🔭 astrophysics

Interlopers as Signal in Line Intensity Mapping

This paper argues that interloping spectral lines in Line Intensity Mapping should be treated as calibrated tracers of large-scale structure rather than mere contaminants, demonstrating that their inclusion in cosmological models can robustly improve distance measurements and transform multiple frequency bands into a transverse BAO distance ladder across a wide redshift range.

Original authors: Anirban Roy

Published 2026-07-29

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Anirban Roy

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine the universe as a giant, three-dimensional party where every galaxy is a guest wearing a specific color of light. Astronomers use a technique called Line Intensity Mapping to take a blurry, wide-angle photo of this party. Instead of trying to see every single guest clearly (which is impossible because they are too far away), they measure the total glow of the entire room. This glow tells them how the guests are clustered together, revealing the hidden structure of the cosmos.

However, there's a tricky problem: the camera lens is a bit like a prism that mixes colors. When astronomers look at a specific "color" (or frequency) of light, they don't just see the guests wearing that exact color. They also see the glow from guests wearing different colors who happen to be standing at different distances (redshifts) but whose light has been stretched to match the camera's setting. In the jargon of the paper, these unwanted, overlapping signals are called "interlopers."

For a long time, the standard rule in this field has been to treat these interlopers like annoying static on a radio. The usual instinct is to scrub them out, mask them, or subtract them before trying to learn anything about the universe's expansion or its contents. But what if those "annoying" signals aren't just noise? What if they are actually secret messages from other parts of the universe's history, waiting to be decoded? This is the question a new study by Anirban Roy tackles, asking whether we should be deleting these signals or listening to them more closely.

The Signal in the Static

In this paper, the author suggests that the old rule of "delete the interlopers" might be throwing away a golden opportunity. Instead of viewing these overlapping signals as contaminants that ruin the data, the paper proposes we treat them as bonus tracers.

Think of it like listening to a choir where everyone is singing different notes at once. The traditional approach is to try to isolate the lead singer (the "target" line) and mute everyone else. But Roy argues that if you understand the music theory well enough, you can actually hear the harmony of the whole choir. The "interlopers" are just singers from different sections of the choir (different redshifts) who have been warped into the same frequency. If you model them correctly, they don't just add noise; they add a whole new layer of information about the shape of the universe at different times.

The Great Decision: Clean or Keep?

The paper frames this as a high-stakes decision problem. It asks: Is it better to clean the data and lose some information, or keep the messy data and try to model the mess?

To answer this, the author runs computer simulations (forecasts) for two future space missions: SPHEREx and FYST.

SPHEREx is like a wide-angle camera looking at a specific type of light (H-alpha) from galaxies about 1 billion light-years away. In this scenario, the "target" signal is loud and clear, but there are fainter "interlopers" from other lines like [O III] and [O II] mixed in.
FYST is a telescope looking at a very faint signal ([C II]) from the early universe (about 5.8 billion light-years away). Here, the "interlopers" (specifically CO lines) are actually much brighter than the target signal itself.

The simulations show that if you simply delete the interlopers, you lose a massive amount of data. But if you keep them and use a smart mathematical model to separate the signals, you can learn a lot more.

What the Simulations Reveal

The results are a mix of good news and cautionary tales, depending on what you are trying to measure:

The Robust Winners: The measurements of the universe's expansion history, specifically the Baryon Acoustic Oscillations (BAO) and the density of matter ( $\Omega_m h^2$ ), are very tough. They remain accurate even if the model of the interlopers isn't perfect. In the FYST simulation, modeling the interlopers turns two observed bands into a "distance ladder" that stretches from $z \approx 0.7$ to $z \approx 5.6$ . This is like building a ruler that can measure distances across almost the entire history of the universe, using only two snapshots of light.
The Sensitive Losers: Other measurements, like the clumping of matter ( $\sigma_8$ ) and the behavior of dark energy ( $w_0$ and $w_a$ ), are much more fragile. These parameters are sensitive to how well we understand the astrophysics of the galaxies. If our model of the interlopers is slightly off (for example, if we guess the brightness of the lines wrong by even a small amount), the results for these parameters can get biased. The paper suggests that while modeling interlopers can shrink the error bars (make the measurement more precise), a wrong guess about the physics could shift the answer in the wrong direction.

The Verdict: A New Strategy

The paper concludes that there is no single "best" way to handle interlopers. It's not a one-size-fits-all rule to "always clean" or "always keep." Instead, it's a strategic choice.

If you are measuring the BAO distance, you should definitely keep the interlopers. They act as extra anchors, turning a single blurry image into a multi-layered map of the cosmos.
If you are measuring dark energy or matter clumping, you have to be careful. You can keep the interlopers to get better precision, but only if you are very confident in your astrophysical models. If your model is shaky, the "noise" might trick you into a false conclusion.

In the end, the paper suggests that future surveys like SPHEREx and FYST shouldn't just treat these overlapping signals as enemies to be defeated. Instead, they should be viewed as calibrated components of the signal. By learning to listen to the whole choir rather than just the lead singer, astronomers might be able to map the universe's structure with a clarity they never thought possible. The "interlopers" aren't just contaminants; they are the key to a much richer, more detailed cosmic story.

Technical Summary: Interlopers as Signal in Line Intensity Mapping

Problem Statement
Line Intensity Mapping (LIM) measures the integrated emission from unresolved galaxies to map large-scale structure over vast cosmic volumes. A fundamental challenge in LIM is that a single observed frequency channel contains emission from multiple spectral lines originating at different redshifts. These "interlopers" are traditionally treated as contaminants that must be masked, deprojected, or subtracted before cosmological inference to ensure a clean map of the target epoch. However, this standard removal strategy may not be optimal for tomographic cosmological constraints. Interlopers are not random noise; they are themselves line intensity maps of large-scale structure at different epochs, merely geometrically warped into the target line's coordinate system. Discarding them potentially throws away valuable cosmological information and sacrifices target modes during the cleaning process.

Methodology
The authors formulate the treatment of interlopers as a Fisher decision problem rather than a fixed preprocessing step. The core question is whether the cosmological information gained by jointly modeling interlopers outweighs the cost of introducing astrophysical nuisance parameters (line amplitudes, biases, and shot noise).

Signal Modeling: The observed power spectrum $P_{obs}(k, \mu)$ is modeled as a projected sum of the target line and all contributing interlopers. Using the Alcock-Paczynski (AP) Jacobian and linear Kaiser redshift-space distortions, each line $L$ is evaluated at its own redshift $z_L$ and projected into the target coordinate system. This creates an anisotropic mapping where the interloper signal carries distinct quadrupole shapes dependent on the distance ratio between the interloper and target redshifts.
Fisher Analysis: The authors compute the Fisher matrix for cosmological parameters ( $\Omega_m h^2, \sigma_8, w_0, w_a$ $Ω_{m} h^{2}, σ_{8}, w_{0}, w_{a}$ ) and astrophysical nuisance parameters. They compare two scenarios:
1. Cleaned Analysis: Interlopers are removed, resulting in a loss of target information ( $F_{lost}$ ) but avoiding astrophysical degeneracies.
2. Joint Modeling: Interlopers are retained and modeled as additional tracers. The effective information gain is $F_{joint} = F_{target} + F_{eff}^{interloper}$ .
Decision Criterion: An estimator is deemed superior if it provides a precision gain ( $G_p = \sigma_{clean}/\sigma_{joint} > 1$ ) while keeping the induced parameter bias below a tolerance threshold ( $B_{max} = 0.5$ ). The bias is calculated assuming the true line amplitude deviates from the assumed calibration by half the prior width.
Survey Forecasts: The study applies this framework to two specific survey configurations:
- SPHEREx: Targeting H $\alpha$ at $z \approx 1$ with [O III], H $\beta$ , and [O II] interlopers.
- FYST (EoR-Spec): Targeting [C II] at $z \approx 5.8$ with a CO ladder interloper.

Key Results

Precision Gains: Joint modeling significantly tightens constraints on cosmological parameters. For the SPHEREx H $\alpha$ case, the uncertainty on $\sigma_8$ is reduced by a factor of $\approx 4$ . For the FYST [C II] case, where the target is faint and interlopers (CO) dominate, the uncertainty on $\sigma_8$ is reduced by a factor of $\approx 190$ , and constraints on dark energy parameters ( $w_0, w_a$ ) improve by up to a factor of 50.
Bias Sensitivity: The robustness of the results depends heavily on the parameter and the astrophysical priors.
- Robust Parameters: The physical matter density ( $\Omega_m h^2$ ) and BAO distance measurements ( $\alpha_\perp, \alpha_\parallel$ ) remain robust to astrophysical calibration errors (e.g., 20% errors in line amplitude produce sub-0.5 $\sigma$ shifts).
- Sensitive Parameters: The clustering amplitude ( $\sigma_8$ ) and dark energy parameters ( $w_0, w_a$ ) are more sensitive. While joint modeling improves their precision, a tight but miscentered astrophysical prior can introduce significant bias, as it restricts the likelihood's ability to self-calibrate interloper amplitudes.
BAO Distance Ladder: Modeling interlopers transforms a single observed frequency band into a transverse BAO distance ladder. By leveraging the fixed rest-frame wavelengths of interlopers, the SPHEREx optical band provides anchors from $z \approx 0.7$ to $3.3$, while the FYST [C II] band extends this to $z \approx 5.6$ . This allows for distance measurements across epochs that are difficult to reach with conventional galaxy surveys.

Significance and Claims
The paper argues that interloper removal is not automatically the optimal strategy for cosmological analysis in LIM. Instead, interlopers should be viewed as a calibrated component of the cosmological signal model. The authors claim that:

Interlopers are Tracers: Once included in the model, interlopers act as additional tracers of large-scale structure at other epochs, projected into the target frame.
Decision-Based Strategy: The optimal approach is a decision problem based on the trade-off between precision gains and bias tolerance, rather than a universal preprocessing rule.
Geometric Leverage: The most robust application of this method is in BAO distance measurements, which provide a clean geometric test and remain stable against astrophysical calibration uncertainties.
Future Implementation: For upcoming surveys like SPHEREx and FYST, the decision criterion (computable via Fisher analysis) can guide whether to excise interlopers or calibrate them as signal, potentially turning a source of contamination into a major source of cosmological leverage, particularly where the target signal is faint.

The authors note that these forecasts rely on empirical and halo-model prescriptions and that real data analyses will require handling foregrounds, masks, and non-Gaussian covariances. However, they assert that these complexities will shift decision boundaries but not alter the central claim: interlopers can be leveraged as cosmological signals under appropriate modeling conditions.

The Signal in the Static

The Great Decision: Clean or Keep?

What the Simulations Reveal

The Verdict: A New Strategy

Technical Summary: Interlopers as Signal in Line Intensity Mapping

More like this