COS 22-4 - A distributional approach to modelling the population dynamics of forest trees

Tuesday, August 9, 2016: 9:00 AM
Palm A, Ft Lauderdale Convention Center
Jessica F. Needham, Department of Plant Sciences, University of Oxford, Oxford, United Kingdom, Cory Merow, Ecology and Evolutionary Biology, University of Connecticut, Storrs, CT and Sean M. McMahon, Smithsonian Environmental Research Center, Smithsonian Institution Global Earth Observatory, Edgewater, MD

In order to accurately describe the population dynamics of slow-growing, long-lived species, such as trees, it is essential that demographic models capture individual differences that persist through time.

In forests, the individuals which make it to the canopy are those that survived and grew consistently fast throughout their lives. These fast-growers define the structure, biomass and many of the ecosystem functions of the forest and their reproductive output contributes disproportionately to population growth rates. Failure to accurately model the dynamics of successful individuals results in skewed estimates of population statistics such as population growth rates.

Integral projection models (IPMs) are a flexible tool that describe how the vital rates of individuals within a population interact to define population level dynamics.

We construct IPMs that explicitly account for temporal autocorrelation in individual vital rates through grouping of individuals into distinct growth classes. Our size-growth-class IPMs describe the transitions of individuals between sizes and growth classes, allowing us to better capture the observed variation in individual pathways through the life cycle.  

Data describing the full life history of tree species is often limited or unavailable. We present methods to inverse model vital rates parameters using population level summary statistics calculated from IPMs.


We present results for three species using data from a 50 ha forest plot on Barro Colorado Island (BCI), Panama. As expected the two example canopy species have much faster growth rates and also more differentiation between growth classes than our example understory species. Survival probabilities were also highest in canopy species.

Population growth rates for each species were calculated from size-growth-class IPMs and were found to be relatively insensitive to fecundity parameters estimated from inverse models. In contrast, population-level statistics such as population growth rate, life expectancies and passage times to threshold sizes (e.g. size at first reproduction) were much more sensitive to the dimensions of the IPMs and the transition probabilities between growth classes. We make recommendations for the number of growth distributions and IPM dimensions to use according to the life history strategy of the focal species.