Eye of the beholder: Observer reference frame bias in Hickson-like compact groups of galaxies

TL;DR

This study quantifies how the identification of Hickson-like compact groups depends on the observer’s reference frame by using a mock lightcone from the Millennium I Simulation with a semi-analytic galaxy formation model. By placing 1000 random observers around each of 7709 CGs and re-applying velocity concordance and compactness criteria, the authors show that the velocity criterion is only mildly observer-dependent ($∼10\%$ fail) while the compactness criterion is highly so ($∼44\%$ fail), with combined cuts capable of yielding up to $84\%$ observer-independent groups at a significant sample-size cost. They further demonstrate that tightening surface brightness and velocity-difference thresholds can balance reliability and sample size, e.g., $μ ≤ 23$ mag arcsec$^{-2}$ reduces compactness dependence to $∼16\%$, and $ΔV ≤ 250$ km s$^{-1}$ reduces velocity-driven failures to $<4\%$, with the highest robustness ($∼84\%$) when both cuts are applied (sample ~7% of original). The analysis also finds bright interlopers seen from different lines of sight have only a minor effect on the main criteria, though local isolation is highly sensitive to projection; overall, observer-frame biases must be accounted for in Hickson-like CG studies. These results provide a practical framework for tailoring CG catalogs to specific science goals by adjusting selection thresholds.

Abstract

[Abridged] We investigate how the identification of Hickson-like CGs depends on the observer's reference frame, quantifying how frequently the same system would be recognised from different vantage points. Using a mock lightcone built from the Millennium I Simulation plus a semi-analytic model of galaxy formation, we identified 7709 CGs when applying the standard Hickson-like criteria. For each CG, we placed 1000 random observers on a surrounding sphere and reapplied the velocity and compactness requirements to test recoverability. We also examined the variation of population and local isolation. The velocity concordance criterion shows modest sensitivity to the observer's location: 10% of CGs fail for some observers, typically groups with members with high peculiar velocities (>1000 km/s). The compactness requirement is far more fragile, as 44% of CGs are missed by most observers, and these systems are very elongated or are chance alignments in real space. Tightening selection limits reduces this dependence. Lowering the surface brightness threshold to $μ\leq 23 \ mag/arcsec^2$ reduces the compactness dependence to 16%, while reducing the velocity limit to $ΔV\leq 250 \ km/s$ lowers velocity-driven failures to less than 4%. Applying both cuts simultaneously yields up to 84% observer-independent groups, although with a substantially smaller sample. Population and isolation are affected by bright interlopers seen from different directions. While such interlopers are common, they have only a minor effect on the compactness and velocity concordance criteria; however, the local isolation is commonly broken. Observer frame effects, dominated by the compactness criterion, can significantly bias Hickson-like CG samples. However, adjusting surface brightness and velocity difference thresholds allows users to balance the physical reliability according to their specific scientific goals.

Figures (9)

• Figure 1: Illustration showing two observers viewing a CG from different directions. The black galaxies represent the triplet in real space, and the white arrows indicate their respective peculiar velocities. Each observer sees the galaxies in redshift space with positions distorted according to the projection of their peculiar velocity vectors along the line of sight (galaxies in colours). Observer A (orange) views the triplet from a direction where the projected area on the sky is smaller, but the system is elongated along the line of sight, whereas Observer B (violet) perceives it as more extended in projection and not as elongated along the radial direction.
• Figure 2: Distribution among the 7709 CGs of the percentage of 1000 virtual observers seeing the CG fail the selection criteria. Top panel: Percentage of all-sky observers who lose the CGs because of the radial velocity filter. Bottom panel: Same but when analysing the compactness criterion (surface brightness, $\mu$). The inset yellow legends in both panels quote the percentage of the total sample of CGs that are lost for at least 5% of randomly orientated observers. Vertical red lines represent the medians (solid) and the 25th and 75th percentiles (dashed) of the distributions for the observer-dependent CGs (grey histograms). The teal bar between 0 and 5% represents the fraction of CGs that are considered robust.
• Figure 3: Scatter plot of the maximum and minimum galaxy peculiar velocity moduli for each CG. The grey dots represent the sample of CGs that are lost by fewer than 5% of the observers when applying the velocity filter, while the coloured dots are the sample of CGs that are lost for larger percentages of random observers. Different colours represent different percentages of observers who lose each CG (inset legends). The right panel displays the histograms of the maximum velocity moduli for non-lost CGs in grey and lost CGs in black (the coloured histograms are sub-samples of the black).
• Figure 4: Scatter plot of the largest 3D real space separation in the CG ($s_{\rm max}$) for the three (in the case of triplets) or four (for the remaining CGs) closest galaxies and the original line-of-sight elongation, features used to classify into Reals and CAs DiazGimenezMamon10. The light green region corresponds to Real CGs, while outside that region, CGs are classified as chance alignments (CAs). Within the CA class, we have also classified CGs as 'Fake' if the maximum 3D separation is larger than $1 \ h^{-1} \, {\rm Mpc}$ (dotted vertical line) or the minimum 3D separation is larger than 200 $h^{-1} \, {\rm kpc}$ (red dots). The top panel shows the distribution of CGs that satisfy the compactness criterion, almost regardless of the position of the observer (for more than 95% of them). The remaining panels display the samples where some observers (more than 5%) lose the CG because they no longer achieve the compactness criterion from their points of view. From the second row to the bottom, the panels display the CGs when the percentage of random observers who lose the CG is between 5 and 25; 25 and 50; 50 and 75; and 75 and 100.
• Figure 5: Percentage of observer-dependent (OD) CGs for four different sub-samples defined by varying the maximum surface brightness or the velocity gap in the line of sight. The left column shows the number of CGs (top panel), the percentages of CGs that are lost due to the velocity filter (middle panel), and $\mu$ criterion (bottom panel) as a function of the original $\mu_{\rm max}$ of the CG, i.e. each bin only takes into account CGs with an original $\mu$ less than the corresponding bin value ($\mu \leq \mu_{\rm max}$). The right column shows the same as the left column, but for samples selected as a function of the maximum velocity difference of galaxy members from the CG centre ($\Delta V \leq \Delta V_{\rm max} )$. In each case, the last bin corresponds to the original values of $\mu_{\rm lim}$ and $\Delta V_{\rm lim}$ of the CG sample.