Interactions between blight resistance and seedling demography limit American chestnut population growth in an integral projection model

Background/Question/Methods

Emerging and invasive pathogens represent a growing threat to global biodiversity and constitute a critical challenge to land and resource management. In addition, because eradication of naturalized pathogens is often impossible, efforts to conserve remnant host stocks and restore self-sustaining host populations following an epidemic frequently occur within a novel ecological regime that includes the pathogen.  Attempts to restore damaged populations and ecosystem services therefore require an integrated approach that accounts for interactions between disease dynamics and the other ecological factors that influence host demography. 

We use integral projection model (IPM) for naturally regenerating stands of American chestnut (Castanea dentata) in New England to explore the dynamics and potential long-term success of hypothetical populations that have different levels of phenotypic resistance to chestnut blight (Cryphonectria parasitica).  Specifically, we hypothesize that the threshold resistance levels at which populations will collapse or become self-sustaining depend on growth and especially on seedling survival, which may be limited by herbivory.   

Vital rates for our IPM were estimated using a continuous-time state-space model.  We compare the results obtained by this approach with traditional, regression-based estimation methods.

Results/Conclusions

In comparison to methods that estimate growth or survival on a fixed interval, vital rate estimation using a continuous-time model allows IPMs to draw on a wider variety of data sources.  Individuals do not need to be surveyed on a regular basis, and the model can easily incorporate age measurements where they are available.  In principle, this means that an IPM could be fit using data from a single cross-sectional snapshot in time.  Embedding the demographic process models within a state space model also has the advantage that it separates measurement error from biological stochasticity, potentially yielding more precise projections.

Preliminary analyses indicate that in the absence of significant blight, chestnut populations in our study sites are expanding with a long-term equilibrium growth rate of 8% (λ = 1.08).  Annual survival is size-dependent; above-ground mortality inversely related to stem diameter, and populations are most sensitive to survival and growth in small trees. Depending on resistance and herbivory, blight-induced mortality can cause model populations to go extinct, or can lead to periodic cycles of regeneration and dieback.  To maximize their chances of success, breeding programs and other efforts to introduce disease resistance need to consider ecological circumstances when identifying phenotypic resistance targets.