Population Modeling

Population Viability Analysis has been called the “flagship technology” of conservation biology (by Michael Soule). Kent Holsinger, Eric Menges, and I collaborated together for some years to develop a novel approach to population viability analysis (PVA): we built a hierarchical Bayesian (hB) model that draws upon data from all parts of the life cycle simultaneously to estimate vital rates and their covariation as a function of time-since-fire and random year effects. A unique feature of this model is a year effect that connects five submodels (generalized linear mixed models) into a single large model (with ~1,340 parameters). This work was published in a series of papers in Population Ecology, Ecological Monographs, and Theoretical Population Biology, and is featured as a case study in Mike Dietze’s book “Ecological Forecasting” (see its Ch. 7). Its descendants include my current work aimed at creating demography-based species distribution models, and the use of a Bayesian models to fuse together tree-ring and forest inventory data to analyze forest carbon dynamics.