no. 15 · Monday 16 November 2026
Central Limit Theorems for Common Zeros via Bergman Kernels
Abstract
We prove central limit theorems as \(N\to\infty\) for common-zero statistics of \(k\) independent standard Gaussian holomorphic sections of \(L^N\), where \(L\) is a positive Hermitian holomorphic line bundle over a compact Kähler manifold of complex dimension \(m\). For fixed real \(\mathscr C^3\) forms \(\varphi\) of bidegree \((m-k,m-k)\) with \(\partial\bar\partial\varphi\not\equiv0\), we prove the CLT in codimensions \(2\leqslant k\leqslant m\) and show that the integrated coefficient in the Shiffman–Zelditch variance formula at scale \(N^{2k-m-2}\) is strictly positive. Shiffman and Zelditch proved the unconditioned smooth CLT for \(k=1\). For \(1\leqslant k\leqslant m\) we also prove a CLT for weighted volumes in fixed domains \(U\) with \(\mathscr C^2\) boundary and fixed real \(\mathscr C^3\) weights \(f\) whose restriction to \(\partial U\) is not identically zero. The variance of a weighted volume is asymptotic to \(N^{2k-m-1/2}ν_{m,k}\int_{\partial U}f^2\,\mathrm dσ\), with \(ν_{m,k}>0\) the Shiffman–Zelditch boundary constant and \(σ\) the induced surface measure. For \(k=m\) and \(f\equiv1\), the statistic counts common zeros in \(U\). Conditioning each section to vanish on the same fixed finite set preserves the leading variance and the Gaussian limit for both classes of tests, including smooth tests for \(k=1\). The set may meet \(\partial U\). A Bergman kernel criterion yields variance bounds and the new unconditioned CLTs from a uniform local \(L^2\) Bargmann–Fock limit and integrable off-diagonal derivative bounds.
- comments
- 67 pages, 1 figure. Substantially revised and retitled. Extends the domain CLT to weighted volumes and identifies their leading variance. Shows that conditioning all sections to vanish on the same fixed finite set preserves the leading variance and Gaussian limit. Adds a Bergman kernel criterion and an explicit lower bound for the integrated smooth variance coefficient
- fields
- Complex Variables, Probability