusion tensor so that it acts primarily along the model's isopycnal surfaces. This largely eliminates the spurious diapycnal flux. Furthermore, GM introduces a term that represents the tendency of eddies to flatten isopycnal slopes, which is parameterized as an "eddy-induced velocity." This elegantly mimics the true behavior of ocean eddies, allowing models to maintain realistic water masses and circulation patterns.

These models are built upon a framework known as the ​​Boussinesq approximation​​. This involves a clever bit of physical reasoning: for the purposes of calculating fluid motion and volume conservation, density variations are assumed to be negligible, allowing the velocity field to be treated as divergence-free ($\nabla \cdot \boldsymbol{u} = 0$). Yet, for calculating the all-important buoyancy forces that drive the flow, those same small density variations are retained. This apparent contradiction is resolved because the impact of density changes on mass balance is much smaller than their impact on gravity. This approximation allows diapycnal mixing to change the density of fluid parcels ($D\rho/Dt \neq 0$) without violating the model's core assumption of volume conservation, providing a self-consistent, though approximate, digital representation of our planet's oceans.

Having journeyed through the intricate principles of diapycnal mixing, we might be tempted to leave it as a curious feature of stratified fluids. But to do so would be to miss the forest for the trees. This seemingly subtle, slow vertical stirring is, in fact, one of the master gears in the grand clockwork of our planet. Its influence extends from the millennial-scale rhythm of global ocean circulation to the very air we breathe and the future climate we will inhabit. Let us now explore how this fundamental process connects to a breathtaking range of phenomena, revealing the profound unity of Earth science.

Imagine the deep, dark abyss of the ocean, miles below the sunlit surface. It is a world of immense pressure and profound cold. How long does it take for a water parcel, having sunk from the polar seas, to make its return journey to the surface? The answer is staggering, and it is dictated almost entirely by diapycnal mixing. The timescale for a signal to diffuse across a vertical distance $H$ with a diffusivity $K_d$ scales as $T \sim H^2/K_d$. For the deep ocean, with $H$ around $4000$ meters and a typical interior diapycnal diffusivity of $K_d \approx 10^{-5} \, \mathrm{m^2/s}$, this simple formula yields a timescale of thousands to tens of thousands of years. The abyss, therefore, does not operate on human timescales; it slumbers on a geological one.

This immense sluggishness is not just a curiosity; it is the linchpin of the global Meridional Overturning Circulation (MOC), the "great ocean conveyor belt." Cold, dense water sinks at high latitudes and spreads throughout the deep ocean basins. But for this circulation to be a continuous loop, that deep water must eventually rise. Buoyancy, however, makes this a profoundly difficult task—it's like trying to lift a heavy weight. Diapycnal mixing, by slowly blending the dense, deep water with the lighter water above, provides the crucial energy source that lightens the abyssal water, allowing it to ascend and complete its journey. Without this tiny, persistent vertical mixing, the global conveyor belt would grind to a halt.

This role as the ocean's pacemaker has direct consequences for climate change. The ocean has absorbed over 90% of the excess heat from anthropogenic global warming, acting as a tremendous buffer for the climate system. But how quickly it takes up this heat, and thus how quickly the surface warms, is not a simple matter. Simple climate models might treat the ocean as a uniform "slab" or bathtub of water that warms passively. Reality is far more interesting. In a full ocean model, the rate of heat uptake depends on complex pathways: rapid advection along outcropping density surfaces ("ventilation") and the slow, inexorable march of heat downwards via diapycnal mixing. The strength of the ocean's vertical stratification, which resists diapycnal mixing, therefore becomes a key regulator of surface warming. A more stratified ocean mixes less efficiently, taking up less heat into its interior and causing the surface to warm faster. Understanding diapycnal mixing is therefore not an academic exercise; it is essential for predicting the trajectory of our planet's temperature.

The slow timescale set by diapycnal mixing has profound implications not just for heat, but for the chemistry of the ocean and the life within it. The "ventilation age" of a water parcel—the time since it was last in contact with the atmosphere—is a direct consequence of these slow transport pathways. As water journeys through the deep ocean for hundreds or thousands of years, it acts as a moving chemical laboratory.

At the sunlit surface, photosynthetic organisms create organic matter, forming the base of the marine food web. When these organisms die, they sink, carrying their carbon with them in a process known as the "biological pump." In the deep, dark ocean, bacteria decompose this falling organic rain, remineralizing it and releasing Dissolved Inorganic Carbon (DIC) back into the water. Because the water is so old, this process has a long time to act. The slow ventilation governed by diapycnal mixing means that DIC accumulates at depth, making the deep ocean a vast reservoir of carbon—far larger than the atmosphere. The older the water, the more remineralized carbon it holds. Diapycnal mixing, by controlling the ventilation age, therefore directly modulates the ocean's capacity to store carbon and regulate atmospheric $\text{CO}_2$.

But what gives life must also take it away. The same process of remineralization that enriches the deep sea in carbon also consumes dissolved oxygen. The slow vertical mixing that isolates deep water from the carbon-poor surface also isolates it from the oxygen-rich atmosphere. This delicate balance creates regions known as Oxygen Minimum Zones (OMZs). As climate change warms the surface ocean, stratification increases, which in turn weakens diapycnal mixing. This reduction in vertical mixing means a smaller supply of oxygen is stirred down from the surface, causing OMZs to expand and intensify. This "deoxygenation" has severe consequences for marine ecosystems, creating vast zones that are inhospitable to fish and other organisms. The same physical knob—diapycnal mixing—that tunes the global carbon cycle also controls the ocean's breath.

How can we possibly capture a process that occurs on scales of centimeters within global climate models whose grid cells are many kilometers wide? This is one of the greatest challenges in computational science, and the story of how we handle diapycnal mixing is a beautiful illustration of the interplay between physics and numerical modeling.

We cannot resolve the tiny turbulent eddies directly, so we must parameterize them. This involves finding a relationship between the small-scale process and the large-scale variables the model does resolve. A cornerstone of this effort is the Osborn formula, which allows us to estimate the diapycnal diffusivity $K_d$ from the rate of turbulent energy dissipation, $\epsilon$, and the stratification, $N$, as $K_d = \Gamma \epsilon / N^2$, where $\Gamma$ is a mixing efficiency factor. This elegant relationship allows us to take measurements from specialized microstructure probes and translate them into a crucial parameter for our global climate models.

But a deeper challenge emerges. Most ocean models historically used a simple, intuitive grid of horizontal levels, like floors in a building (a "$z$-coordinate" system). A problem arises because density surfaces in the ocean are not flat; they are gently sloped. When water flows along these sloped density surfaces, a model with a "staircase" grid forces the flow to artificially step up or down between grid levels. This numerical process creates a catastrophic amount of artificial, or "spurious," diapycnal mixing. This numerical artifact can be hundreds of times larger than the true physical mixing, completely corrupting the model's simulation of ocean properties and circulation. The model becomes a leaky sieve when the ocean is, in reality, a nearly watertight vessel.

The solution is as elegant as the problem is vexing: if the flow follows density surfaces, then let the model's coordinate system do the same. This led to the development of "isopycnal-coordinate" and "hybrid-coordinate" models, where the grid layers themselves bend and deform to follow the ocean's density structure. By aligning the computational grid with the natural pathways of the flow, spurious mixing is dramatically reduced, and the model can faithfully represent the near-adiabatic nature of the ocean interior. This is a powerful lesson: to build a true-to-life simulation of our world, the very architecture of our code must be guided by a deep understanding of the underlying physics.

Finally, we come to a connection that reveals the truly fundamental nature of diapycnal mixing. In a rotating, stratified fluid like the ocean, there exists a quantity of profound importance: Ertel's Potential Vorticity (PV). You can think of PV as the fluid's dynamical "genetic code." In an idealized, frictionless, and adiabatic (no heating or mixing) ocean, PV is conserved following a fluid parcel. It constrains where water can and cannot go, shaping the paths of ocean currents and eddies.

One might think that mixing, which is a thermodynamic process of stirring tracers like heat and salt, would be separate from the dynamics of motion governed by PV. But this is not so. The very act of diapycnal mixing—the stirring of buoyancy across density surfaces—acts as a source or a sink of Potential Vorticity. Even in an otherwise frictionless fluid, simply mixing a tracer can generate or destroy PV. This means that diapycnal mixing doesn't just redistribute properties within the flow; it can fundamentally alter the dynamical character of the flow itself. It is a stunning example of the deep and often surprising unity of physics, where the laws of thermodynamics and the laws of motion are inextricably intertwined.

From setting the timescale of the ice ages to dictating the future of marine life and challenging the ingenuity of our brightest computational scientists, diapycnal mixing is a thread woven through the very fabric of the Earth system. Its study is a constant reminder that in nature, the largest and most dramatic phenomena are often governed by the quiet, persistent action of the very small.