Abstract:Feng-Yun-Zhang have proved a function field analogue of the arithmetic Siegel-Weil formula, relating special cycles on moduli spaces of unitary shtukas to higher derivatives of Eisenstein series. We prove an extension of this formula in a low-dimensional case, and deduce from it a Gross-Zagier style formula expressing intersection multiplicities of cycles in terms of higher derivatives of base-change $L$-functions.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is the problem of the intersection formula between special cycles on the moduli space, especially the relationship between these intersection numbers and the higher - order derivatives of certain L - functions. Specifically, the authors focus on the special cycle classes defined on the unit shtuka moduli space and attempt to prove that the intersection numbers of these special cycles can be expressed by the higher - order derivatives of certain L - functions. This can be regarded as a generalization of the Gross - Zagier - type formula, that is, a way to establish a connection between arithmetic geometry and automorphic forms.
### Background of the paper
1. **Special cycle classes**: Feng, Yun, and Zhang defined special cycle classes on the unit shtuka moduli space and proved an arithmetic Siegel - Weil formula, which links these special cycles to the higher - order derivatives of the Fourier coefficients of Siegel Eisenstein series.
2. **Gross - Zagier - type formula**: Usually, one hopes to use this connection to prove a Gross - Zagier - type formula, that is, to link the intersection numbers of special cycles to the higher - order derivatives of Langlands L - functions. However, since these moduli spaces are almost always non - complete, this makes the problem very difficult.
### Goals of the paper
The main goal of the paper is to prove a Gross - Zagier - type formula in the low - dimensional case by bypassing the problem of non - completeness of the moduli space. Specifically, the authors conducted research in the following situations:
- **Low - dimensional case**: When \(n = 2m\) and \(m = 1\), that is, consider the special cycles on the \(U(2)\) moduli space.
- **Completeness assumption**: Assume that certain special cycles are complete, so that the intersection theory in the Chow group can be used.
### Main results
1. **Theorem 1.2.2**: For a rank - 1 vector bundle \(E_2\) and a Hermitian isomorphism \(a_2: E_2\cong\sigma^*E_2^{\vee}\), the author proves that there exists a \(K_1\)-invariant automorphic form \(D\), whose geometric Fourier coefficients can be expressed as:
\[
D(E_1, a_1)=\chi(\det(E_1))q^{d(E_1)}\deg([Z_r^{E_1}(a_1)]\cdot[Z_r^{E_2}(a_2)])
\]
where \(E_1\) is an arbitrary rank - 1 vector bundle, and \(a_1: E_1\rightarrow\sigma^*E_1^{\vee}\) is an arbitrary Hermitian map.
2. **Theorem 1.3.3**: In the case where \(E_2\) is a line bundle and \(a_2: E_2\cong\sigma^*E_2^{\vee}\) is a Hermitian isomorphism, the author proves the Gross - Zagier - type intersection formula:
\[
\deg(\vartheta_r^{\chi}(f)\cdot[Z_r^{E_2}(a_2)]) = f(E_2, - a_2)q^{m d(E_2)}(\log q)^r\frac{d^r}{ds^r}\left(q^{ns\deg(\omega_X)}L(s + 1/2, BC(\pi)\otimes\chi)\right)\bigg|_{s = 0}
\]
where \(f\) is a \(K_m\)-invariant automorphic form, and \(\vartheta_r^{\chi}(f)\) is the arithmetic theta lift.
### Conclusion
Through these results, the authors not only prove the validity of the Gross - Zagier - type formula in the low - dimensional case, but also provide a theoretical basis for further generalizing this formula. These results are of great significance in the study of arithmetic geometry and automorphic forms.