On the equivalence of distributional and synthetic Ricci curvature lower bounds
Andrea Mondino, Vanessa Ryborz
TL;DR
The paper proves an equivalence between distributional and synthetic lower bounds for Bakry–Émery Ricci curvature on weighted manifolds with a continuous metric and low-regularity connections. It develops a comprehensive non-smooth calculus, including a weak Bochner formula, and establishes a robust first- and second-order framework that aligns distributional Ricci bounds with the RCD/CD notions via BE inequalities. The main contributions are (i) showing that $\mathsf{RCD}^*(K,N)$ (or $\mathsf{RCD}(K,\infty)$) implies a distributional $\mathrm{Ric}_{\mu,N}\ge K g$ and (ii) proving the reverse implication under a volume-growth condition, thereby unifying distributional and synthetic curvature notions on spaces with low regularity. This bridges classical geometric analysis with modern metric-measure theory, enabling curvature-based methods on spaces with minimal smoothness and weighted measures.
Abstract
The goal of the paper is to prove the equivalence of distributional and synthetic Ricci curvature lower bounds for a weighted Riemannian manifold with continuous metric tensor having Christoffel symbols in $L^2_{\rm loc}$, and with weight in $C^0\cap W^{1,2}_{\rm loc}$.
