Weak Lensing by Photometric Density Ridges

This paper reports a high-significance detection of weak gravitational lensing by photometric density ridges in Dark Energy Survey Year 3 data, utilizing improved algorithms to demonstrate that the signal primarily depends on the cosmological parameter $S_8$.

The Gist

Here is an explanation of the paper "Weak Lensing by Photometric Density Ridges," translated into everyday language with some creative analogies.

The Big Picture: Seeing the Invisible Web

Imagine the universe isn't just a bunch of scattered stars and galaxies. Instead, think of it as a giant, invisible spiderweb made of dark matter (the "cosmic web"). Galaxies don't just float randomly; they sit along the strands of this web, like dewdrops on a spider's thread.

We know this web exists, but we can't see the threads directly because dark matter doesn't emit light. However, we know that massive things bend light (like a funhouse mirror). So, if we look at the light coming from distant galaxies behind these threads, the threads should slightly distort the shapes of those background galaxies. This is called gravitational lensing.

The Problem: Finding the Threads

For years, astronomers have tried to find these threads.

• The Old Way: They looked for specific, bright galaxies (like Luminous Red Galaxies) and drew straight lines between them, assuming the dark matter threads were straight roads connecting two cities.
• The Flaw: Real cosmic threads aren't straight highways; they are winding, twisting, and curvy like mountain rivers. Drawing straight lines misses a lot of the action.

The New Solution: The "Mean Shift" Algorithm

This paper introduces a new, smarter way to find these threads using data from the Dark Energy Survey (DES).

The Analogy: The Crowd at a Concert
Imagine you are at a massive concert. You can't see the stage lights clearly, but you can see where the crowd is densest.

1. The Data: The "crowd" is the foreground galaxies. They are clumped together along the invisible dark matter threads.
2. The Algorithm (SCMS): The authors use a computer program called Subspace-Constrained Mean Shift (SCMS). Think of this as a magical wind that blows thousands of invisible "dust motes" (mesh points) across the sky.
   • These dust motes drift toward areas where the crowd (galaxies) is thickest.
   • However, they are "constrained." They can't just pile up on the single loudest singer (a single galaxy); they are forced to line up along the ridges of the crowd.
   • Eventually, the dust motes settle into perfect lines that trace the shape of the invisible threads.

The Result: Instead of guessing straight lines, the computer draws the actual, winding, curvy paths of the cosmic web.

The Experiment: Measuring the Bend

Once the computer has mapped these "ridges" (the threads), the team looks at the background galaxies.

• The Test: Do the background galaxies look slightly stretched or squashed in a specific direction relative to these newly found threads?
• The Finding: Yes! They found a clear signal. The light from background galaxies is indeed being bent by the mass of these photometric ridges.

The "Catch" (Why it's not a magic bullet yet)

The authors are honest about the limitations.

• The Overlap: Because these threads connect galaxies, the lensing signal they see is a mix of the thread itself and the galaxies at the ends of the thread. It's hard to separate the "rope" from the "knots" at the end.
• The Analogy: Imagine trying to measure the weight of a rope by weighing the whole bundle of rope plus the heavy hooks at the ends. You know the rope is there, but you can't get a precise weight of just the rope without subtracting the hooks.
• Current Status: They detected the effect with high confidence (a "5-sigma" detection, which is a gold standard in science), but it's not yet independent enough to give us brand-new, precise numbers about the universe's composition.

The Secret Ingredient: $S_8$

The paper found that the strength of this bending signal depends heavily on a specific cosmological number called $S_8$.

• What is $S_8$? Think of it as the "Clumpiness Score" of the universe. It tells us how much matter likes to clump together to form structures like galaxies and filaments.
• The Discovery: The more "clumpy" the universe is (higher $S_8$), the stronger the lensing signal from these ridges. This confirms that the method is sensitive to the fundamental physics of how our universe grows.

The Conclusion: A New Tool in the Box

This paper is a proof-of-concept. It says:

"We have built a new tool (the SCMS algorithm) that can trace the winding roads of the cosmic web more accurately than before. We used it to find a new type of lensing signal in real data. It works! It depends on the universe's clumpiness. Now, we just need to refine the tool to separate the 'rope' from the 'hooks' so we can use it to measure the universe with extreme precision."

In short: They found a way to trace the invisible skeleton of the universe using a smart computer algorithm, proved that this skeleton bends light, and showed that the strength of that bend tells us how "clumpy" our universe is.

Technical Summary

Here is a detailed technical summary of the paper "Weak Lensing by Photometric Density Ridges" by Nikjoo, Zuntz, and Moews.

1. Problem Statement

Gravitational lensing is a primary tool for mapping the distribution of dark matter and probing cosmology. While standard weak lensing analyses rely on two-point statistics (cosmic shear and galaxy-galaxy lensing), these capture only Gaussian information. The non-linear evolution of the cosmic web introduces higher-order information contained in the topology of the large-scale structure, specifically in filaments.

Previous attempts to detect lensing by filaments have largely relied on:

• Spectroscopic data to identify 3D filaments.
• Assuming filaments are straight lines connecting pairs of galaxies (e.g., Luminous Red Galaxies), which simplifies the geometry but ignores the true curved nature of the cosmic web.
• "Nulling" techniques that remove the signal from the galaxies at the filament ends, making the measurement independent of standard galaxy-galaxy lensing but potentially losing signal-to-noise.

The Gap: There is a need to detect lensing signals from photometrically identified ridges (2D projections of 3D filaments) without assuming they are straight lines. This approach utilizes the high density of foreground galaxies in photometric surveys but risks mixing the filament signal with the halo signals of the galaxies themselves.

2. Methodology

The authors propose a pipeline to identify density ridges in foreground galaxy maps and measure the tangential shear of background galaxies around them.

A. Ridge Identification: Subspace-Constrained Mean Shift (SCMS)

The core of the methodology is the SCMS algorithm, implemented via the DREDGE package.

• Concept: The algorithm treats the galaxy distribution as a density field. It iteratively moves a mesh of points toward high-density regions but constrains them to move along the subspace of the ridge (where the Hessian matrix has $D-1$ negative eigenvalues), effectively tracing the "crests" of the density field rather than just peaks.
• Improvements: The authors optimized the algorithm for large datasets ($>10^6$ points) by:
  • Parallelizing over mesh points using MPI.
  • Implementing individual point convergence criteria (stopping updates for points that have converged) rather than global ensemble convergence.
  • Using Ball Trees for efficient nearest-neighbor searches.
  • Adding Just-In-Time (JIT) compilation via Numba.

B. Segmentation (Ridge to Filament)

The raw output of SCMS is a continuous set of ridge points. To analyze them as discrete filaments:

1. Minimum Spanning Tree (MST): Constructed to connect all ridge points.
2. Branching Removal: Nodes with a degree $>2$ (junctions) are identified and removed to split the MST into linear segments.
3. Clustering: The DBSCAN algorithm is applied within each segment to group points into physically coherent filaments and remove noise.

C. Shear Measurement

• Geometry: For each background source galaxy, the algorithm identifies the nearest point on a foreground filament segment.
• Transformation: The source ellipticity is rotated by the position angle $\phi$ relative to the filament to compute the tangential shear ($\gamma_+$) and cross shear ($\gamma_\times$).
• Data Selection:
  • Lenses (Foreground): Photometric redshift $z < 0.4$.
  • Sources (Background): Photometric redshift $z > 0.4$.
  • Data Source: Dark Energy Survey (DES) Year 3 (Y3) GOLD catalog (Metacalibration shapes and MagLim lenses).

3. Key Contributions

1. Novel Metric: Introduction of "ridge lensing" as a new statistical metric that utilizes the full 2D photometric density field without assuming straight-line geometries.
2. Algorithmic Optimization: Significant speed-ups and robustness improvements to the SCMS algorithm, enabling its application to massive photometric surveys.
3. High Significance Detection: The first detection of weak lensing by photometrically defined ridges in real survey data (DES Y3) with high statistical significance.
4. Cosmological Sensitivity Analysis: Demonstration that the ridge lensing signal is primarily sensitive to the structure growth parameter $S_8 = \sigma_8 (\Omega_m/0.3)^{1/2}$.

4. Results

A. Signal Characteristics (Noiseless Simulations)

• Signal Morphology: The tangential shear ($\gamma_+$) shows a positive signal at large separations and a negative signal at small separations (below the smoothing bandwidth). The negative signal arises because background galaxies aligned perpendicular to the ridge are lensed by structures along the ridge.
• Parameter Dependence:
  • $S_8$: The signal amplitude scales monotonically with $S_8$. Higher $S_8$ yields a stronger signal.
  • $\Omega_m$ and $\sigma_8$: The signal shows intermediate dependence on these parameters individually, consistent with their combination in $S_8$.
  • Algorithmic Parameters: The signal is robust to changes in mesh size and convergence thresholds. However, the bandwidth (smoothing scale) significantly affects the signal structure; smaller bandwidths make the signal resemble standard galaxy-galaxy lensing (dominated by individual halos), while larger bandwidths smooth out small-scale features.

B. Real Data Detection (DES Y3)

• Detection: Using the fiducial parameters (Bandwidth = 6 arcmin, Density Threshold = 15th percentile), the authors detected a clear tangential shear signal in DES Y3 data.
• Significance:
  • Simulations: Signal-to-Noise Ratio (S/N) $\approx 57$.
  • Real Data: S/N $\approx 47$.
  • The cross-shear ($\gamma_\times$) is consistent with zero, validating the absence of systematic biases.
• Density Threshold: The detection significance peaks at the 55th percentile density threshold, balancing the selection of high-density (stronger lensing) ridges against the reduction in the number of available structures.

C. Independence from Galaxy-Galaxy Lensing

The authors note that while the signal morphology differs from standard galaxy-galaxy lensing, the two are not independent.

• At small bandwidths or large scales, the ridge signal converges toward the galaxy-galaxy lensing signal.
• The method does not "null" the signal from the galaxies at the ends of the filaments, meaning the detection is partially driven by the same halo-lensing physics as traditional galaxy-galaxy lensing.

5. Significance and Future Work

• Cosmological Utility: The detection confirms that photometric ridges are effective tracers of the 3D cosmic web and contain real cosmological information, specifically sensitive to $S_8$.
• Limitations: Currently, the method is not independent of galaxy-galaxy lensing, limiting its utility as a purely "higher-order" statistic for breaking degeneracies.
• Future Directions: To achieve precision parameter estimation, the authors propose:
  • Implementing more realistic redshift cuts in simulations.
  • Incorporating lens sample weights to account for survey depth variations.
  • Developing methods to subtract the galaxy-galaxy lensing component to isolate the pure filamentary signal.
  • Using simulation-based inference (emulation) to constrain cosmological parameters directly from ridge lensing.

In conclusion, this work establishes a robust framework for measuring weak lensing around the curved, photometrically identified filaments of the cosmic web, achieving a high-significance detection in DES data and demonstrating its sensitivity to key cosmological parameters.