Title: Distribution of primes of split reductions for abelian surfaces

Abstract: Let $A$ be an absolutely simple abelian surface defined over a number field $K$ with a commutative (geometric) endomorphism ring. Let $\pi_{A, \text{split}}(x)$ denote the number of primes $\mathfrak{p}$ in $K$ such that each prime has absolute norm bounded by $x$, of good reduction for $A$, and the reduction of $A$ at $\mathfrak{p}$ splits. It is known that the density of such primes is zero. Under a Generalized Riemann Hypothesis (GRH) for Dedekind zeta functions and certain Galois image assumptions, we prove that $\pi_{A, \text{split}}(x) \ll_{A, K} x^{\frac{41}{42}}\log x$ if the endomorphism ring of $A$ is trivial; $\pi_{A, \text{split}}(x) \ll_{A, E, K} \frac{x^{\frac{11}{12}}}{(\log x)^{\frac{2}{3}}}$ if $A$ has real multiplication by $E$; $\pi_{A, \text{split}}(x) \ll_{A, E, K} x^{\frac{2}{3}}(\log x)^{\frac{1}{3}}$ if $A$ has complex multiplication by $E$. These results improve the bounds by D. Zywina in 2008 and J. Achter in 2012. Under other credible conjectures, we improve the first two bounds to $x^{\frac{5}{6}}\log x$ and $x^{\frac{2}{3}}\log x$, respectively. As an application, we also obtain upper bounds for the number of extremal primes for abelian surfaces.