A reaction-advection-diffusion system modeling the competition for two complementary resources with seasonality in a flowing habitat (Q2347129)

This paper is concerned with a one-dimensional mathematical model for the competition of two species \(u_1\) and \(u_2\) for two growth-limiting nutrients \(R\) and \(P\) in a periodic bio-reactor (for instance, a chemostat) of length \(L\): \[ \partial_t R=\delta \partial_{xx}R-\nu\partial_x R-q_{1r}h_1(R,P)u_1-q_{2r}h_2(R,P)u_2, \] \[ \partial_t P=\delta \partial_{xx}P-\nu\partial_x P-q_{1p}h_1(R,P)u_1-q_{2p}h_2(R,P)u_2, \] \[ \partial_t u_1 = \delta \partial_{xx}u_1-\nu \partial_x u_1+h_1(R,P)u_1, \] \[ \partial_t u_2 = \delta \partial_{xx}u_2-\nu \partial_x u_2+h_2(R,P)u_2, \] together with nonnegative initial conditions and boundary conditions at \(x\in \{0,L\}\): \[ \nu R(0,t)-\delta \partial_x R(0,t)=\nu R^{(0)}(t),~\partial_x R(L,t)=0, \] \[ \nu P(0,t)-\delta \partial_x P(0,t)=\nu P^{(0)}(t),~\partial_x P(L,t)=0, \] \[ \nu u_i(0,t)-\delta \partial_x u_i(0,t)=\partial_x u_i(L,t)=0. \] Above, \(\delta>0\) denotes the diffusivity and \(\nu>0\) is the rate of the flow through the chemostat. The influx of nutrients \(R^{(0)}(t)\) and \(P^{(0)}(t)\) at \(x=0\) is assumed to be periodic over time. In this model, the response functions \(h_i(R,P)\) are defined by \[ h_i(R,P)=\min(f_{ir}(R),f_{ip}(P))\qquad (f_i(0)=0,~f_i'>0), \] so population growth only occurs if both types of nutrient are present. In Section 2 of the paper, the model above with a single population \(u\) present only is investigated. The following results are obtained: The quantities \(R(x,t)+q_r u(x,t)\) and \(P(x,t)+q_p u(x,t)\) approach periodic functions \({R}^*(x,t)\) and \({P}^*(x,t)\) as \(t\to\infty\), respectively, uniformly in \(x\in[0,L]\). Denote by \(\mu^*\) the principal eigenvalue of the system replacing \(R\) and \(P\) by \({R}^*\) and \({P}^*\), respectively. If \(\mu^*>0\), the population will be washed out: \(u(x,t)\to 0\) as \(t\to\infty\), uniformly in \(x\). If \(\mu^*<0\), the population persists as there exists a periodic solution \((\hat{R},\hat{P},\hat{u})\) which is approached by \((R,P,u)\) as \(t\to\infty\). Section 3 is concerned with the full model. The authors prove that, if the parameters allow for persistence in both single-species models, one also has persistence in the full model: There exists a positive periodic solution and \(\liminf\limits_{t\to\infty} u_i(x,t)\) is \(x\)-uniformly positive. Convergence to the periodic solution is not proven for this system. Numerical simulations are also given, see Section 4 of the paper.