Random catastrophes, the sudden loss of a significant fraction of a population due to an external event, is one type of noise that has received little attention and yet may dramatically impact population growth patterns and our ability to reconstruct them. In particular, random catastrophes may dampen boom and bust patterns of population growth, leading us to underestimate the underlying dynamics of a population. Here we explore the effect of random catastrophes and other kinds of stochasticity (demographic, environmental) on our ability to reconstruct the dynamics of populations exhibiting an Allee effect and forming part of a metapopulation. In a metapopulation system we can collect data to reconstruct functions by sampling multiple populations at two different time steps. The main parameter of interest is the slope of the growth function as it intercepts the identity line. A value of -1 or less indicates unstable population dynamics. We hypothesized that catastrophes would cause a systematic underestimation of the slope. Our simulation is inspired by social spiders, but should be relevant to other organisms with a metapopulation structure and an Allee effect. We simulate two different population sizes with two different Allee effect sizes subject to five different levels of environmental noise and six levels of catastrophes for a total of 120 different scenarios. We assessed the accuracy and bias of parameter estimations against the known simulation values.
Results/Conclusions
Even small catastrophes caused a strong bias to underestimate slopes. Bias in parameter estimation for small populations was most dependent on the catastrophe size and the amount of environmental noise did not seem important in the weak Allee effect populations. Large populations with weak Allee effects were extremely robust to catastrophes and environmental noise, but large populations with strong Allee effects were the most sensitive to all forms of noise. Populations with a strong Allee effects had the most complex interaction with catastrophes and environmental noise. At moderate levels of noise small population parameter estimations were robust to catastrophes, but as noise increased any catastrophe caused underestimation. Large populations with strong Allee effects were the most sensitive to both noise and catastrophes. In conclusion we show that catastrophes are a type of noise that will systematically interfere with population dynamic reconstruction. A small catastrophe can have far greater impact than even unrealistically high levels of environmental noise. These catastrophes can interact with strong Allee effects leading to extremely poor functional reconstructions.