Extreme Value Statistics is a branch of statistics concerned with the most extreme deviations from a given distribution. In (Lovell et al., 2023) we used EVS to predict the most massive haloes and galaxies in the high redshift Universe. We found tension with some of the earliest JWST results, but over time those tensions have diminished as the calibration has improved, redshift outliers have been identified, and stellar mass estimates have been refined.
EVS constraints on the most massive galaxy compared to the latest version of the Labbe+22 results.
The EVS technique is now being used to evaluate candidates from Euclid and other observatories (missing reference). Christian Kragh Jespersen, a student at Princeton, has also worked on including the effects of cosmic variance in an EVS-like scheme, finding important deviations from the fiducial results for individual JWST pointings (Kragh Jespersen et al., 2024).
References
2024
-
On the Significance of Rare Objects at High Redshift: The Impact of Cosmic Variance
Christian Kragh Jespersen, Charles L. Steinhardt, Rachel S. Somerville, and 1 more author
Feb 2024
Publication Title: arXiv e-prints ADS Bibcode: 2024arXiv240300050K
The discovery of extremely luminous galaxies at ultra-high redshifts (\zrsim 8\) has posed a challenge for galaxy formation models. Most statistical analyses of this tension to date have not properly accounted for the variance due to field-to-field clustering, which causes the number counts of galaxies to vary from field to field, greatly in excess of Poisson noise. This super-Poissonian variance is often referred to as cosmic variance. Since cosmic variance increases rapidly as a function of mass, redshift, and for small observing areas, the most massive objects in deep }textit{JWST} surveys are severely impacted by cosmic variance. In this paper, we introduce a simple model to predict the distribution of the mass of the most massive galaxy found for different survey designs, which includes cosmic variance. The distributions differ significantly from previous predictions using the Extreme Value Statistics formalism, changing both the position and shape of the distribution of most massive galaxies in a counter-intuitive way. We test our model using the }texttt{UniverseMachine} simulations, where the predicted effects of including cosmic variance are clearly identifiable. Moreover, we find that the highly significant skew in the distributions of galaxy number counts for typical deep }textit{JWST} surveys lead to a high "variance on the variance", which greatly impacts the calculation of the cosmic variance itself. We conclude that it is crucial to accurately account for the impact of cosmic variance in any future analysis of tension between extreme galaxies in the early universe and galaxy formation models.
2023
-
Extreme value statistics of the halo and stellar mass distributions at high redshift: are JWST results in tension with ΛCDM?
Christopher C. Lovell, Ian Harrison, Yuichi Harikane, and 2 more authors
MNRAS, Jan 2023
ADS Bibcode: 2023MNRAS.518.2511L
The distribution of dark matter halo masses can be accurately predicted in the lambda cold dark matter (ΛCDM) cosmology. The presence of a single massive halo or galaxy at a particular redshift, assuming some baryon and stellar fraction for the latter, can therefore be used to test the underlying cosmological model. A number of recent measurements of very large galaxy stellar masses at high redshift (z \textgreater 8) motivate an investigation into whether any of these objects are in tension with ΛCDM. We use extreme value statistics to generate confidence regions in the mass-redshift plane for the most extreme mass haloes and galaxies. Tests against numerical models show no tension, neither in their dark matter halo masses nor their galaxy stellar masses. However, we find tentative \textgreater3σ tension with recent observational determinations of galaxy masses at high redshift from both Hubble Space Telescope and James Webb Space Telescope, despite using conservative estimates for the stellar fraction (f⋆ ~ 1). Either these galaxies are in tension with ΛCDM, or there are unaccounted for uncertainties in their stellar mass or redshift estimates.
