4MOST Cosmology Redshift Survey (CRS)

• 4MOST CRS is a spectroscopic survey that maps cosmic structure by targeting bright galaxies, LRGs, and quasars from z = 0.15 to 3.5 over thousands of square degrees.
• It employs rigorous target selection, advanced tiling algorithms, and joint fibre configurations to optimize redshift completeness and minimize systematic errors.
• Forecasts indicate that CRS achieves ~3% BAO precision and enhances multi-probe cosmology when combined with datasets from DES, KiDS, Euclid, and other surveys.

The 4MOST Cosmology Redshift Survey (CRS) is Consortium Survey 8 of the 4-metre Multi-Object Spectroscopic Telescope, designed as the principal southern-hemisphere spectroscopic large-scale-structure programme within the 4MOST survey suite. In its original consortium definition, CRS was described as a multi-tracer survey of bright galaxies, luminous red galaxies, emission-line galaxies, and quasars, totalling about 8 million objects over the redshift range $z=0.15$ to $3.5$ across 7500 deg$^2$; later target-selection and clustering papers based on DESI Legacy Survey DR10.1 presented an operational configuration of nearly 5.4 million spectroscopic redshifts over $\sim 5700$ deg$^2$, with bright galaxies and luminous red galaxies forming the low- and intermediate-redshift backbone and quasars providing the higher-redshift extension (Richard et al., 2019, Verdier et al., 10 Aug 2025, Bandi et al., 2 Oct 2025).

1. Survey identity and programmatic evolution

CRS was conceived as the major southern spectroscopic large-scale-structure dataset at intermediate redshift, explicitly optimized for overlap with southern weak-lensing imaging and other cosmological probes. The 2019 consortium survey paper positioned it as the principal cosmology and large-scale-structure redshift survey in the southern hemisphere within 4MOST, with strong overlap with DES, KiDS, LSST, Euclid, SKA, and southern CMB experiments. Later implementation papers retained the same cosmological role but reformulated the target catalogues on DESI Legacy Survey DR10.1 imaging, with updated footprint, masking, and target-density choices (Richard et al., 2019, Verdier et al., 10 Aug 2025).

Feature	Original consortium definition	Later implementation papers
Area	7500 deg$^2$ main area, plus a 1000 deg$^2$ ELG subset	$\approx 5700$ deg$^2$
Scale	about 8 million objects	nearly 5.4 million spectroscopic redshifts
Main tracers	BG, LRG, ELG, QSO, QSO–Ly$\alpha$	BG, LRG, QSO

The two descriptions correspond to different stages of the programme. The original Messenger survey note emphasized the full multi-tracer design, including an emission-line galaxy component over a 1000 deg$3.5$0 sub-area. The later DR10.1-based work described a DESI-aligned but 4MOST-optimized implementation in which BG and LRG catalogues were specified in detail and validated, and the ELG sample was dropped because of fibre constraints and Euclid’s forthcoming ELG spectroscopy (Richard et al., 2019, Verdier et al., 10 Aug 2025).

CRS is therefore best understood as both a survey concept and an evolving operational realization. Its stable defining features are its southern footprint, spectroscopic clustering focus, wide redshift coverage, and explicit role in joint analyses with lensing, CMB, radio, and X-ray datasets.

2. Tracer hierarchy and sample definition

The original CRS architecture was explicitly multi-tracer. The 2019 survey description defined bright galaxies at $3.5$1 with density $3.5$2 over 7500 deg$3.5$3, luminous red galaxies at $3.5$4 with density $3.5$5 over 7500 deg$3.5$6, emission-line galaxies at $3.5$7 with density $3.5$8 over a 1000 deg$3.5$9 subset, low-$^2$0 quasars at $^2$1 with density $^2$2, and a QSO–Ly$^2$3 subset at $^2$4 with density $^2$5 (Richard et al., 2019).

That original target hierarchy encoded the survey’s cosmological logic. BGs were the low-$^2$6 clustering and galaxy–galaxy lensing anchor; LRGs were the intermediate-$^2$7 high-bias tracer for BAO and RSD; ELGs supplied a denser star-forming population at $^2$8–1.1; low-$^2$9 QSOs extended the matter-tracer field into the quasar regime; and QSO–Ly$\sim 5700$0 targets enabled Ly$\sim 5700$1-forest cosmology and high-$\sim 5700$2 BAO (Richard et al., 2019).

The later DR10.1 implementation sharpened this architecture for the galaxy samples. BG selection was defined as an $\sim 5700$3 magnitude-limited sample in the redshift range $\sim 5700$4, with CRS density reduced to $\sim 5700$5. The key additional BG colour cuts were

$\sim 5700$6

combined with Gaia- and Tycho-based star removal, Gaia astrometric filtering, fibre-magnitude cuts, and standard Legacy Survey masks. LRG selection targeted $\sim 5700$7 at $\sim 5700$8, using $\sim 5700$9 colours together with cuts such as

$^2$0

$^2$1

$^2$2

and the fibre condition

$^2$3

The same paper states that the ELG sample is not included in this later implementation (Verdier et al., 10 Aug 2025).

These definitions place CRS deliberately close to DESI in imaging basis and target philosophy, while retaining 4MOST-specific density and completeness trade-offs. A plausible implication is that joint 4MOST–DESI analyses are simplified not only by overlapping cosmological objectives but also by deliberately similar sample construction.

3. Instrumentation, operations, and tiling

CRS is enabled by 4MOST as a permanent wide-field, high-multiplex spectroscopic facility on VISTA. In the later CRS target-selection work, 4MOST is described as a new 4 m, high-multiplex spectroscopic facility on VISTA, with 2436 fibres over 4.2 deg$^2$4, running a 5-year multi-survey program. More generally, 4MOST was designed for a wide field $^2$5 deg$^2$6 with goal $^2$7 deg$^2$8, high multiplex $^2$9 fibres with goal 3000, and low-resolution spectroscopy at $^2$0 over 400–900 nm with goal coverage of 390–1050 nm; the facility design explicitly included “redshift surveys to $^2$1,” “Large Scale Structure of galaxies (e.g. BAO),” and the “sample of >100,000 galaxy redshifts needed to calibrate the Euclid photometric redshifts” [(Verdier et al., 10 Aug 2025); (Jong et al., 2012)].

Operationally, CRS does not receive an isolated observing stream. The 4MOST survey strategy merged most science programmes into one joint survey, observed simultaneously through shared fibre configurations and Observation Blocks. Default fibre-configuration exposures are $^2$2 minutes, with a maximum of 30 minutes to avoid fibre drift from differential refraction; the fiducial strategy observes every field twice, with Observation Blocks containing three fibre configurations of about 20 minutes each, giving a typical total integration of about 2 hours per field. In the survey-strategy paper, CRS itself appears as Survey S8 with 8.0 million targets in the brightness range $^2$3 (Guiglion et al., 2019).

The resulting tiling problem is non-trivial because target density, required exposure time, and sky-brightness preference vary strongly across the joint 4MOST programme. The dedicated 4MOST tiling algorithm models the tiling pattern as a marked point process with probability density

$^2$4

where the energy function is decomposed as

$^2$5

These terms respectively encode target-completion efficiency, overhead minimization, geometric interactions between Observation Blocks, and the distribution of Bright/Grey/Dark tiles. Optimization is performed with simulated annealing and Metropolis–Hastings birth, death, and change moves. For the combined 4MOST mock catalogues, the default optimized solution used 40,503 tiles and 12,375 Observation Blocks, with 9,694 hours of exposure time and 13,386 hours of total telescope time, corresponding to a 72.4% observing fraction (Tempel et al., 2020).

For cosmological analyses this infrastructure is not merely logistical. Tile geometry, repeated passes, sky-brightness allocation, and probabilistic fibre assignment all enter the CRS angular and radial selection functions. The strategy papers therefore treat completeness, exposure accumulation, and field overlap as core survey-design variables rather than downstream nuisances.

4. Cosmological observables and forecasted performance

CRS was defined to “perform stringent cosmological tests via spectroscopic clustering measurements” and to allow “definitive tests of gravitational physics” by combining redshift-space clustering with weak lensing, CMB, and other southern surveys. The original survey description emphasized BAO, RSD, curvature and consistency tests of the expansion history, high-$^2$6 Ly$^2$7-forest science, and joint tests of growth and geometry with lensing. It stated that CRS is expected to deliver distance and effective expansion-rate measurements at the 1–5% level in redshift bins of width $^2$8 up to $^2$9, and noted that for Ly$^2$0-forest studies 4MOST has almost twice the resolution of DESI (Richard et al., 2019).

Later target-selection forecasting focused on the BG and LRG components. Using the DR10.1-based BG and LRG catalogues, the 2025 selection paper forecasted that 4MOST-CRS alone should deliver about 3% precision on BAO and about 25% on $^2$1, and that combining 4MOST-CRS with DESI improves the constraints from DESI alone by 12% in the $^2$2 redshift range. In that formulation the BG sample provides dense low-shot-noise clustering at $^2$3, while the LRG sample provides a highly biased tracer out to $^2$4 (Verdier et al., 10 Aug 2025).

The survey’s multi-tracer role becomes still sharper in cross-facility forecasts. A 2025 SKA–ESO cosmology study treated a 4MOST CRS LRG sample with $^2$5 over 7500 deg$^2$6, in a bin centred at $^2$7 with $^2$8, and forecasted for the joint multi-tracer combination with SKA-Mid HI intensity mapping

$^2$9

The same study quoted $\approx 5700$0 for the SKA-Mid+4MOST multi-tracer combination (Santos et al., 5 Mar 2025).

These forecasts illustrate two distinct but connected uses of CRS. In its internal logic, it is a self-contained spectroscopic clustering survey. In the wider southern cosmology ecosystem, it is a spectroscopic anchor that converts imaging- and intensity-mapping datasets into joint geometric and growth probes with substantially stronger degeneracy breaking.

5. Survey ecosystem and complementary probes

CRS was designed from the outset as a survey of overlaps. The 2019 survey paper emphasized direct overlap with DES, KiDS, LSST, Euclid, SKA and precursors, eROSITA, and CMB Stage 4, and highlighted cross-correlation redshift calibration as a central use case, with a requirement of at least 1000 deg$\approx 5700$1 per target class for photometric-redshift calibration by cross-correlation. The underlying 4MOST facility design had already identified Euclid photometric-redshift calibration, eROSITA clusters and AGN, and galaxy large-scale structure as core science drivers [(Richard et al., 2019); (Jong et al., 2012)].

Within the broader 4MOST cosmology programme, CRS is complemented by the eROSITA Galaxy Cluster Redshift Survey. That survey was described as one of the core building blocks of the wider 4MOST cosmology programme, and in particular of CRS, providing the cluster- and group-based cosmology component that complements the galaxy-redshift and BAO-focused parts of CRS. It targets $\approx 5700$2 groups and clusters in the German eROSITA sky up to $\approx 5700$3, aims for dynamical and caustic mass measurements for about 10,000 clusters at $\approx 5700$4, and is explicitly framed as the halo-based cosmology anchor of the wider programme (Finoguenov et al., 2019).

Time-domain complementarity is provided by TiDES, the 4MOST Time-Domain Extragalactic Survey. TiDES was designed to classify live transients down to $\approx 5700$5, obtain spectra for up to 30,000 live transients to $\approx 5700$6, measure redshifts for up to 50,000 transient host galaxies to $\approx 5700$7, and monitor around 700 active galactic nuclei to $\approx 5700$8. In the 4MOST cosmology architecture, TiDES adds host-galaxy redshifts for supernova cosmology and AGN reverberation mapping, whereas CRS supplies the static large-scale-structure map (Swann et al., 2019).

A broader implication of this architecture is that CRS is not only a redshift survey in the narrow sense. It is the spectroscopic backbone for a southern multi-probe programme in which galaxy clustering, cluster abundances, weak lensing, SN host spectroscopy, radio intensity mapping, AGN, and CMB cross-correlations are meant to be analyzed as a connected system.

6. Validation, systematics, and present survey characterization

The later CRS literature shifted from top-level design to target-catalogue validation. The BG/LRG selection paper quantified stellar contamination of both galaxy samples at $\approx 5700$9–4% using DESI DR1, showed with angular clustering that both samples are robust against imaging systematics, and reported OpSim-predicted completeness of about 87% for BG and 67% for LRG after 5 years. The same work emphasized that completeness is not spatially uniform because of overlapping 4MOST surveys, so detailed selection functions remain necessary for clustering analyses (Verdier et al., 10 Aug 2025).

The 2025 clustering-validation paper further tested photometry, masking, angular clustering, and redshift distributions of the BG and LRG catalogues. For BG, it adopted Legacy Survey MASKBITS that veto bright stars, SGA large galaxies, and globular clusters, and showed through a Limber-scaling test across $^2$0-band magnitude slices that, after scaling, the $^2$1 curves collapse to a near-common power law over the fitted angular range, with consistency between the North and South Galactic Caps. For LRGs, angular clustering in photo-$^2$2 slices $^2$3 was found to be mutually consistent between the DECaLS and DES footprints at fixed $^2$4, and the final adopted mask combined Legacy Survey masks with the full unWISE W1 artefact mask (Bandi et al., 2 Oct 2025).

That same clustering paper also carried the survey characterization into halo modeling. HOD fits to the LRG projected correlation function yielded satellite fractions of roughly 10%, characteristic halo masses $^2$5–13.0, and linear bias rising from $^2$6 at $^2$7 to $^2$8–2.6 at $^2$9. The authors concluded that the masks and target selections yield uniform clustering statistics, supporting precision large-scale-structure analyses with 4MOST CRS (Bandi et al., 2 Oct 2025).

Taken together, these validation studies indicate that CRS has moved from a consortium-level survey blueprint to a quantitatively specified and tested spectroscopic programme with explicit masking, colour-space selection, angular-clustering diagnostics, HOD-based tracer characterization, and forecasted completeness. In that sense, CRS now occupies a dual status: it remains the southern counterpart to DESI in programme logic, while its practical realization is increasingly defined by DR10.1-based target catalogues, 4MOST operational constraints, and the requirement of precision-ready selection functions for joint cosmological inference.