Variations in the distribution of principally polarized abelian varieties among isogeny classes (Q2165699)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Variations in the distribution of principally polarized abelian varieties among isogeny classes
scientific article

    Statements

    Variations in the distribution of principally polarized abelian varieties among isogeny classes (English)
    0 references
    0 references
    22 August 2022
    0 references
    A challenge: over a finite field, estimate the number of (simple, ordinary) principally polarized abelian varieties \(A\) in a given isogeny class. This is a little too naive: one should instead look at a stratum inside an isogeny class in which the endomorphism ring of \(A\) is fixed. This paper gives an answer for a significant class of rings. The most immediately applicable version of the result is this: fix \(\dim A=n\) and for each \(m>0\) fix a finite field \({\mathbb F}_{q_m}\), such that \(q_m\to \infty\), and an isogeny class \({\mathcal C}_m\) with Frobenius angles \(\theta_{m,i}\). Then (under mild conditions) the number \(P_m\) of principally polarized abelian varieties in the stratum of \({\mathcal C}_m\) corresponding to the ring generated by Frobenius and Verschiebung is very approximately \[ q_m^{n(n+1)/4}\prod_{i<j}(\cos \theta_{m,i} - \cos \theta_{m,j})\prod_i\sin \theta_{m,i}. \] The meaning of ``very approximately'' is made precise in the paper. It means in particular that the trigonometric terms are unimportant if the \(\theta_{m,i}\) differ from one another and from \(0\) and \(\pi\) by some fixed positive amount. However this need not be the case, and the paper gives examples of families of surfaces where \(P_m\) is not even very approximately \(q_m^{3/2}\) but rather some lower power of \(q_m\). Combined with a result of \textit{S. G. Vlăduţ} [Mosc. Math. J. 1, No. 1, 125--139 (2001; Zbl 0987.11042)] this formula also gives a heuristic explanation of a result of \textit{N. M. Katz} and \textit{P. Sarnak} [Random matrices, Frobenius eigenvalues, and monodromy. Providence, RI: American Mathematical Society (1999; Zbl 0958.11004)] on the limiting distribution of the Frobenius angles, but there seems to be no way of constructing a proof along these lines. The most general result here is less explicit because it involves a ratio of class numbers associated with the ring, but the rings for which it is applicable are described precisely.
    0 references
    abelian variety
    0 references
    Frobenius eigenvalue
    0 references
    distribution
    0 references
    isogeny
    0 references
    complex multiplication
    0 references
    Katz-Sarnak
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references

    Identifiers

    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references