ScalingStacks

7.2. The skeletal measure [017W]

Original official author HTML, exact retained edition. Historical TeX conversion verdicts remain unchanged. Cited-edition alignment and mathematical self-containment are not assessed.

Complete original source context · Original author HTML

7.2. The skeletal measure

Consider the skeletal measure μlog⁡|η|#\mu_{\log|\eta|^{\#}} on Sk⁡(X)\operatorname{Sk}(X). Choose an snc model 𝒳{\mathcal{X}}, and write as usual 𝒳0=∑i∈Ibi​Ei{\mathcal{X}}_{0}=\sum_{i\in I}b_{i}E_{i}. The form η\eta defines an identification K𝒳/Slog=∑i∈Iai​EiK^{\mathrm{log}}_{{\mathcal{X}}/S}=\sum_{i\in I}a_{i}E_{i}, and Proposition 5.10 yields

κmin=mini⁡aibi.\kappa_{\min}=\min_{i}\frac{a_{i}}{b_{i}}. (7.1)

If κmin∈ℤ\kappa_{\min}\in{\mathbb{Z}}, then

ω:=d​ttκmin+1∧η\omega:=\frac{d{t}}{{t}^{\kappa_{\min}+1}}\wedge\eta

is a logarithmic form on 𝒳{\mathcal{X}}. For each face σ\sigma of Δ⁡(ℒ)\Delta({\mathcal{L}}), ordering the set J⊂IJ\subset I of components cutting out the stratum Y=YσY=Y_{\sigma} yields a well-defined Poincaré residue ResY⁡(ω)\operatorname{Res}_{Y}(\omega). By Lemma 5.12, ResY⁡(ω)\operatorname{Res}_{Y}(\omega) is a rational section of KYK_{Y}, with divisor

−BYℒ=∑i∉J(ai−κmin​bi−1)​Ei|Y-B^{\mathcal{L}}_{Y}=\sum_{i\notin J}(a_{i}-\kappa_{\min}b_{i}-1)E_{i}|_{Y}

When σ\sigma is a maximal face, ResY⁡(ω)\operatorname{Res}_{Y}(\omega) is thus a holomorphic form on YY; using the formulas in §6.2, it is easy to see that the residual measure on YY is given by

ResY⁡(log⁡|η|#)=|ResY⁡(ω)|2.\operatorname{Res}_{Y}(\log|\eta|^{\#})=|\operatorname{Res}_{Y}(\omega)|^{2}.

The following result corresponds to Theorem C in the introduction.

Theorem 7.1.

Assume that XX is maximally degenerate, i.e. dimSk⁡(X)=n\dim\operatorname{Sk}(X)=n. If XX has semistable reduction, then the skeletal measure μlog⁡|η|#\mu_{\log|\eta|^{\#}} is a multiple of the integral Lebesgue measure of Sk⁡(X)\operatorname{Sk}(X).

Proof.

Let 𝒳{\mathcal{X}} be a semistable model, i.e. 𝒳{\mathcal{X}} is snc with 𝒳0{\mathcal{X}}_{0} reduced. By (7.1), we have κmin∈ℤ\kappa_{\min}\in{\mathbb{Z}}. Since some non-empty EJE_{J} might have several components, the dual complex Δ⁡(𝒳)\Delta({\mathcal{X}}) is possibly not a triangulation of Sk⁡(X)\operatorname{Sk}(X). However, the barycentric subdivision Δ′\Delta^{\prime} of Δ⁡(𝒳)\Delta({\mathcal{X}}) is a triangulation; the corresponding toroidal modification 𝒳′{\mathcal{X}}^{\prime} is snc, with 𝒳0′{\mathcal{X}}^{\prime}_{0} is possibly non-reduced, but bσ=1b_{\sigma}=1 for each nn-simplex σ\sigma of Δ′\Delta^{\prime}. Applying the above discussion to 𝒳′{\mathcal{X}}^{\prime}, we infer

μlog⁡|η|#=∑σ|Resyσ⁡(ω)|2​λσ,\mu_{\log|\eta|^{\#}}=\sum_{\sigma}|\operatorname{Res}_{y_{\sigma}}(\omega)|^{2}\lambda_{\sigma},

with σ\sigma ranging over the nn-dimensional faces of Δ′\Delta^{\prime}, with corresponding strata yσ∈𝒳0′y_{\sigma}\in{\mathcal{X}}^{\prime}_{0} reduced to single points. It will thus be enough to show that |Resyσ⁡(ω)||\operatorname{Res}_{y_{\sigma}}(\omega)| is independent of σ\sigma.

By the strong connectedness property, any two nn-simplices σ\sigma, σ′\sigma^{\prime} of Δ′\Delta^{\prime} can be joined by a chain of nn-simplices σ=σ1,…,σN=σ′\sigma=\sigma_{1},\dots,\sigma_{N}=\sigma^{\prime} with σi\sigma_{i} and σi+1\sigma_{i+1} sharing a common (n−1)(n-1)-face τi\tau_{i}. Denoting by yi=yσiy_{i}=y_{\sigma_{i}} and Yi=YτiY_{i}=Y_{\tau_{i}} the corresponding strata in 𝒳′{\mathcal{X}}^{\prime}, we thus have yi,yi+1∈Yiy_{i},y_{i+1}\in Y_{i}. Further, the Poincaré residue ResYi⁡(ω)\operatorname{Res}_{Y_{i}}(\omega) has poles precisely at yi,yi+1y_{i},y_{i+1}, since any other pole would correspond to an nn-simplex of Δ′\Delta^{\prime} containing τi\tau_{i}, contradicting the non-branching property. Since Resyi⁡ResY⁡(ω)=Resyi⁡(ω)\operatorname{Res}_{y_{i}}\operatorname{Res}_{Y}(\omega)=\operatorname{Res}_{y_{i}}(\omega), the residue theorem applied to the Riemann surface YiY_{i} yields Resyi⁡(ω)+Resyi+1⁡(ω)=0\operatorname{Res}_{y_{i}}(\omega)+\operatorname{Res}_{y_{i+1}}(\omega)=0, and hence |Resy1⁡(ω)|=⋯=|ResyN⁡(ω)||\operatorname{Res}_{y_{1}}(\omega)|=\dots=|\operatorname{Res}_{y_{N}}(\omega)|. ∎

Remark 7.2.

Theorem 7.1 fails in general when XX does not have semistable reduction. Indeed, the semistable reduction theorem [KKMS] shows that the base change p:X′→Xp\colon X^{\prime}\to X to ℂ⁡((t1/m)){\mathbb{C}}(\!({t}^{1/m})\!) has semistable reduction for some mm divisible enough. By Lemma 5.14, dimSk⁡(X′)=n\dim\operatorname{Sk}(X^{\prime})=n, and μlog⁡|η′|#\mu_{\log|\eta^{\prime}|^{\#}} is thus a multiple of the integral Lebesgue measure λ′\lambda^{\prime} of Sk⁡(X′)\operatorname{Sk}(X^{\prime}), by Theorem 7.1. By Theorem 6.7, μlog⁡|η|#=m−n​p∗​λ′\mu_{\log|\eta|^{\#}}=m^{-n}p_{*}\lambda^{\prime}. However, p∗​λ′p_{*}\lambda^{\prime} is not proportional to the integral Lebesgue measure λ\lambda of Sk⁡(X)\operatorname{Sk}(X) in general. Indeed, for each nn-simplex σ\sigma of Δ⁡(ℒ)\Delta({\mathcal{L}}), Lemma 5.13 shows that (p∗​λ′)σ=mn​bσ​λσ(p_{*}\lambda^{\prime})_{\sigma}=m^{n}b_{\sigma}\lambda_{\sigma}, and bσb_{\sigma} is in general not independent of σ\sigma.

Original mathematics by the credited authors. Source-backed reader collection; mathematical self-containment is not assessed.