Wave-mean decomposition of scale-dependent kinetic energy from surface drifters

This paper introduces a generalized Lagrangian mean framework to separate wave and mean-flow contributions to surface kinetic energy using drifter data, revealing that mean flows dominate rotational energy at scales larger than 1 km while divergent and rotational components are equipartitioned at smaller scales, with winter exhibiting more active mean flows and enhanced downscale wave energy transfer compared to summer.

The Gist

The Big Picture: Untangling the Ocean's "Noise" and "Signal"

Imagine the surface of the ocean is a busy dance floor. There are two types of movement happening at the same time:

1. The "Mean Flow" (The Slow Dance): Large, slow-moving currents and eddies that carry water (and anything floating in it) from one place to another over days or weeks. This is the "signal" or the steady rhythm.
2. The "Waves" (The Fast Jitter): High-frequency wiggles, ripples, and internal waves that shake things up quickly. This is the "noise" or the jittery movement.

The challenge for ocean scientists is that these two movements are mixed together. If you just look at a floating object (a drifter), you see a chaotic mess of both the slow drift and the fast shaking. It's hard to tell how much energy belongs to the slow currents versus the fast waves.

This paper introduces a new way to separate these two movements using data from thousands of floating drifters in the Gulf of Mexico.

The Tool: The "Lagrangian Filter" (The Moving Camera)

To separate the dance from the jitter, the authors used a technique called Lagrangian filtering.

• The Old Way (Eulerian): Imagine standing on a pier watching the ocean. You see a wave crash, then a current, then another wave. But because the current is moving, it makes the waves look faster or slower than they really are (like a Doppler shift). It's hard to tell where the wave ends and the current begins.
• The New Way (Lagrangian): Imagine you are on a surfboard, riding along with the slow current. From your perspective, the slow current feels like you are standing still. The fast waves, however, still zip past you. By filtering the data from the perspective of the moving surfboard (the "mean trajectory"), the authors can cleanly separate the slow drift from the fast waves.

The Key Innovation: The authors didn't just filter the speed; they filtered the path. They calculated where the drifters would have gone if they only followed the slow currents (the "mean trajectory"). Then, they measured the fast waves relative to that smooth path, rather than the jagged, real path the drifter actually took. This is like measuring how much a passenger is jiggling in a car seat relative to the car's smooth path, rather than relative to the bumpy road.

What They Found: The "Gulf of Mexico" Dance Floor

Using data from two different times of year (Summer 2012 and Winter 2016), they broke down the energy of the ocean surface.

1. The Size Matters (Scale)

• Big Scales (Larger than 10 km): The ocean is dominated by the Slow Dance (Mean Flow). The energy here is mostly rotational (spinning like a top), which is typical for large ocean currents.
• Small Scales (Smaller than 1 km): The Fast Jitter (Waves) takes over. Here, the energy is split almost evenly between spinning (rotational) and stretching/squeezing (divergent).

2. The Seasonal Difference

• Winter (LASER): The "Slow Dance" was more active and energetic in the winter, especially in the middle-sized zones (submesoscale). The "Fast Jitter" was concentrated in very small, tight spots. The authors suggest the stronger winter currents might be "shredding" the waves, breaking their energy down into smaller and smaller scales.
• Summer (GLAD): The "Slow Dance" was less active. The "Fast Jitter" was spread out over larger areas.

3. The "Divergent" Surprise
One of the most interesting findings is about the Mean Flow at small scales (under 1 km).

• Usually, we think of slow currents as just spinning (rotational).
• But the authors found that at small scales, the slow currents are also stretching and squeezing (divergent) just as much as they are spinning.
• Why this matters: Stretching and squeezing water horizontally forces water to move up or down vertically. This suggests that even the "slow" currents are driving vertical mixing, which is crucial for moving nutrients and heat in the ocean.

The "Helmholtz" Trap: Don't Just Look at the Spin

The paper also warns against a common shortcut scientists used to take.

• The Shortcut: Many researchers assumed that if they saw "spinning" motion, it was a slow current, and if they saw "stretching" motion, it was a wave. They used a math trick called Helmholtz decomposition on the raw, unfiltered data to make this guess.
• The Problem: The authors show this shortcut is often wrong. If you don't filter out the waves first, the "spinning" you see might actually be a mix of slow currents and fast waves.
• The Lesson: You have to separate the waves from the currents before you try to figure out if the currents are spinning or stretching. Otherwise, you are trying to read a book while someone is shaking the pages.

Summary in a Nutshell

The authors built a better "mathematical sieve" to separate the ocean's slow, steady currents from its fast, jittery waves. They found that:

1. Big currents are mostly spinning.
2. Small currents (under 1 km) are surprisingly active in both spinning and stretching, which helps mix the ocean vertically.
3. Winter currents are more energetic and break waves into smaller pieces than summer currents.
4. Old methods that didn't separate waves first were likely misinterpreting the ocean's energy.

This study gives us a clearer picture of how energy moves through the ocean surface, helping us understand how the ocean transports heat and nutrients.

Technical Summary

Technical Summary: Wave-Mean Decomposition of Scale-Dependent Kinetic Energy from Surface Drifters

Problem Statement
Separating oceanic motions into "balanced" (slow, geostrophic, or submesoscale currents) and "unbalanced" (fast, internal gravity waves) regimes is a fundamental challenge in ocean dynamics. While the Helmholtz decomposition (separating flows into rotational and divergent components) is a common heuristic—associating rotation with balanced flows and divergence with unbalanced waves—it is not a rigorous dynamical decomposition. Balanced flows can be divergent (e.g., submesoscale fronts), and unbalanced flows can be rotational (e.g., near-inertial waves). Furthermore, Eulerian temporal filtering is often compromised by Doppler shifting from balanced motions, making clean separation difficult. Lagrangian filtering offers a promising alternative by exploiting the cleaner time-scale separation between balanced and unbalanced motions in the flow-following frame, but its application to sparse, unstructured surface drifter data to derive scale-dependent statistics has been limited.

Methodology
The authors apply a Generalized Lagrangian Mean (GLM) framework to surface drifter data from two Gulf of Mexico campaigns: the Grand Lagrangian Deployment (GLAD, summer 2012) and the Lagrangian Submesoscale Experiment (LASER, winter 2016).

1. GLM-Inspired Lagrangian Filtering:

   • Drifter time series are filtered using a 5th-order Butterworth lowpass filter with a 1.5-day cutoff (near the inertial period).
   • Key Methodological Choice: The authors implement the filtering within the GLM framework. This means both the highpass (wave) and lowpass (mean flow) velocity components are attributed to the lowpass-filtered (mean) trajectory, rather than the raw particle trajectory. This choice is motivated by the need to remove Doppler shifts induced by mean-flow advection of wave energy, ensuring that the diagnosed mean flow represents the true balanced motion.
   • Highpass velocities are defined as the difference between unfiltered and lowpass velocities.

2. Scale-Dependent Statistics (SF2s):

   • Second-order velocity structure functions (SF2s) are computed for longitudinal and transverse velocity differences.
   • Unlike previous studies that assumed isotropy, the authors expand SF2s into azimuthal Fourier modes, focusing on the $n=0$ (angularly averaged) mode to handle anisotropic sampling without assuming isotropy.
   • SF2s are computed for:
     • Unfiltered velocities ($D_{LLo}, D_{TTo}$).
     • Lowpass (mean flow) velocities ($D_{LLs}, D_{TTs}$).
     • Highpass (wave) velocities ($D_{LLf}, D_{TTf}$), evaluated as a function of the lowpass separation vector $r_s$.

3. Helmholtz Decomposition of Filtered SF2s:

   • The authors apply Helmholtz decomposition formulae to the filtered SF2s to separate rotational ($D_{\psi\psi}$) and divergent ($D_{\phi\phi}$) contributions for both mean and wave components.
   • This allows for a direct comparison between the "wave-mean" decomposition (via Lagrangian filtering) and the "rotational-divergent" decomposition (via Helmholtz) to test the validity of using Helmholtz alone as a proxy for wave-mean separation.

Key Contributions

• First GLM-Inspired Wave-Mean Decomposition from Observations: This study presents the first application of a GLM-based wave-mean decomposition to oceanic observational data, specifically deriving scale-dependent kinetic energy statistics from surface drifters.
• Methodological Validation: The paper demonstrates that attributing filtered velocities to the mean trajectory (GLM framework) yields more physically interpretable results than attributing them to the raw particle trajectory. Specifically, the particle trajectory frame can introduce artifacts where mean-flow advection leaks into the diagnosed "wave" component, leading to unphysical rotational dominance in highpass SF2s.
• Critique of Helmholtz-Only Decomposition: The authors show that Helmholtz decomposition of unfiltered statistics alone is insufficient to diagnose the scale-dependent partition between balanced and unbalanced motions. While rotational dominance often correlates with balanced flows at large scales, the relationship breaks down at smaller scales where unbalanced motions can be rotational and balanced motions can be divergent.

Key Results

• Scale-Dependent Partitioning:
  • Large Scales (>10 km): Balanced (lowpass) motions dominate the kinetic energy, and rotational components dominate the lowpass SF2s, consistent with geostrophic balance.
  • Small Scales (<1 km): Unbalanced (highpass/wave) motions dominate the kinetic energy.
  • Transition: A transition from wave dominance to mean-flow dominance occurs at spatial separations of 5–10 km.
• Helmholtz Decomposition of Waves (Highpass):
  • Highpass SF2s are broadly consistent with linear internal waves, showing that divergent contributions are generally larger than or comparable to rotational contributions ($D_{\phi\phi f} \ge D_{\psi\psi f}$).
  • In winter (LASER), near-equipartition between rotational and divergent highpass SF2s extends to smaller scales (<10 km), suggesting enhanced near-inertial wave activity at submesoscales, potentially driven by stronger wave-mean interactions.
• Helmholtz Decomposition of Mean Flows (Lowpass):
  • Large Scales: Rotational mean flows dominate.
  • Submesoscales (<1 km): A remarkable equipartition between rotational and divergent lowpass SF2s is observed in both summer and winter. This suggests that low-frequency divergent motions (associated with vertical exchange) are as active as rotational motions at these scales, a finding not previously quantified in surface mean flows.
• Seasonal Differences:
  • Winter (LASER): Mean flows are more energetic and concentrated at submesoscales (500 m–10 km) compared to summer. Wave kinetic energy is concentrated at smaller spatial scales in winter, possibly reflecting enhanced downscale transfer by stronger submesoscale mean flows.
  • Summer (GLAD): Highpass motions show a broader decorrelation scale (>100 km) compared to winter (~1 km), indicating less concentration of wave energy at small scales.

Significance and Claims
The paper claims that its primary significance lies in providing a robust, observation-based method to quantify the contributions of waves and mean flows to surface kinetic energy across scales. By adopting the GLM framework, the authors argue they have achieved a cleaner separation of wave and mean dynamics than previous approaches, avoiding the artifacts associated with particle-trajectory-based filtering.

The results challenge the simplistic view that Helmholtz decomposition alone can separate balanced and unbalanced flows, particularly at submesoscales. The finding of rotational-divergent equipartition in mean flows at scales <1 km suggests that low-frequency divergent motions are a significant component of surface dynamics, with implications for vertical transport and tracer exchange. The study also highlights that oceanic states with similar mean-flow statistics can exhibit vastly different wave statistics, emphasizing the need for wave-mean separation in interpreting ocean observations, particularly for upcoming satellite missions like SWOT that sample these co-existing scales.