The Local Distance Network: a community consensus report on the measurement of the Hubble constant at 1% precision

TL;DR

The paper presents a community-driven framework, the Local Distance Network (DN), to compute the local Hubble constant $H_0$ with 1% precision by covariantly combining multiple distance indicators. It introduces a formal network of anchors, calibrators, and tracers, solved via generalized least squares with a full covariance treatment to exploit redundancy and guard against systematics. The baseline DN returns $H_0 = 73.50 \pm 0.81$ km s$^{-1}$ Mpc$^{-1}$, with strong internal consistency across a wide array of datasets (Cepheids, TRGB, SNe Ia/II, SBF, FP, TF, masers, Gaia parallaxes, etc.), and its robustness is demonstrated through numerous variant analyses. The results remain in tension with early-Universe inferences from Planck/ΛCDM and related constraints, reinforcing the case for new physics or revised cosmology, while providing a transparent, repeatable methodology and open-source tools for future refinements.

Abstract

The direct, empirical determination of the local value of the Hubble constant (H0) has markedly advanced thanks to improved instrumentation, measurement techniques, and distance estimators. However, combining determinations from different estimators is non-trivial, due to correlated calibrations and different analysis methodologies. Using covariance weighting and leveraging the broad and comprehensive community of experts, we constructed a rigorous and transparent Distance Network (DN) to find a consensus value and uncertainty for the local H0. All critically reviewed the available data sets, spanning parallaxes, detached eclipsing binaries, masers, Cepheids, the TRGB, Miras, JAGB stars, SN Ia, Surface Brightness Fluctuations, SN II, the Fundamental Plane, and Tully-Fisher relations and voted for indicators to define a `baseline' DN and others to assess robustness and sensitivity of the results. We provide open-source software and data products to support full transparency and future extensions of this effort. Our conclusions: 1) Local H0 is robustly determined, with first-rank indicators internally consistent within their uncertainties; 2) A covariance-weighted combination yields an uncertainty of 1.1% (baseline) or 0.9% (all estimators); 3) The contribution from SNe Ia is consistent across four current compilations of optical magnitudes or using NIR-only magnitudes; 4) Removing either Cepheids or TRGB has minimal effect; 5) Replacing SNe Ia with galaxy-based indicators changes H0 by less than 0.1 km/s/Mpc, while doubling its uncertainty; 6) The baseline result is H0=73.50+/-0.81 km/s/Mpc. Compared to early Universe results, our result differs by 7.1sigma from flat ΛCDM with Planck+SPT+ACT and 5.0 sigma with BBN+BAO (DESI2). A networked approach is invaluable for enabling further progress in accuracy and precision without overreliance on any single method, sample or group.

Figures (18)

• Figure 1: Conceptual overview of the Local Distance Network, a many routes approach. Different methods for distance determination may connect the absolute scale determined by geometric means to $H_0$. A non-exhaustive list of baseline linkages discussed in the literature or the paper is labeled on the right. Links to geometric distances provided by Masers, DEB, and Parallax are indicated as available in our analysis. Background rectangles in orange, light blue, and gray indicate where Rung 1, Rung 2, and Rung 3 of a traditional distance ladder would fall. Unlabeled tickmarks represent Groups (Fornax & Virgo for the TRGB to SBF, Coma for FP). Example references: 2022ApJ...934L...7R, 2025ApJ...985..203F, 2021AJ....162...80A, 2024ApJ...973...83A, 2020ApJ...890..118P, 2022MNRAS.514.4620D, 2020ApJ...902..145K, 2025MNRAS.539.3627S, 2025AA...702A..41V. Appendix \ref{['app:replication']} replicates a subset of these routes.
• Figure 2: The complete Distance Network, with all possible pathways illustrated. Anchors are objects that establish an absolute scale based on the methods shown to their left. Primary distance indicators (Cepheids, TRGB, Miras, JAGB) transfer the absolute scale to hosts (i.e., galaxies), the ensemble of which calibrates secondary distance indicators in the Hubble flow (tracers). Exceptions are Megamasers and astrophysically modeled SNe II, both of which serve as primary distance indicators and are capable of reaching the Hubble flow without intermediate steps. Green arrows illustrate direct connections between anchors or tracers and the method used to determine the absolute scale. Blue, violet, yellow, and red arrows show which calibrators constrain host distances; line width qualitatively distinguishes the attainable precision. Among hosts, rectangles qualitatively indicate overlap among objects measured via multiple methods. Diamond shapes represent groups. Dark gray arrows tie subsets of hosts whose distance is constrained by different calibrators to tracers. Any given arrow may represent multiple data sets, e.g., HST or JWST photometry of Cepheids, or optical vs. infrared photometry of SNe Ia. The number of hosts is labeled for Cepheids, TRGB, JAGB, and Miras, with the number of hosts exclusively available to each method shown in parentheses.
• Figure 3: The Baseline Distance Network, illustrated analogously to Fig. \ref{['fig:DN_full']}
• Figure 4: Residuals for each category of host distance measurements from the Baseline solution. Each panel represents a group of measurements of host distances that share the same method, anchor, and authors, and shows the deviation of those measured host distances from the full distance network value. Error bars represent the individual uncertainty of each measurement, while the shaded regions for each group shows the common (fully correlated) uncertainty due to the reference system.
• Figure 5: Residuals in distance modulus as a function of redshift for objects in the Hubble flow in the Baseline solution. Error bars reflect the individual source scatter, without the common calibration uncertainty. The shaded region in each plot indicates the effect of a velocity uncertainty of $250 \, \mathrm{km \, s^{-1}}$. The bars at the right show the mean and dispersion for each group of sources and the calibration uncertainty, which is a common error mode for all points in each panel.