Seminar:Community Assembly and Disassembly in Stochastic Environments

Thời gian: 10:00 đến 11:00 Ngày 17/07/2026

Địa điểm: Phòng B102, VIASM

Báo cáo viên: Prof.Nguyễn Hải Đăng, Đại học Alabama, Hoa Kỳ

Tóm tắt: Understanding how biodiversity is assembled and lost is a fundamental challenge in ecology. Local communities are continually shaped by species invasions from regional pools, while extinctions may occur gradually or trigger cascades of secondary extinctions. These assembly and disassembly processes are driven by complex species interactions and environmental fluctuations, motivating the need for mathematical frameworks that explicitly incorporate stochasticity. In this talk, I will present recent progress toward a unified theory of stochastic community assembly based on **invasion graphs**. These graphs encode transitions among ecological communities through single- or multiple-species invasions and are defined using invasion growth rates (Lyapunov exponents). I will discuss conjectures and recent results suggesting that invasion graphs characterize the long-term behavior of broad classes of stochastic ecological models, extending classical persistence theory beyond the Hofbauer criterion. For deterministic models, we will prove that there is a unique subgraph of the deterministic invasion graph that accounts for small, random environmental perturbations. This result filters out biologically unrealistic predictions that depend on noise-free dynamics, revealing ecologically meaningful long-term outcomes that are robust to environmental variability. Using the invasion graphs, we introduce two complementary labeled graphs. The community assembly graph characterizes community assembly through single species invasions and reveals when species introductions cause species loss. The community disassembly graph characterizes when species removals trigger secondary extinctions. These methods also identify priority effects (when assembly order determines outcomes) and inverse priority effects (when disassembly order determines outcomes).