Speaker
Description
Biodiversity varies in response to natural and anthropogenic drivers. Yet, most analytical tools for quantifying driver impact remain correlational, irrespective of the sophistication of the underlying statistical methodology. Unfortunately, at present, few patterns in biodiversity science have been subject to robust causal analyses, leaving even well-studied empirical patterns victim to opaque interpretations. Here we develop a causal modelling framework for one of biodiversity’s oldest questions: what generates the island species-area relationship (ISAR) in island biogeography? We encode the causal structure of island assembly in a directed acyclic graph (DAG), and build a phylogenetically-informed generative model that explicitly separates two scales: lineage colonization of archipelagos, and within-archipelago species’ occurrence on individual islands. Stage 1 models colonization of 32 archipelagos for terrestrial birds as a function of isolation, area, and age; stage 2 models species’ occurrence on islands as a function of island area. The ISAR emerges as a derived relationship defined by the causal question and data generating process, not from a regression applied to emergent phenomena. The two-stage architecture reveals how passive sampling operates across scales. At the archipelago scale, larger source pools increase colonization probability, determining which lineages enter the regional pool available to colonize islands. At the island scale, the size of the filtered pool sets the expected richness of a reference island (ISAR intercept), and area governs how richness scales with island area from that baseline, generating the ISAR slope. Our framework results in a causally derived ISAR slope lower than many classical estimates that conflate these processes, a discrepancy with direct implications for how species-area theory is applied to habitat loss and fragmentation. By anchoring inference to an explicit DAG, the framework clarifies what can be identified from observational data, and points toward the causal infrastructure needed to estimate intervention effects for biodiversity science.
| Status Group | Postdoctoral Researcher |
|---|---|
| FOR TALKS: Poster Presentation Option | No, I prefer to present only as a talk. |