Dark Energy Surveys

Dark Energy Surveys

Knowledge Ark · Observing the universe

Tracing
cosmic
accelerationDark energy surveys

An exploding star, a faintly stretched galaxy, a pattern across millions of galaxies: each carries a different clue to how the universe expands. Together, they help us investigate a cause we still do not understand.

DistancesGeometryStructure growth
Four ways to investigate cosmic acceleration Supernova brightness and redshift constrain relative distances. The apparent scale of the baryon acoustic pattern constrains geometry and expansion. Gravitational lensing uses correlated galaxy-shape distortions to study mass and geometry. Cluster counts and calibrated masses constrain the growth of structure. These schematic symbols are a guide to the methods, not survey measurements or a time sequence. FOUR SIGNALS IN THE SKY SupernovaeAcoustic pattern · BAOGravitational lensingGalaxy clusters Brightness + redshift → distancesPattern scale → cosmic geometryShape distortions → mass + geometryCounts + masses → structure growth Schematic guide · No measured data shown
Four complementary observing methods. The BAO symbol represents a statistical separation scale, not a literal ring of galaxies.[1][2]
How did space expand?

Distances and redshifts constrain the universe’s expansion history.

How did matter gather?

Lensing and clusters reveal how cosmic structure developed.

Does one model fit?

The measurements must agree on the same account of the universe.[3]

Reading faint differences

How do you study something you cannot photograph?

A dark energy survey observes stars, galaxies and the light travelling between them. The aim is to find which explanation of cosmic acceleration can account for all those observations.

The methods are strongest together because each measures something different and faces different sources of error. This guide follows those measurements from the telescope to the scientific claim, with survey status and results checked through 8 September 2026.

01

The scientific question

Acceleration is observed. Its cause is still open.

In 1998, two independent supernova teams found evidence that cosmic expansion was accelerating. Distant Type Ia supernovae were fainter than expected in the decelerating models being tested, placing them farther away at a given redshift. The landmark papers appeared in 1998 and 1999.[4][5]

Within general relativity, a component with sufficiently negative pressure can drive acceleration. The simplest candidate is a cosmological constant, Λ, with constant energy density. ΛCDM combines it with cold dark matter and the universe’s other ingredients.[3]

Alternatives include a changing energy component or modified gravity. Surveys measure their proposed consequences; they have not isolated a dark-energy substance.[3]

For the broader physical picture, continue with Dark Energy: Accelerating Expansion.

02

Light as a distance indicator

Supernovae: calibrated flashes across time

Type Ia supernovae are thermonuclear explosions involving white dwarfs. Their peak brightnesses are not identical. Astronomers standardize them using relationships involving the shape of the light curve, its colour and properties of the host galaxy. That makes them useful distance indicators.[7]

Measure the light

Repeated observations record how an explosion brightens and fades in different wavelength bands.

Measure redshift

The stretching of spectral wavelengths supplies the redshift of the supernova or its host galaxy.

Compare distances

Standardized brightnesses and redshifts build a Hubble diagram: a distance–redshift relation.[7]

Distances reflect expansion along the light’s journey, rather than its rate at one instant. The absolute scale also requires calibration: the diagram’s shape can constrain expansion history without independently fixing today’s Hubble constant.[3]

The difficult part is consistency. Dust, calibration offsets, selection effects and changes in the observed supernova population can all shift inferred distances. A sample with many more explosions is only as useful as the understanding of those effects.[7]

03

A pattern inherited from the early universe

BAO: a ruler written into galaxy clustering

Before atoms formed, pressure waves travelled through the hot mixture of ordinary matter and radiation. Their legacy survives as a slight preference for matter to be separated by a particular scale. Today, that statistical pattern can be measured across large galaxy samples. These are baryon acoustic oscillations, or BAO.[2]

Astronomers compare the pattern’s apparent size across the sky and along the redshift direction. The two measurements constrain cosmic distances and the expansion rate relative to the ruler’s scale. It is the collective pattern that matters, not a fixed spacing between every neighbouring pair of galaxies.[8]

This is why a galaxy redshift survey such as DESI belongs at the centre of dark-energy research. Read more in Baryon Acoustic Oscillations and Redshift Surveys and Mapping the Universe.

04

Gravity changes the view

Lensing connects mass with geometry.

Matter along the line of sight bends the paths of light. In weak gravitational lensing, the distortions of individual galaxy images are usually too small to distinguish from their natural shapes. By studying correlated distortions across many galaxies, astronomers recover a statistical signal called cosmic shear.[6]

The signal depends on both the distribution of matter and the distances between observer, lens and source. Dividing source galaxies into redshift groups adds information about different depths along the line of sight. This lets lensing test the growth of structure alongside cosmic geometry.[9]

The measurement challenge

A stretched image has several possible causes.

Telescope optics, detectors and natural alignments between galaxies can imitate or alter the signal. Those effects must be modelled, along with uncertainty in galaxy redshifts.[6]

Strong lensing

One source. Several arrival times.

A strongly lensed variable source can appear in multiple images whose brightness changes arrive at different times. Measured delays, combined with a model of the lens and its environment, constrain distances and the Hubble constant. Assumptions about the lens mass remain important.[10]

Explore Gravitational Lensing: A Natural Cosmic Telescope for the wider phenomenon.

05

The growth of massive structures

Clusters: count them, then weigh them.

Galaxy clusters are enormous gravitationally bound systems containing dark matter, hot gas and galaxies. Their abundance as a function of mass and redshift depends on how matter fluctuations grew, as well as on the cosmic volume being surveyed.[11]

X-ray emission

Hot intracluster gas makes clusters visible at X-ray wavelengths.

The SZ effect

Hot electrons scatter cosmic microwave background photons, changing the radiation’s spectrum.

Galaxy concentrations

Optical and infrared surveys identify groups of galaxies likely to occupy the same system.[11]

Brightness, gas signal and galaxy number are useful clues to mass, but none is a perfect scale. Researchers must calibrate those relationships and model which objects a survey detects or misses. Weak lensing helps estimate cluster masses through their effect on background galaxies.[11]

An example: a 2024 eROSITA cosmology analysis used 5,259 clusters, with weak-lensing information helping calibrate their masses. Its results were consistent with a cosmological constant in the model tested. That tests one part of the cosmological picture; it does not identify the physical nature of dark energy.[12]

06

The observing landscape · September 2026

Different instruments. Complementary views.

A survey’s observing period, public data releases and cosmology papers are separate milestones. Years of calibration and analysis can follow the last observation.

Five major programs and what they contribute
Program Main contribution Status at this update
DES Supernovae, galaxy clustering, lensing and clusters, using imaging from Chile. Observations ran from 2013 to 2019. The final six-year galaxy-clustering and lensing analysis was released in January 2026.[1][13]
DESI Spectroscopic mapping for BAO and the growth of structure. Completed its planned five-year observations in April 2026; observations continue. The first full five-year cosmological findings are expected in 2027.[14]
Euclid Space imaging and spectroscopy for weak lensing and galaxy clustering. Operating in space. Following earlier releases, DR1 Foundation is planned for November 2026; complete DR1 is planned for mid-2027.[15]
Rubin / LSST Repeated wide-field imaging for transient discoveries, galaxy shapes and clustering. The ten-year Legacy Survey of Space and Time began in June 2026.[16]
Roman Planned infrared surveys for lensing, galaxy mapping and supernova distances. Launched on 30 August 2026 and is commissioning. Its science surveys lie ahead.[17][18]

Planned release dates and observing schedules can change. A public image release is not necessarily a release of measurements ready for dark-energy inference.

Why repeated observations matter

A single image records where an object is and how bright it appears. Revisiting the same field reveals changes: a supernova rising and fading, or a source varying over time. Roman’s planned high-latitude time-domain survey uses this repeated-view approach, complementing its wide-area mapping.[18]

The number of detections or alerts is not the number of usable cosmological distance measurements. Sources still need classification, adequate observations and quality checks.[7]

07

The evidence is developing

Is dark energy changing?

Several combinations of galaxy mapping, supernova distances and the cosmic microwave background favour models in which dark energy evolves. The strength of that preference changes with the datasets and analysis. These are significant questions for follow-up, not an established identification of new physics.[20]

The equation of state

w = −1

The parameter w is pressure divided by energy density. A cosmological constant has w = −1. A measurement close to that value is consistent with Λ; it does not prove that Λ is the underlying cause.[22]

Allowing change over time

More freedom changes the question.

A common model uses w(a) = w0 + wa(1 − a), where a is the cosmic scale factor, equal to 1 today. Λ corresponds to w0 = −1 and wa = 0. This is a useful test model, not a complete physical theory.[22]

What changed between the headlines?

  • DESI’s 2025 results: BAO measurements alone were consistent with flat ΛCDM. Combining them with the microwave background and particular supernova compilations produced stronger preferences for evolving dark energy. The chosen combination mattered.[8]
  • Supernova recalibration: a revised DES analysis reduced a widely quoted preference from 4.2 to 3.2 standard deviations for its specified supernova + DESI + microwave-background combination. This is a concrete example of calibration changing the interpretation.[19]
  • Additional DES measurements: a May 2026 analysis combining DES lensing, clustering, BAO and recalibrated supernovae found a modest preference for evolution using DES measurements alone. It remained dependent on the evolving-dark-energy model being tested.[22]
  • New DESI information in July 2026: an analysis of hydrogen absorption in distant quasar spectra added a different geometric constraint. That new measurement lay closer to ΛCDM, while some combined fits still favoured evolution. The evidence does not simply strengthen with every new dataset.[20]

A very recent cross-check: a 4 September 2026 preprint combined Pantheon+ and DES supernova data with consistent treatment of calibration and selection. With BAO and microwave-background information, its authors reported only weak Bayesian preference for evolution.[21]

There is therefore no single context-free answer such as “w is known to one per cent.” Fixing w to be constant, allowing it to evolve, and adding other measurements are different statistical problems. DES’s six-year results illustrate how much the uncertainty changes between a survey-only fit and a specified joint analysis.[13]

08

Turning precision into confidence

What would make the answer convincing?

The strongest result would explain distances and structure growth together, remain stable under reasonable analysis choices, and survive checks with new observations. Each probe brings a different vulnerability into that test.

Where small errors can enter
Method A key uncertainty What researchers check
Supernovae Brightness calibration, dust and selection. Whether different samples give consistent distances after common corrections.[19]
BAO The ruler calibration and interpretation of the observed pattern. Dependence on early-universe assumptions and the cosmological model.[8]
Weak lensing Image distortions, intrinsic alignments and redshifts. Whether the same shear signal survives alternative calibration and modelling choices.[9]
Clusters Mass estimates and survey completeness. Agreement with lensing calibration and a model of the detection process.[12]

Combined evidence must account for shared data.

Two catalogues can contain some of the same supernovae. Lensing and galaxy-clustering measurements can use the same sky and galaxies. Treating those measurements as wholly independent would overstate the information available; joint analyses have to account for shared uncertainties and correlations.[21][13]

A preference is not a probability that a theory is true.

A quoted number of standard deviations describes a comparison made with a particular statistical procedure and model. Bayesian model comparisons ask a related but different question and depend on the range of possibilities allowed beforehand. Neither is a direct percentage chance that dark energy evolves.[19][21]

Strong-lensing distances also connect this research with the Hubble tension, the disagreement over today’s expansion rate. That is a related question, but a hint of changing dark energy does not automatically resolve it.[10]

The sky as an experiment

Every method adds another test.

The discovery of cosmic acceleration began with unexpectedly faint supernovae. The investigation now reaches across galaxy maps, distorted light, massive clusters and the early universe’s surviving patterns.

The next advance may strengthen the simplest explanation or force a more complicated one. What matters is whether the same account keeps working when the universe is measured in another way.

Follow the evidence

Sources & further reading

Original papers, collaboration analyses and official mission updates. Recent preprints are identified; forecasts and release schedules are distinguished from completed measurements.

  1. NOIRLab Astro Data Lab. Dark Energy Survey.Survey methods, observing period and archived imaging.
  2. Eisenstein et al. · 2005. Detection of the Baryon Acoustic Peak in SDSS Luminous Red Galaxies.The acoustic feature as a statistical scale in galaxy clustering.
  3. Weinberg & White · Particle Data Group · 2025. Dark Energy.Expansion, structure growth and the relation between observations and cosmological models.
  4. Riess et al. · 1998. Observational Evidence from Supernovae for an Accelerating Universe and a Cosmological Constant.The High-Z Supernova Search Team’s original acceleration evidence.
  5. Perlmutter et al. · 1999. Measurements of Omega and Lambda from 42 High-Redshift Supernovae.The Supernova Cosmology Project’s independent evidence.
  6. Euclid Consortium · 2023. Mapping the dark Universe with gravitational weak lensing.Correlated distortions, matter mapping and observational contaminants.
  7. DES Collaboration · 2024. The Dark Energy Survey: Cosmology Results with the Full Five-Year Supernova Sample.Standardized distances, sample selection and systematic uncertainties.
  8. DESI Collaboration · 2025. DESI DR2 Results II: BAO Measurements and Cosmological Constraints.Three-year distance measurements and dataset-dependent dark-energy comparisons.
  9. Bartelmann & Schneider · 2001. Weak Gravitational Lensing.The physical and statistical framework for lensing cosmology.
  10. TDCOSMO Collaboration · 2025. Cosmological constraints from strong lensing time delays.Distance inference and the importance of lens-mass modelling.
  11. Allen, Evrard & Mantz · 2011. Cosmological Parameters from Observations of Galaxy Clusters.Cluster detection, mass calibration, selection and cosmological interpretation.
  12. Ghirardini et al. · 2024. The SRG/eROSITA all-sky survey: Cosmology constraints from cluster abundances.The 5,259-cluster analysis and weak-lensing mass calibration.
  13. DES Collaboration · 2026. Dark Energy Survey Year 6: Galaxy Clustering and Weak Lensing Cosmology.Final six-year measurements and model-dependent joint constraints.
  14. Berkeley Lab · 15 April 2026. DESI Completes Planned 3D Map of the Universe and Continues Exploring.Completion of planned observations and the expected full-survey analysis.
  15. Euclid NASA Science Center at IPAC · updated August 2026. Euclid Data Release Timeline.The revised Foundation and complete DR1 schedule.
  16. Rubin Observatory · 30 June 2026. Rubin Observatory begins the Legacy Survey of Space and Time.The start of the ten-year observing program.
  17. NASA · 30 August 2026. Nancy Grace Roman Space Telescope Launches.Launch and commissioning status.
  18. NASA Goddard · 2025. Roman’s Core Surveys.Planned wide-area and time-domain observing strategies.
  19. Popovic et al. / DES Collaboration · 2026. DES-Dovekie: Reanalysis with Updated Supernova Calibration.Calibration changes and their effects on inferred dark-energy evolution.
  20. DESI Collaboration · July 2026 preprint. DESI DR2 Results IV: Alcock–Paczyński Measurements from the Lyman Alpha Forest.Additional geometric information and revised combined constraints.
  21. Camilleri et al. · 4 September 2026 preprint. Supernovae Unite: Combining Pantheon+ and DES-SN5YR.A consistently modelled combined sample and differing statistical comparisons.
  22. DES Collaboration · May 2026 preprint. Constraints on Dynamical Dark Energy from Multiple Probes in the Full Dark Energy Survey.The evolving-equation-of-state model and combined DES measurements.
Continue exploring cosmology
Back to blog