Scalar curvature and \(Q\)-curvature of random metrics (Q5891887)

Let \((M,g_0)\) be a compact Riemannian manifold with nonvanishing scalar curvature, and consider the problem of determining the probability that a conformal change \(e^f g_0\) makes the scalar curvature change sign. This paper analyzes this interesting question (as well as other similar questions in ``random'' conformal geometry), using a Gaussian probability measure on conformal classes. Let \(g_1(a)=e^{a f} g_0\) be a random metric on a surface \((M,g_0)\), where \(f: M\to \mathbb R\) is a random function conveniently written as a power series in terms of eigenfunctions of the Laplacian of \(g_0\), with coefficients that are Gaussian random variables. The authors prove certain lower and upper bounds (in terms of \(a\)) for the probability that the scalar curvature of \(g_1(a)\) changes sign somewhere in the surface \(M\). In particular, they estimate the rate at which this probability goes to zero as \(a\to 0\). From a point of view of Global Analysis, this provides some information on the local geometry of the space of Riemannian metrics on \(M\) near \(g_0\), concerning the relative volume of the metrics whose scalar curvature changes sign. In even dimensions \(>2\), the authors establish analogous estimates for the probability that a conformal change makes the \(Q\)-curvature of a metric change sign. The paper under review has a large overlap with a previously published paper, with the same title, by the same authors [Electron. Res. Announc. Math. Sci. 17, 43--56 (2010; Zbl 1208.53051)], where the ``results were announced''.