Vital processes (i.e., survival, growth and reproduction) of many species vary between its individuals because of differences in attributes such as size or age. To account for these differences, integral projection models (IPMs) describe each vital rate as a function of these attributes and then integrate them into a single model. However, individuals do not display vital processes independently, but jointly, and can exhibit trade-offs between these processes that, ultimately, may limit the set of demographic behaviors a population can display. To evaluate how trade-offs could influence demography, we took the population dynamics of disparate species from the COMPADRE/COMADRE databases, and simulated demographic data following these dynamics, modifying the degree of trade-off between the vital rates, as well as other demographic aspects: variation of the trade-off within and between individuals, and individual time-series length. We applied both the traditional independent modeling approach and ours, and compared estimated vital-rate functions and projected population trends.
Results/Conclusions
As expected, assuming independence between vital rates that actually exhibit trade-offs affects the results derived from IPMs, both at the vital-rate, and population structure and growth level. In the latter case, failure to include trade-offs became more significant when these existed between processes having large elasticity values. However, detectability of the existence of trade-offs plays an important role on results: relatively long individual time-series can be required to estimate individual-level trade-offs, particularly when large within-individual variation in such trade-offs exist; in turn, large between-individual variation in trade-offs reduces their effect on population trends. IPMs are becoming important modeling tools to study the dynamics of populations and project their growth over time. Not accounting for the existence of trade-offs in IPMs can translate into proposed population growth trajectories that, due to these trade-offs, would not be displayed; fortunately, including them does not require further fieldwork, as these can be estimated from existing survey data.