Local Limit Theorem for Branching Processes in Varying Environments

Abstract

Let $Z_{i,n}$ be the population at time $n$ descended from one particle at time $i$ in a branching process in a varying environment. Under two-sided bounds on products of offspring means, an integrable uniform second-moment tail condition, and nondegeneracy and unit positive-support gcd of all weak limits, we prove $\mathbb{P}(Z_{i,n}=k)\le C(n-i+1)^{-2}$ uniformly in $0\le i<n$ and $k\ge1$, and obtain a local limit theorem as $n-i,k\to\infty$ with $k\le L_0(n-i)$, without third-moment assumptions. These results extend the classical local probability estimates for critical Galton--Watson processes to varying environments.