Reconstructing Our Cosmic Neighborhood
The Universe around us is shaped by an invisible web of dark matter — a vast scaffolding that governs how galaxies form, move, and cluster together. We cannot observe this web directly, but its fingerprints are imprinted on the galaxies we can see. The Manticore Project uses Bayesian inference to work backwards from galaxy redshift surveys, inferring the initial conditions that evolved into the structures we observe today and recovering the hidden dark-matter distribution that underlies them.
Unlike traditional cosmological simulations that begin from random initial conditions, each Manticore reconstruction is constrained by real data. The result is not a single best-guess model but a full ensemble of possible cosmic histories — dozens of independent realisations that together capture the uncertainties in both observation and theory. This allows us to map dark-matter density and velocity fields, identify galaxy clusters and cosmic voids, and assign rigorous statistical significance to every structure we detect.
By comparing these reconstructions directly against observations, Manticore provides one of the most stringent field-level tests of the standard cosmological model. All data products — density fields, velocity fields, cluster and void catalogues — are publicly released to support further research by the wider community.
Bayesian Reconstruction
Field-level inference from galaxy redshift surveys to recover the initial conditions and dark-matter distribution of our local Universe.
Full Posterior Ensembles
Multiple independent realisations that capture uncertainties in dark-matter density fields, velocity fields, and structure properties.
Open Science
All reconstructions, halo and void catalogues, and derived data products are freely available for scientific use.
