Strong competition with refuges

DOI10.1016/S0362-546X(96)00160-5MaRDI QIDQ4378771zbMATH OpenOpenAlexFDO

Publication date 30 March 1999

Published in Nonlinear Analysis: Theory, Methods & Applications (Search for Journal in Brave)

Full work available at URL https://doi.org/10.1016/s0362-546x(96)00160-5

Nonlinear elliptic equations (35J60) General theory of partial differential operators (47F05) Applications of operator theory in chemistry and life sciences (47N60) Ecology (92D40)

Review text

Here, the problem of permanence of \(N \geq 2\) competing species in the case when their interactions are very strong is analyzed. For this analysis the following model is used: \[ (1.1)\qquad \partial_tu_k+{\mathcal L}_ku_k=\lambda_ku_k-c_{kk}(x)u^2_k-\sum_{j=1,j\neq k}^N s_{kj}c_{kj}(x)u_ku_j\quad\text{in\quad\(\Omega\times(0,\infty)\)}, \] \[ {\mathcal B}_k u_k=0\quad\text{on\quad \(\partial\Omega\times(0,\infty)\)},\quad u_k(x,0)=u_{0k}(x)\quad \text{in\quad \(\Omega\)},\qquad\text{for \(1 \leq k \leq N\)}, \] \[ \; \] where \(\Omega\) is a bounded domain of \({\mathbb R}^d, d \geq 2\), with boundary \(\partial\Omega\) of class \(C^{2+\nu}\) for some \(\nu\in(0,1)\) and \({\mathcal L}_k, 1 \leq k \leq N\), are second order uniformly elliptic operators \[ {\mathcal L}_k:= - \sum_{i,j=1}^d\alpha_{ij}^k(x)D_iD_j+\sum_{i=1}^d\alpha_i^k(x)D_i+\alpha_0^k(x),\qquad 1\leq k \leq N, \] with coefficients in the Banach space of Hölder continuous functions \(V:=C^\nu(\overline\Omega)\). The \({\alpha_k}\)'s and \({s_{kj}}\)'s are assumed to be real numbers such that \(s_{kj}>0\) for all k, j and the \({c_{kj}(x)}\)'s to be non-negative functions of V such that the sets \[ (1.2)\quad\{x\in\Omega : c_{kk}(x)=0\},\quad 1 \leq k \leq N,\quad\text{are finite}. \] Each of the boundary operators \({\mathcal B}_k\) is assumed to be of any of the standard types; namely, either Neumann, Dirichlet, Robin, or mixed type. Condition (1.2) ensures that each population grows accordingly to a logistic law in the absence of the remaining populations. The functions \(c_{kj}(x)\), \(j \neq k\), describe the subdomains of \(\Omega\) where interactions of \(u_j\) on \(u_k\) take place. The species \(u_k\) will be free from the action of \(u_j\) if \(c_{kj}\) vanishes, and it will have to compete against \(u_j\) in the subdomains of \(\Omega\) where \(c_{kj}>0\). So, the \({c_{kj}}\)'s can be regarded as a sort of ``maps of operations, or maps of interactions. The parameters \(s_{kj}>0\) measure the struggle of the species \(u_j\) on \(u_k\). The main result reads as follows: THEOREM 1.1. Let \({\mathcal B}_{k0}\) denote the boundary operator on \(\partial {\mathcal R}_k\) defined by \({\mathcal B}\) on \(\partial\Omega\), and by \({\mathcal D}\) that on \(\partial {\mathcal R}_k\setminus\partial\Omega\), where \({\mathcal D}\) stands for the Dirichlet operator. Then: (i) If the following inequalities are satisfied \[ (1.3)\qquad \lambda_k\geq\sigma_1^{{\mathcal R}_k}[{\mathcal L}_k;{\mathcal B}_{k0}],\quad 1 \leq k \leq N, \] then (1.1) is permanent. (ii) If, instead of (1.3), it is assumed that there exist \(k \in \{1,...,N\}\) for which \[ (1.4)\qquad \lambda_k < \sigma_1^{{\mathcal R}_k}[{\mathcal L}_k;{\mathcal B}_{k0}],\quad \lambda_j \geq \sigma_1^{{\mathcal R}_j}[{\mathcal L}_j;{\mathcal B}_{j0}],\quad j \neq k, \] then for each \(u_0=(u_{01},...,u_{0N}) \in V^N\) with \(u_{j0} > 0\), \(\;1 \leq j \leq N\), there exist \({s_{kj}^*=s_{kj}(u_0)>0}\), ine \( \;j \neq k\), such that \[ (1.5)\qquad \lim_{t\to\infty} u_k(\cdot,t;u_0)=0, \] uniformly in \(\overline \Omega\), provided \(s_{kj} \geq s_{kj}^*,\;j \neq k\). In other words, the species \(u_k\) is driven to extinction by the \(u_j,\;j \neq k\). Theorem 1.1 (i) establishes that if the birth rate of each of the species is sufficiently large, then the species are permanent, and this independently on how strong are their interactions. Moreover, how large each \(\lambda_k\) has to be is measured by means of \(\sigma_1^{{\mathcal R}_k}[{\mathcal L}_k;{\mathcal B}_{k0}].\) Theorem 1.1 (ii) establishes the optimality of condition (1.3) in the sense that if some \(\lambda_k\) does not reach the value \(\sigma_1^{{\mathcal R}_k}[{\mathcal L}_k;{\mathcal B}_{k0}]\), while the remaining species do, then \(u_k\) is driven to extinction by the remaining species if the strenght of each \(u_j,\;j \neq k\), on \(u_k\) is sufficiently large. The organization of the paper is as follows. Section 2 analyzes the evolution of each of the species in the absence of the remaining ones. The results are well known in the case when the \({s_{kk}}\)'s are positive and bounded away from zero, but are new when some of these coefficients vanish, since in this case the usual way of finding supersolutions does not work. Further the concept of permanence used in this work is established and a sufficient condition for it is given. Although this sufficient condition is more or less folklore, a short proof of it is given, as it will be used to prove the main theorems. The asymptotic behavior of the principal eigenvalues of a general class of second order uniformly elliptic operators as the amplitude of a potential vanishing on a subdomain of \(\Omega\) grows up to infinity is established. This result is the key to complete the proof of Theorem 1.1. First the asymptotic behavior of the positive solutions of a general class of logistic boundary value problems is analyzed as the amplitude of a potential vanishing on a subdomain of \(\Omega\) grows up to infinity. Then, this result is used to give the distribution of each of the species when the interactions of the remaining ones on it are very strong. It is shown how the population concentrates on its corresponding refuge while it is driven to extinction outside.

Recommendations

Cites work

Cited in

(11)

This page was built for publication: Strong competition with refuges

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4378771)