Let $(M^n,g)$ be a closed (compact, without boundary) Riemannian manifold and let $p\in M$. Let $$ \square: =c\Delta+S $$ be the conformal Laplacian of $(M,g)$. Here $c=4\frac{n-1}{n-2}$ is a constant, $\Delta$ is the Laplacian and $S$ is the scalar curvature defined by $g$. Lastly, let $G_p:M\setminus \{p\}\to \mathbb{R}$ be the Green function of $\square$, so $\square\, G_p=0$ everywhere on $M\setminus \{p\}$. See here for a discussion of the conformal Laplacian $\square$ and its Green function $G_p$.

If $(M,g)$ has positive scalar curvature, then the generalized stereographic projection of $(M,g)$ is defined to be the Riemannian manifold $(\hat{M},\hat{g})$, where $$ \hat{M}=M\setminus \{p\} \qquad \text{and} \qquad \hat{g}=G_p^{k-2}g, $$ and $k=2n/(n-2)$ is a constant. This is well defined when $(M,g)$ has positive scalar curvature, since in this case the Green function $G_p$ is strictly positive.

Question 1: If $n\geq 4$, is the Riemannian metric $\hat{g}$ on $\hat{M}$ complete?

I suspect this is the case as Lee and Parker prove that metric $\hat{g}$ is asymptotically flat of order at least 2.

My second question is related to the curvature properties of $(\hat{M},\hat{g})$. Note that $(\hat{M},\hat{g})$ has zero scalar curvature, as follows from the fact that $\square\, G_p=0$.

Question 2: How is the Ricci curvature of $\hat{g}$ related to the Ricci curvature of $g$? More specifically, if $g$ has Ricci curvature bounded above or below, does this give bounds on the Ricci curvature of $\hat{g}$?

  • $\begingroup$ A naive question : is classical stereographic projection from the sphere minus one of its pole to a plane really a particular case of what you explain ? $\endgroup$ – Jean Marie Jan 12 at 22:40
  • 1
    $\begingroup$ You can establish an asymptotic estimate for $G_p$ near $p,$ which in particular lets you bound it from below by $r^k$ for some constant $k(n).$ (You should definitely be able to find this in Lee-Parker.) You can then verify completeness of the metric by using this bound to show that the length of any curve approaching $p$ is infinite. $\endgroup$ – Anthony Carapetis Jan 13 at 7:08
  • $\begingroup$ @JeanMarie: this is explained very nicely in Section 6 of Lee and Parker's paper (see hyperlink in my question). $\endgroup$ – rpf Jan 16 at 15:30

Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Browse other questions tagged or ask your own question.