Population viability analysis (PVA) has been shown to be a reliable tool for ranking management options for species despite parameter uncertainty. No studies have yet investigated whether this holds true for model uncertainty for long-lived species with complex life histories and for multiple threats. We ask whether a range of model structures give similar rankings of management and/or threat scenarios for two long-lived plant species: an obligate seeding shrub and a facultative resprouting shrub. These are each exposed to altered fire regimes and an additional, species-specific threat: high post-fire predation rates or a fatal root pathogen. Long-term demographic datasets are used to construct an individual-based model (IBM), a complex stage-based model, and a simple matrix model that subsumes all life stages into 2 or 3 stages. A range of threat and management scenarios is systematically applied to the three model types for both species. Pearson product-moment and Spearman rank correlation coefficients between model structures were calculated for rankings of scenarios, based on extinction risk or final median abundance (FMA).
Results/Conclusions
For the obligate seeder, when fire and predation management scenarios were considered altogether strong positive correlations across the three model structures were observed. This reflects the dramatic and consistent effect of predation management on extinction risks across the three model types. The 2-stage and complex matrix models had the highest correlations, followed by the 2-stage and individual-based models, and finally the complex matrix and IBM. As the predation rate decreased so did the correlation coefficients for comparisons involving the IBM. For the facultative resprouter, correlation coefficients were high when all scenarios were considered together, reflecting the strong effect of disease and very frequent fire on FMA. Of these, the 3-stage and complex matrix models compared most favorably. When fire management was considered under the “no disease” scenario, correlation coefficients were positive and high. As disease increased, correlations became weaker, with the greatest number of negative correlations observed under the high disease scenario. For both species, correlations decline as the scenarios deviate from baseline conditions, likely the result of a number of factors related to life history complexity and how it is represented in models. While PVA can be an invaluable tool for integrating data and understanding species’ responses to threats and management strategies, this is best achieved in the context of decision support for adaptive management alongside multiple lines of evidence and expert critique of model construction and output.