The Marked Power Spectrum as a Practical Bispectrum Measure for Galaxy Redshift Surveys

This paper demonstrates that the marked power spectrum is a practical and efficient alternative to the bispectrum for extracting non-Gaussian information from galaxy redshift surveys, as it effectively breaks parameter degeneracies, accommodates standard survey geometry treatments, and allows for cosmological inference through smooth interpolation rather than complex analytical modeling.

The Gist

Imagine you are trying to understand the layout of a massive, invisible city built out of stars and galaxies. For a long time, astronomers have been mapping this city by measuring the average distance between buildings. They ask, "On average, how far apart are these galaxies?" This is like measuring the Power Spectrum. It's a great tool, but it's a bit like looking at a city from a high-altitude drone: you see the general shape, but you miss the details of the neighborhoods, the traffic patterns, and the unique quirks of specific districts.

The problem is that this "average distance" view hides some secrets. Different combinations of cosmic rules (like how much dark matter there is or how fast the universe is expanding) can create the exact same average distances. It's like two different cities having the same average distance between houses, but one is a grid of skyscrapers and the other is a sprawling suburb. You can't tell them apart just by looking at the average.

To solve this, scientists need to look at higher-order patterns—how groups of three or four galaxies cluster together. This is usually done with a tool called the Bispectrum. Think of the Bispectrum as a super-detailed 3D scan of the city. It's incredibly powerful, but it's also heavy, slow, and complicated to process. It requires massive computers and huge amounts of data storage, making it hard to use for the next generation of giant sky surveys.

The "Marked" Solution: A Smart Filter

This paper introduces a clever new tool called the Marked Power Spectrum (MPS).

Imagine you are a real estate agent looking at a city map. Instead of just measuring distances, you decide to highlight specific types of neighborhoods.

• The Old Way (Power Spectrum): You measure the distance between every house, regardless of whether it's a mansion or a shack.
• The New Way (Marked Power Spectrum): You decide to give a "bonus weight" or a "mark" to houses in empty, quiet neighborhoods (voids) and ignore the crowded city centers.

By "marking" these empty spaces, you change the data. You aren't just measuring distance anymore; you are measuring the relationship between the emptiness and the density.

Why is this a big deal?

1. It's Lighter: Unlike the heavy 3D scan (Bispectrum), the Marked Power Spectrum still looks like a simple 2D map (a Power Spectrum). It fits into the existing software and computers astronomers already use. It's like upgrading your drone software to highlight empty lots without needing a new, massive supercomputer.
2. It Breaks the Tie: Because it focuses on the empty spaces, it reacts differently to the rules of the universe than the standard map does. If two different cosmic theories predict the same average distances, they will likely predict different patterns in the empty spaces. This allows scientists to tell the theories apart, breaking the "degeneracy" (the tie) that has held them back.
3. It's Practical: The authors show that you can apply this to real survey data (like the DESI survey) without getting bogged down by the complex shapes of the sky being observed. They proved that the "window" of the telescope doesn't break the math.

The "Smooth" Secret

One of the biggest hurdles in cosmology is that the universe is messy. When you try to model it, you often have to guess about the "noise" (random errors) or the tiny, chaotic details that our math can't quite reach.

The authors discovered something magical about the Marked Power Spectrum: its behavior is incredibly smooth.

Imagine trying to guess the temperature of a room. If you have a thermometer that jumps wildly from 60°F to 90°F to 40°F, you can't trust it. But if the thermometer moves smoothly from 60 to 61 to 62, you can predict the future temperature easily.

The paper shows that the Marked Power Spectrum changes smoothly as you change the cosmological rules. This means scientists don't need to build a complex, theoretical equation for every single possible universe. Instead, they can just measure a few key points and interpolate (draw a smooth line between them) to find the answer. It turns a complex math problem into a simple "connect the dots" game.

The Bottom Line

This paper is like handing astronomers a smart filter for their data.

• Before: They had a blurry photo (Power Spectrum) that couldn't tell two cities apart, or a heavy, slow 3D scanner (Bispectrum) that was too hard to use.
• Now: They have a tool that takes the easy-to-use photo but adds a "highlighter" that reveals the hidden details of the empty spaces.

This allows them to use the massive data from upcoming telescopes to finally answer deep questions about the universe—like the weight of neutrinos or the nature of dark energy—without needing to wait for supercomputers to catch up. It's a practical, powerful step forward in decoding the universe's blueprint.

Technical Summary

Here is a detailed technical summary of the paper "The Marked Power Spectrum as a Practical Bispectrum Measure for Galaxy Redshift Surveys" by Haruki Ebina, Martin White, and Edmond Chaussidon.

1. Problem Statement

Modern cosmological surveys (e.g., DESI) have reached a precision where extracting information from higher-order statistics (non-Gaussianity) is essential to break parameter degeneracies and constrain cosmological models (e.g., neutrino mass, modified gravity).

• The Challenge: Traditional higher-order statistics, such as the bispectrum (3-point function), offer this information but suffer from significant practical drawbacks:
  • Computational Complexity: Calculating the bispectrum is computationally expensive.
  • Data Vector Size: The number of triangle configurations creates a massive data vector, requiring large simulation suites to estimate the covariance matrix.
  • Modeling Difficulties: Handling survey window functions, systematics, and nuisance parameters for the bispectrum is more complex than for the standard power spectrum (2-point function).
• The Goal: Develop a method to access bispectrum-level information while retaining the computational efficiency, small data vector size, and established modeling infrastructure of the power spectrum.

2. Methodology

The authors propose and refine the Marked Power Spectrum (MPS) as a practical alternative. The core idea is to weight the galaxy density field by a "mark" function dependent on the local smoothed density, thereby encoding higher-order correlations into a 2-point statistic.

A. Theoretical Restructuring

• Definition: The marked density field is defined as $\rho_M(x) = m[\delta_{g,R}(x)] \rho_g(x)$, where $m$ is a mark function of the smoothed overdensity $\delta_{g,R}$.
• Isolation of Higher-Order Info: The authors restructure the MPS to explicitly separate the standard power spectrum information from the new higher-order information.
  • They define the MPS $M$ specifically as the cross-spectrum between the unmarked field and a linear mark ($m = 1 + \delta_{g,R}$).
  • This isolates terms proportional to the bispectrum (specifically the tree-level bispectrum $B_{112}$) and removes the dominant two-point contributions that merely duplicate the standard power spectrum.
• Perturbative Modeling:
  • The MPS is modeled using Eulerian Perturbation Theory (EPT) up to one-loop order.
  • Crucially, the smoothing kernel ($W_R$) prevents UV-divergent contact terms, placing the MPS on the same theoretical footing as the standard power spectrum regarding renormalization.
  • Nuisance Parameters: The model introduces only a small set of new nuisance parameters ($B_{shot}, A_{shot}, A_0$) related to stochasticity and higher-loop corrections, which are manageable and largely degenerate with existing parameters.

B. Survey Geometry and Window Functions

• A major hurdle for applying MPS to real data is the complex survey footprint (mask).
• The authors demonstrate that the MPS can be modeled using the same window matrix formalism as the standard power spectrum.
• By defining a modified FKP field ($F_M = m n_g - \bar{m} \alpha n_r$), they show that the survey geometry effects can be convolved with the theoretical model using existing codes (e.g., pypower), avoiding the need to rebuild window function formalisms from scratch.

C. Covariance and Alcock-Paczynski (A-P) Effects

• Covariance: The authors derive the covariance matrix for the MPS and its cross-covariance with the power spectrum.
  • Unlike the bispectrum, where cross-covariance is often suppressed, the MPS cross-covariance is not strongly suppressed due to the integration over the inner momentum $p$ in the smoothing kernel. However, the Gaussian approximation is sufficient to capture the diagonal terms.
• A-P Effect: Instead of analytically deriving the complex coordinate transformation for the MPS (which involves the bispectrum integral and smoothing), the authors show that the cosmological dependence of the MPS is smooth. This allows for efficient interpolation between cosmologies during MCMC inference, bypassing expensive analytical calculations.

3. Key Contributions

1. Theoretical Refinement: A rigorous redefinition of the MPS to isolate pure higher-order information (bispectrum terms) while minimizing overlap with the power spectrum.
2. Practical Implementation: A complete framework for applying MPS to real survey data, including:
   • Integration of survey window functions using standard 2-point function infrastructure.
   • Perturbative modeling including stochasticity and nuisance parameters suitable for next-generation surveys.
3. Covariance Analysis: An analytical derivation showing that while MPS cross-covariance is not suppressed like the bispectrum, it is manageable, and the Gaussian approximation holds for diagonal terms.
4. Validation: Extensive validation on mock catalogs (periodic boxes and DESI DR1 cutsky mocks) demonstrating that the theory matches simulations within $1\sigma$.

4. Results

• Degeneracy Breaking: The MPS is highly sensitive to the quadratic bias parameter $b_2$.
  • Figure 2 demonstrates that a $\sim 1\sigma$ shift in $b_2$ causes a change in the MPS amplitude by a factor of 2, whereas the standard power spectrum multipoles ($P_0, P_2$) remain nearly indistinguishable.
  • This signal is an order of magnitude larger than observational errors and higher-order theoretical uncertainties.
• Stochasticity: The authors confirm that the stochasticity introduced by estimating the density field from a finite number of galaxies is degenerate with existing shot-noise parameters and does not require new free parameters for current survey densities.
• Mock Validation:
  • Periodic Boxes: Joint fits of Power Spectrum ($P$) and MPS ($M$) to 25 AbacusSummit mocks achieved $\lesssim 1\sigma$ residuals for various smoothing scales ($R=10, 15, 20 \, h^{-1}\text{Mpc}$) and number densities.
  • Cutsky Mocks: Applied to DESI DR1 cutsky mocks (simulating the survey mask). The joint fit of $P$ and $M$ remained consistent within $1\sigma$, proving the window function implementation is robust.
• A-P Interpolation: The residual variation of the MPS under different cosmologies ($\Omega_m, w_0, w_a$) was found to be smooth, validating the interpolation strategy for MCMC chains.

5. Significance

This work establishes the Marked Power Spectrum as a practical, high-performance tool for next-generation cosmological analyses (e.g., DESI, Euclid, Rubin).

• Efficiency: It offers the degeneracy-breaking power of the bispectrum (specifically for bias parameters like $b_2$) with the computational cost and data vector size of the power spectrum.
• Infrastructure Reuse: By leveraging existing 2-point function pipelines (window functions, covariance estimation, bias modeling), it lowers the barrier to entry for incorporating higher-order statistics into standard cosmological inference.
• Future Outlook: The authors provide a clear path forward, noting that while fiber assignment effects and full covariance validation with larger simulation suites are future work, the current framework is ready for application to near-future datasets.

In summary, the paper transforms the Marked Power Spectrum from a theoretical concept into a robust, observationally ready statistic that bridges the gap between the simplicity of 2-point analyses and the information richness of 3-point statistics.