Light from distant galaxies passes through warped space before reaching our telescopes. Massive clumps of matter act like uneven glass: galaxy images get slightly stretched and rotated. This phenomenon is called weak gravitational lensing. From it, astronomers infer how invisible mass is distributed across the Universe and ultimately compute an important number, S₈. It links the amount of matter and how strongly it clumps together.
But there's a catch. On relatively small scales, violent processes come into play: supernova explosions, winds from black holes, and heating of gas. They stir up ordinary matter and alter the lensing signal in unpredictable ways. To avoid guesswork, scientists devised a way to "cut off" these scales, like a baker sifting flour to remove lumps and working only with the clean, uniform fraction.
In this work, they applied a clever "k-cut" method with the BNT transform for the first time. It removes information from scales smaller than about 3 megaparsecs (roughly 10 million light-years — the distance to our neighboring large galaxy, the Andromeda Galaxy). After cutting, the result barely changed: S₈ = 0.798 ± 0.045. This means unaccounted astrophysics wasn't spoiling previous conclusions.
But when they used the observed data themselves instead of theory to estimate uncertainties, S₈ dropped to 0.717. And the larger the scale they considered, the lower the value fell. The difference between the extremes reached 1.8 sigma — not a discovery, but a curiosity. Computer simulations showed that neither the data nor the calculation method alone is to blame: the effect arises from their complex interplay.
So the k-cut method has become a dual-purpose tool. On one hand, it shields measurements from murky astrophysical "sediment." On the other, it highlights hidden inconsistencies in the processing pipeline itself. Perhaps by sifting data through ever-finer sieves, we'll catch the moment when the standard cosmological model starts to crack.
as written by the authors arXiv:2607.04384
We perform the first $$k$$-cut cosmic shear analysis of the KiDS-Legacy survey. This method uses the Bernardeau-Nishimichi-Taruya (BNT) transform to construct weak-lensing kernels that are more localised than conventional ones, and remove information from selected physical scales while retaining the constraining power of the targeted range. Removing the scale of $$k \geq 0.33~\mathrm{Mpc}^{-1}$$ from the KiDS-Legacy pseudo-$$C_\ell$$ data vector, and using a covariance matrix whose Gaussian component is computed from the theoretical data vector, we find $$S_8 = 0.798 \pm 0.045$$. This agrees with both the fiducial KiDS-Legacy bandpower result and our no-$$k$$-cut pseudo-$$C_\ell$$ posterior to within $$0.1\sigma$$, indicating no significant bias from nonlinear astrophysical feedback at the precision of KiDS-Legacy. We also study the case in which the Gaussian covariance is computed from the observed data vector. In this setup, the same scale cut of $$k < 0.33~\mathrm{Mpc}^{-1}$$ gives a much lower $$S_8=0.717_{-0.046}^{+0.047}$$. Further $$k$$-cut tests reveal a mild scale-dependent trend, with larger physical scales preferring lower $$S_8$$ values and a maximum low- versus high-$$k$$ deviation of $$1.80\sigma$$. Mock tests show that this behaviour is not produced by the covariance prescription or data vector alone, but may arise from their interplay. These results show that BNT $$k$$-cuts provide both a mitigation strategy for nonlinear systematics and a diagnostic of weak-lensing inference pipelines.
Weak gravitational lensing alters the apparent shapes of galaxies by only a few percent, like a perfect circle turning into a barely noticeable oval. To detect this, you need images of millions of galaxies and very precise mathematics.