The initial data of effective field theories of relativistic viscous fluids and gravity

This paper proposes a "reduction of order" approach applied specifically to initial data to uniquely determine unphysical degrees of freedom in relativistic viscous fluids and gravitational effective field theories, arguing that the resulting apparent Lorentz invariance breaking is acceptable within valid effective field theory frames.

The Gist

The Big Picture: Fixing the "Ghost" in the Machine

Imagine you are trying to predict the weather. You have a super-complex computer model that includes wind, rain, temperature, and even the movement of tiny dust particles. This model is mathematically perfect and "well-posed," meaning if you give it the right starting numbers, it will give you a unique, stable answer.

But here's the catch: The model is too detailed. It includes "ghost" variables—things like the exact position of every dust particle—that aren't actually part of the weather we care about. In physics, we call these unphysical degrees of freedom.

If you start your weather simulation with a random guess for where the dust is, the computer might go crazy. It might start vibrating wildly or exploding with noise that has nothing to do with real weather. This is a problem for scientists trying to simulate things like viscous fluids (like hot plasma in stars) or gravity (like black holes).

This paper proposes a clever solution: Don't change the rules of the game; just fix the starting line.

---

The Problem: The "Heavy" Backpack

In modern physics, scientists use something called Effective Field Theory (EFT). Think of EFT as a map.

• The Map: It shows the main roads (the "light" or "slow" things we care about, like fluid flow or gravity waves).
• The Detail: To make the map accurate, it also draws every single pebble and blade of grass (the "heavy" or "fast" things).

The problem is that the "pebbles" (unphysical modes) are so heavy and fast that they don't belong in the long-term picture.

• In Fluids: If you start with the wrong "pebble" data, the fluid might shake violently for a split second before settling down.
• In Gravity: If you start with the wrong "pebble" data, the shaking never stops. It creates high-frequency oscillations that ruin the simulation of black holes or the universe's expansion.

Scientists need to know: "What is the only correct way to set the starting numbers so these ghosts don't show up?"

The Solution: The "Reduction of Order" Trick

The authors suggest a method called "Reduction of Order," but with a twist.

Usually, when people try to fix this, they rewrite the entire equation to remove the "pebbles" entirely. But this is like trying to redraw the whole map to remove the grass. It's messy, breaks the rules of symmetry (Lorentz invariance), and can make the math unstable.

The Authors' New Idea:
Keep the complex, perfect map (the full equations) exactly as it is. Only use the "Reduction of Order" trick to decide how to set the starting numbers.

The Analogy: The Moving Train

Imagine a train moving at a constant speed $v$.

• The Physics: The train's position ($x$) and its speed ($\dot{x}$) are linked. If you know the train is moving at speed $v$, you know exactly how its position is changing.
• The Mistake: If you tell the computer, "The train is at position $x$, but its speed is randomly 100 mph," you are introducing a "ghost." The computer will try to fix this by making the train jerk violently to match the speed.
• The Fix: The authors say: "Don't guess the speed. Calculate it based on the position and the known rules of the train."
  • Rule: Speed = $v \times$ Position Change.
  • Action: When you start the simulation, you set the position. Then, you force the speed to be exactly what the position implies. You don't let the speed be a free variable.

By doing this, you ensure the "ghost" (the extra speed) is zero from the very first second. The simulation starts clean.

Why This Doesn't Break the Universe (The "Frame" Argument)

A critic might say: "Wait! If you use a specific rule to set the starting speed, aren't you breaking the laws of physics? Physics should look the same to everyone, no matter how fast they are moving!"

The authors say: No, not if you play by the rules of the "Effective Theory."

Think of EFT like a low-resolution camera.

• If you take a photo of a fast-moving car with a low-resolution camera, the car looks blurry. You can't see the details of the wheels spinning.
• If you try to analyze the photo, you assume the car is moving smoothly.
• If you zoom in too much (or look at the photo from a weird angle where the car looks super fast), the blur turns into static noise, and your "smooth car" theory breaks down.

The authors argue:

1. The Rule: You are only allowed to use this "Reduction of Order" trick if you are in a "frame of reference" where the physics is clear (the camera isn't too blurry).
2. The Result: If you are in a valid frame, the "error" you introduce by fixing the starting numbers is tiny—about the same size as the "blur" (the errors) already present in your low-resolution theory.
3. The Conclusion: You aren't breaking the laws of physics; you are just being consistent with the limits of your own map.

Real-World Applications

The paper shows how to do this for two big things:

1. Hot, Sticky Fluids (Hydrodynamics):

   • Imagine a neutron star (a super-dense, spinning ball of hot soup).
   • To simulate heat flowing through it, you need to know the temperature and how fast the temperature is changing.
   • The authors give a recipe: "Set the temperature. Then, calculate the rate of change using the heat flow rules. Do not guess the rate of change." This stops the simulation from exploding with fake noise.

2. Gravity (General Relativity):

   • Imagine simulating two black holes colliding.
   • Modern gravity theories add "corrections" (like adding a little extra spice to a recipe) to make them more accurate. But these corrections add "ghost" variables.
   • The authors show how to set the initial shape of space and its "speed" (how fast it's changing) so that the ghosts don't appear. This allows for cleaner, more accurate simulations of the universe.

The Takeaway

This paper is a "user manual" for the next generation of physics simulations.

• Old Way: "Here is a complex equation. Give me any starting numbers, and I'll try to solve it." (Result: Often crashes or produces fake noise).
• New Way: "Here is a complex equation. Here is the only correct way to set the starting numbers so the fake noise never appears. Now, let's run the simulation."

It's like telling a chef: "You can use this fancy, complex recipe, but you must measure the salt exactly based on the amount of flour. If you guess the salt, the dish will be ruined."

By fixing the starting data, they ensure that the "ghosts" of unphysical physics stay in the box, leaving only the real, beautiful physics to shine.

Technical Summary

1. Problem Statement

The paper addresses a critical issue in modern Effective Field Theories (EFTs) for relativistic viscous hydrodynamics (specifically the BDNK formalism) and gravitational EFTs (extensions of General Relativity).

• Well-Posedness vs. Unphysical Modes: Recent breakthroughs (BDNK for fluids, FHK for gravity) have established that these theories can be formulated as well-posed initial value problems. However, achieving well-posedness often requires introducing unphysical degrees of freedom (fast/heavy modes).
  • In BDNK hydrodynamics, equations become second-order in time (unlike the first-order Landau frame), requiring initial data for both the field and its time derivative.
  • In gravitational EFTs, equations involve higher-than-second derivatives of the metric, requiring initial data for higher time derivatives.
• The Initialization Ambiguity: While the physical ("slow") degrees of freedom are determined by the physical state, the initial data for the unphysical ("fast") modes is not uniquely determined by the physical state alone.
• Consequences of Improper Initialization:
  • Dissipative Systems (Fluids): Improperly initialized fast modes decay rapidly but induce spurious transient relaxation, potentially corrupting numerical simulations before the system settles.
  • Non-Dissipative Systems (Gravity): Without dissipation, improperly initialized fast modes persist indefinitely, manifesting as high-frequency oscillations or exponential growth that "infects" the physical solution, rendering the simulation unphysical.
• The Dilemma: Standard methods to eliminate these modes (reducing the order of the equations of motion) break Lorentz invariance (or general covariance) and often destroy the well-posedness of the system, making them unsuitable for numerical relativity or hydrodynamics simulations.

2. Methodology

The authors propose a "Reduction of Order" applied exclusively to Initial Data, rather than to the equations of motion.

• Core Concept: Retain the full, covariant, well-posed higher-order equations of motion. However, use the perturbative nature of the EFT (expansion in a small parameter $\lambda$, the UV cutoff scale) to constrain the initial data for the unphysical modes.
• The Procedure:
  1. Identify the physical (slow) modes and the unphysical (fast) modes.
  2. Use the lower-order (physical) equations of motion to express the highest time derivatives of the fields in terms of spatial derivatives and lower-order time derivatives.
  3. Substitute these expressions into the initial data requirements for the higher-order terms.
  4. This uniquely fixes the initial data for the unphysical modes in terms of the physical modes, effectively setting the amplitude of the fast modes to zero (or to their physically consistent value) at $t=0$.
• Addressing Lorentz Invariance: The procedure requires choosing a specific time coordinate (foliation) to perform the reduction, which formally breaks Lorentz invariance. The authors argue that this breaking is physically negligible provided the observer is in a frame where the EFT assumptions hold (i.e., gradients are small compared to the UV scale, $\lambda/L \ll 1$). In such frames, the error introduced by the reduction is of the same order as the truncation error of the EFT itself.

3. Key Contributions and Theoretical Framework

A. Toy Models (Section II)

The authors demonstrate the necessity and efficacy of their method using linear toy models:

• Causal Heat Conduction: They show that initializing the BDNK heat equation with arbitrary $\partial_t T$ leads to transient relaxation. Only the "order-reduced" initial data (where $\partial_t T$ is fixed by the spatial profile) yields a solution that matches the physical, non-transient evolution immediately.
• Dispersive Sound: In a non-dissipative model, improper initialization leads to persistent, unphysical high-frequency oscillations. The order-reduced initial data eliminates these oscillations entirely, recovering the correct infrared physics.

B. General Theorems (Section III)

The paper provides rigorous mathematical proofs for the validity of the approach in linear theories:

• Theorem 1 (Fast Modes): Proves that in a linear EFT where the highest derivative term is suppressed by $\lambda$, the extra degrees of freedom correspond to frequencies that diverge as $\lambda \to 0$. Thus, they are inherently "fast" modes outside the EFT validity regime.
• Theorem 2 (Existence of Reduction): Proves that an iterative differential operator exists that can systematically reduce the time-order of the equations, allowing the elimination of higher time derivatives in favor of spatial derivatives without changing the physical content of the theory up to the truncation order.
• Theorem 3 (Consistency): Demonstrates that the slow dispersion relations (physical modes) of the original high-order equation and the order-reduced equation agree to first order in $\lambda$.

C. Covariance Analysis (Section IV)

• Lorentz Invariance: The authors argue that while the reduction of order picks a specific time slicing, the resulting error is of order $(\lambda/L)^{m+1}$. If two observers (Alice and Bob) are in frames where the EFT is valid, their respective order-reduced initial data will differ only by terms of this order. Due to the well-posedness (continuous dependence on initial data), the resulting spacetime solutions will be indistinguishable within the accuracy of the EFT.
• General Covariance: The argument is extended to gravity, showing that diffeomorphism invariance is preserved modulo the EFT truncation error.

4. Applications (Section V)

The authors apply their framework to three specific, non-linear scenarios:

1. Causal Heat Conduction in Rotating Stars:

   • They derive the explicit initial data prescription for the BDNK heat equation in a rotating neutron star background.
   • The prescription fixes $\partial_t T_R$ (time derivative of redshifted temperature) in terms of spatial derivatives of $T_R$, ensuring no spurious transients in numerical simulations of rotating stars.

2. BDNK Hydrodynamics:

   • They derive the complex, non-linear expression required to initialize the BDNK equations for a relativistic viscous fluid.
   • They show that while a simplified initialization (using only the zeroth-order relation) is common, it is insufficient for achieving the full order-$\lambda$ precision promised by the BDNK theory. The full reduction (Eq. 33) is necessary for high-precision numerical relativity.

3. Regularized Einstein-Gauss-Bonnet (EGB) Gravity:

   • They address the initialization of gravitational EFTs with constraints.
   • They show how to use order reduction to fix the initial data for higher time derivatives of the extrinsic curvature ($\mathcal{L}_n K$) in terms of the metric and extrinsic curvature.
   • They discuss how to solve the resulting "order-reduced constraints" (which are higher-order elliptic PDEs) using perturbative expansions or conformal methods, noting that constraint damping techniques can handle the residual errors.

5. Results and Significance

• Resolution of Initialization Ambiguity: The paper provides a unique, physically motivated prescription for initializing the unphysical degrees of freedom in well-posed EFTs. This removes a long-standing ambiguity in numerical simulations of BDNK hydrodynamics and gravitational EFTs.
• Preservation of Physicality: By enforcing these specific initial conditions, the "fast" unphysical modes are not excited. The system evolves purely on the physical (slow) manifold, avoiding spurious transients (in fluids) or persistent high-frequency noise (in gravity).
• Compatibility with Numerical Relativity: Unlike previous methods that modified the equations of motion (breaking covariance and well-posedness), this method keeps the equations of motion intact. This allows the use of standard, stable numerical evolution schemes while ensuring the initial data is physically consistent.
• Theoretical Rigor: The paper bridges the gap between formal EFT reasoning (which assumes no heavy modes in initial states) and the practical requirements of solving partial differential equations. It proves that the "reduction of order" is not just a heuristic but a mathematically consistent procedure for defining the Cauchy problem in EFTs.

Conclusion:
The paper establishes that the "reduction of order" should be applied only to the initial data of relativistic viscous fluids and gravitational EFTs. This approach uniquely determines the unphysical degrees of freedom, ensuring that numerical simulations start in a state consistent with the low-energy effective theory, thereby eliminating unphysical artifacts while maintaining the mathematical well-posedness and covariance of the evolution equations.