The prescribed cross curvature problem on the three-sphere (Q6156069)

Following the seminal work of \textit{J. L. Kazdan} and \textit{F. W. Warner} [Ann. Math. (2) 99, 14--47 (1974; Zbl 0273.53034)] on the admissibility of smooth functions as scalar curvature for closed Riemannian manifolds, the next natural step would be: What can we say about the prescription of Ricci curvature? Prescribing curvature notions is almost always related to understanding the existence of solutions to an overdetermined system of PDEs. The Yamabe problem of understanding constant scalar curvature metrics on conformal classes is a major example of the difficult task of understanding the correct hypotheses aiming to circumvent this overdetermined constraint and critical exponents appearing in this PDE. To the best of this reviewer's knowledge, and as also pointed out in this paper, the best result on the admissibility of obstruction of Ricci tensors on smooth manifolds is local (thanks to the work by \textit{D. M. DeTurck} [Invent. Math. 65, 179--207 (1981; Zbl 0489.53014)]). In this paper, the authors study the prescribing of cross curvature on the 3-dimensional sphere. The paper is complete and also presents the appropriate historic background to this and related problems. The cross curvature prescription problem can be seen as letting \(M\) be a smooth manifold. Once chosen a Riemannian metric \(\mathrm g\) in \(M\), we can always consider the Einstein tensor \(\mathrm{Einstein}(\mathrm g) := \mathrm{Ric}(\mathrm g)-\mathrm{S}(\mathrm g)\frac{\mathrm g}{2}\), where \(\mathrm{Ric}(\mathrm{g})\) stands for the Ricci curvature of \(\mathrm g\) and \(S(\mathrm{g})\) its scalar curvature. The Einstein tensor, via composing it with the \textit{sharp operator} \(\sharp\), defines a \((1,1)\)-tensor field depending on \(\mathrm g\), which the authors denote by \(\mathcal{E}(\mathrm g)\). If \(\mathcal{E}(\mathrm g)\) is invertible, then \(\mathrm{X}(\mathrm g) := \mathrm{det}(\mathcal E(\mathrm g))(\mathcal{E}^{-1})^{\flat}\) is named \textit{cross curvature of \(\mathrm{g}\)}. The prescription of cross curvature is hence equivalent to the problem, given a \((0,2)\) tensor field \(Y\), to find \(\mathrm g\) such that \(X(\mathrm g) = Y\). A conjecture attributed to \textit{R. Hamilton} [Publ., Math. Sci. Res. Inst. 2, 47--72 (1984; Zbl 0557.53018)] asks for the uniqueness of the solution on \(\mathrm{SU}(2)\), further imposing that \(Y\) is positive-definite. The present paper studies the plausibility of this conjecture, obtaining a range of assumptions where it could hold but showing that uniqueness fails in full generality. The authors' results are deeply based on DeTurck's approach to prescribe the Ricci curvature, so it relies heavily on the local nature in some aspects. More precisely, they obtain existence and regularity for the prescribed cross curvature problem in the neighborhood of a point in \(\mathrm{SU}(2)\). They also study the existence of solutions to metrics close (in some adapted Sobolev space norm) to the round metric in the sphere. It is worth pointing out that this is very related to what is known as the Nirenberg problem in prescribing scalar curvature. The paper is well organized and well written, further promoting plenty of references, making it a very good reference for this and related problems.