A systolic inequality for geodesic flows on the two-sphere (Q514363): Difference between revisions

The authors show that for a sufficiently pinched Riemannian metric \(g\) on \(S^2\) the following inequalities hold: \[ l_{\min}^2(g)\leq\pi\text{Area}(S^2,g)\leq l_{\max}^2(g) \] where \(l_{\min}\) and \(l_{\text{max}}\) denote, respectively, the lengths of the shortest and longest simple closed geodesics. It is also shown that the inequalities become equalities if and only if the metric is Zoll. As a corollary the authors solve positively the conjecture of Babenko and \textit{F. Balacheff} [Geom. Dedicata 121, 61--71 (2006; Zbl 1109.53044)] which says that the systolic ratio of the round sphere is a local (non-strict) maximum in the space of Riemannian metrics on the 2-sphere. The paper is well written and is highly recommended to those researchers interested in differential geometry (including every branch: Riemannian, contact, symplectic,\dots). It is almost self-contained and it also includes original proofs of several known but non-trivial results. The paper is organized as follows: In the first section, the authors explain the main results and the proof strategy. In the second section, they present a class of diffeomorphisms \(\Phi:S\to S\), called \(\mathcal{D}_L(S,\omega)\), defined in the infinite closed strip \(S=\mathbb{R}\times[0,\pi]\) so that \(\Phi(x+L,y)=(L,0)+\Phi(x,y)\), \(\Phi\) preserves the boundaries and preserves the form \(\omega=\sin y\, dx\wedge dy\). These diffeomorphisms can be defined at the level of the annulus \(\mathbb{R}/L\mathbb{Z}\times[0,\pi]\). For these diffeomorphisms the flux and the Calabi invariant [\textit{E. Calabi}, in: Probl. Analysis. Sympos. in Honor of Salomon Bochner, Princeton Univ. 1969, 1--26 (1970; Zbl 0209.25801)] (for the case of zero flux) can be defined. The authors restrict the study to diffeomorphisms in \(\mathcal{D}_L(S,\omega)\) which are \textit{monotone} (Definition 2.8). In this restricted class the important Theorem 2.12 holds: Let \(\Phi\) be a monotone and zero-flux element of \(\mathcal{D}_L(S,\omega)\) with negative (resp. positive) Calabi invariant then \(\Phi\) has an interior fixed point with negative (resp. positive) action. In the third section the authors work with a sufficiently pinched metric on the sphere and the Birkhoff annuli relative to the shortest (and longest) simple closed geodesics, and they check (with full detail) that, under good coordinates, the first return map to these annuli can be lifted to a monotone, zero-flux diffeomorphism in \(\mathcal{D}_L(S,\omega)\) (Theorem 3.8). The other key point is the relation between the area and the Calabi invariant of this lift, more precisely: \[ \pi\,\text{Area}(S^2,g)=L^2+L\cdot\text{CAL}(\Phi), \] where \(L\) is the length of the simple closed geodesic used to define the Birkhoff annuli. From these ingredients the main theorems follow readily. Two valuable appendices are included in the paper. Appendix A includes a proof of a version of Topogonov's theorem in order to estimate the perimeter of a convex geodesic polygon on positively curved spheres. Appendix B deals with the geodesic flow on Zoll metrics and their symplectic relation with the geodesic flow on the round sphere.