Pome

23.2.25 Volume equations in Gauduchon gauge29CDAA9CF041

Let $X$ again be a compact LCK manifold of dimension $n\geq 2$. Prescribing the volume of an LCK metric in Gauduchon gauge couples two equations: one for the determinant and one for the conformal factor. To compare metrics with a fixed Lee class, we first identify the twisted groups for cohomologous Lee forms. For a closed real one-form $\lambda $, smooth real functions $a,u$ and a smooth differential form $\eta $, multiplication by $\mathrm{e}^{-a}$ gives

\[ \mathrm{d}_\lambda (\mathrm{e}^{-a}\eta ) =\mathrm{e}^{-a}\mathrm{d}_{\lambda +\mathrm{d}a}\eta , \qquad D_\lambda (\mathrm{e}^{-a}u)=\mathrm{e}^{-a}D_{\lambda +\mathrm{d}a}u. \]

We use this identification when comparing classes; the additive constant in $a$ corresponds to a constant rescaling of the weight bundle.

Let $\omega _1$ be an LCK Gauduchon metric with Lee form $\lambda $, and let $\omega _2$ be an LCK metric with Lee form $\lambda +\mathrm{d}a$. If their twisted classes agree under the chosen identification, then

\[ \mathrm{e}^{-a}\omega _2=\omega _1+D_\lambda u=:A>0 \]

for a smooth real function $u$. Substituting $\omega _2=\mathrm{e}^ aA$ shows that $\omega _2$ has the same volume form as $\omega _1$ and is Gauduchon exactly when

C2B05F8B6DCA 23.2.26
\begin{equation} \begin{cases} A^ n=\mathrm{e}^{-na}\omega _1^ n,\\ \mathrm{dd}^{\mathrm{c}}\bigl (\mathrm{e}^{(n-1)a}A^{n-1}\bigr )=0. \end{cases} \end{equation}

The scalar Hermitian equation of section 14.1 does not impose the second equation, and it uses $\mathrm{dd}^{\mathrm{c}}u$ in place of $D_\lambda u$. The prescribed-volume Gauduchon theorem, theorem 14.10.1, does not assert that its solution is LCK or lies in the given twisted class.

Problem 23.2.27.3B2DB14C0DFF Does a compact LCK manifold have at most one Gauduchon LCK metric with a given Lee class, a given twisted Bott–Chern class and a given smooth positive volume form?

The identification of classes makes this uniqueness problem precise. The question concerns metrics: uniqueness of a scalar potential $u$ would additionally require control of the kernel of $D_\lambda $. The counterexamples with degenerate densities in section 14.20 concern ordinary potentials and do not decide this question.

Cite this page

Comments

Corrections, questions, and mathematical discussion are welcome. Sign in with GitHub to comment. Comments are public.