Metapopulation Allee effects, habitat destruction, and extinction in metacommunities (Q2297277)

The article introduces a model of \(n\) competitor species such that the metapopulation dynamics of the \(i\)-th species (where \(i=1,\ldots,n\)) are described by the differential equations \[ \frac{dp_{i}}{dt}=\frac{c_{i}p_{i}}{\frac{1}{2}(1+e^{-k_{i}(p_{i}-a_{i})})}\left(1-D-\sum_{j=1}^{i}p_{j}\right)-m_{i}p_{i}-p_{i}\sum_{j=1}^{i-1}\frac{c_{j}p_{j}}{\frac{1}{2}(1+e^{-k_{j}(p_{j}-a_{j})})}. \] It is well known that metapopulations models assume that the habitat is fragmented in a number of patches and the focus is on the number of patches occupied instead of the number of individual within patches. In this article the authors consider a degrading habitat with \(n\) species occupying the patches. Moreover, along with the vital dynamics in each patch, each species can leave a patch and colonize another one. In the above context: \begin{itemize} \item \(p_{i}\) is the proportion of patches occupied by the \(i\)-th species, \item \(c_{i}\) is the colonization rate of the \(i\)-th species, \item \(m_{i}\) is the extinction rate of the \(i\)-th species, \item \(k_{i}\geq 0\) and \(a_{i}\in [0,1]\) are parameters that control the intensity and location of a species, \item \(D\) is the proportion of sites destroyed. \end{itemize} Moreover, it is important to emphasize that the species are ranked by its competitive ability from \(i=1\) (best competitor) to \(i=n\) (worst competitor) and that there exists an inverse relationship between competitive ability and colonization rate, that is \(c_{1}\leq c_{2} \leq \cdots \leq c_{n}\) (the best competitor is the worst colonizer and vice versa). The model introduced by the authors generalizes several previous works. In fact, if \(k_{i}=0\) for any \(i=1,\ldots,n\) we obtain the ordinary differential equations: \[ \frac{dp_{i}}{dt}=c_{i}p_{i}\left(1-D-\sum_{j=1}^{i}p_{j}\right)-m_{i}p_{i}-p_{i}\sum_{j=1}^{i-1}c_{j}p_{j}, \] which have been introduced by \textit{D. Tilman} et al. in [``Habitat destruction and the extinction debt'', Nature 371, 65--66 (1994; \url{doi:10.1038/371065a0})]. One of the main conclusions of this model is that ``\dots habitat destruction would lead to the extinction of species in order from the best competitor to the worst, thus leaving only the weakest competitors (but most prolific dispersers) to remain.'' In addition, if \(n=1\), the Tilman's model becomes \[ \frac{dp}{dt}=cp\left(1-D-p\right)-mp, \] and notice that if we assume no habitat destruction, that is \(D=0\) we obtain the seminal model introduced by \textit{R. Levins} \[ \frac{dp}{dt}=cp\left(1-p\right)-mp. \] in [``Some demographic and genetic consequences of environmental heterogeneity for biological control'', Bull. Entomol. Soc. Am. 15, No. 3, 237--240 (1969; \url{doi:10.1093/besa/15.3.237})]. The fundamental difference between the model introduced by the authors compared with Tilman's model is given by the colonization rate: while Tilman's model have a logistic type growth, the authors introduce Allee effects. The section 3 of the article is devoted to study some qualitative properties of the model: firstly, the authors describe the properties of the equilibria and explain the difficulty of obtain an explicit characterization. Secondly, the authors study how the Allee effects can drastically affect the predictions stated by Tilman's model. The study of the differential equations is mainly geometrical and numerical but provides scenarios where Allee effect can reverse the order of extinction compared with Tilman's model.