The volume measure of the Brownian sphere is a Hausdorff measure (Q2082674)

The Brownian sphere, also known as the Brownian map, is a random compact metric space providing a universal model of two-dimensional random geometry. The Brownian sphere has been shown to be the scaling limit in the Gromov-Hausdorff sense of many different classes of random planar maps such as triangulations or quadrangulations of the sphere, see [\textit{L. Addario-Berry} and \textit{M. Albenque}, Ann. Henri Lebesgue 4, 653--683 (2021; Zbl 1483.05138)] and [\textit{C. Marzouk}, Ann. Henri Lebesgue 5, 317--386 (2022; Zbl 1487.05074)]. The first construction the Brownian sphere relied on Brownian motion indexed by the Brownian tree, or equivalently on the Brownian snake driven by a normalized Brownian excursion [the author, Invent. Math. 169, No. 3, 621--670 (2007; Zbl 1132.60013)]. In this construction, the Brownian sphere is constructed as a quotient space of the interval \([0, 1]\) and is naturally equipped with a volume measure defined as the pushforward of Lebesgue measure on \([0, 1]\) under the canonical projection. This volume measure appears as the limit of counting measures on vertices when the Brownian sphere is written as the Gromov-Hausdorff-Prokhorov limit of large random planar maps, for the case of quadrangulations see [the author, Ann. Inst. Henri Poincaré, Probab. Stat. 55, No. 1, 237--313 (2019; Zbl 1466.60021), Theorem 7] and [Marzouk, loc. cit., Theorem 1.2] for the more general case of bipartite planar maps with a prescribed degree sequence. The Brownian sphere is known to be homeomorphic to the \(2\)-sphere, but its Hausdorff dimension is equal to 4, see [the author, Zbl 1466.60021]. The original motivation of the present work was to determine an exact Hausdorff measure function for the Brownian sphere. The main result of the paper solves this problem, and also provides a natural interpretation of the volume measure. Denote the Brownian sphere by \(m_{\infty}\), and write \(\mathrm{Vol}\) for the volume measure on \(m_{\infty}\). For any gauge function \(h\), denote by \(m_{\infty}\) the associated Hausdorff measure. Theorem. For every \(r\in (0, 1/4)\), set \(h(r) = r^4\, \log\log(1/r)\). There exists a constant \(\kappa>0\) such that we have almost surely for every Borel subset \(A\) of \(m_{\infty}\): \(m^h(A) = \kappa \mathrm{Vol}(A)\). This theorem shows that the volume measure \(\mathrm{Vol}\) is completely determined by the metric on \(m_{\infty}\). Although this result seems plausible, it was not obvious from the construction of the Brownian sphere in terms of Brownian motion indexed by the Brownian tree. Note that other models of random geometry such as the Brownian plane and the Brownian disk have been investigated in recent papers, see [the author, Zbl 1466.60021] and also correspond to specific quantum surfaces, see [\textit{J. Miller} and \textit{S. Sheffield}, Ann. Probab. 49, No. 6, 2732--2829 (2021; Zbl 1478.60045), Corollary 1.5]. The paper is organized as follows. Section 2 contains a number of preliminaries. The author recalls the construction of the Brownian sphere, and the spine decomposition of the Brownian snake conditioned on its minimum. The author also states the comparison lemma for Hausdorff measures that plays a central role in the proofs. Section 3 proves estimates on moments of the volume of balls centered at \(x_*\), from which it is relatively easy to derive the lower bound \(m^h(A)\geq k_1 \mathrm{Vol}(A)\). The proof of the corresponding upper bound is given in Section 4. Finally, the zero-one law argument needed to establish the main result is presented in Section 5.