How spatial heterogeneity affects transient behavior in reaction-diffusion systems for ecological interactions? (Q2008248)

Motivated by [\textit{M. G. Neubert} and \textit{H. F. Caswell}, ``Alternatives to resilience for measuring the responses of ecological systems to perturbations'', Ecology 78, No. 3, 653--665 (1997; \url{doi:10.1890/0012-9658})], who defined the concepts of reactivity and amplification envelope in the context of non-spatial ODEs, the authors introduce a rigorous theoretical framework which is appropriate for the study of possible spatial aspects of transient behavior for reaction-diffusion systems with Neumann boundary conditions, along with operator-theoretic definitions of these concepts for infinite dimensional systems. From a practical viewpoint, this corresponds to predicting the impact of perturbations or control measures upon real-world ecosystems on comparatively shorter time scales, aspect which is not covered by the study of the asymptotic behavior, but is relevant to real-life problems. First, the authors illustrate numerically several possible spatial aspects of transient behavior (amplification and attenuation, respectively) for a showcasing \(2\)-dimensional spatially distributed predator-prey system and provide a comprehensive overview of the state-of-art in the study of transient dynamics. Then, they consider a general model of \(n\) populations which interact locally and diffuse randomly within a habitat, in the form \begin{align*} \frac{\partial \omega_1}{\partial t}&= g_1(\omega_1) + f_1(\omega_1, \omega_2, \ldots , \omega_n) + d_1\Delta \omega_1,\\ \frac{\partial \omega_2}{\partial t}&= g_2(\omega_1) + f_2(\omega_1, \omega_2, \ldots , \omega_n) + d_2\Delta \omega_2,\\ &\vdots\\ \frac{\partial \omega_n}{\partial t}&= g_n(\omega_1) + f_n(\omega_1, \omega_2, \ldots , \omega_n) + d_n\Delta \omega_n. \end{align*} In the above model, \(\omega_i(x,t)\) represents the density of species \(i\), \(1\leq i\leq n\), at spatial location \(x\) and time \(t\), while \(g_i\) and \(f_i\) quantify the effects of intraspecific and interspecific interactions, respectively, and \(d_i\) represents the random diffusion rate of species \(i\). The model is considered on a bounded domain \(\Omega \subset \mathbb{R}^n\), with Neumann boundary conditions for each species, being also assumed that the solutions of the system tend asymptotically to a spatially homogeneous steady state \(\omega ^*\). The authors consider the linearization of the system at the steady state and define the reactivity by means of a Rayleigh quotient for this linearized system. They also show that this definition is independent of the spatial configuration and that, for finite dimensional systems, it agrees with the results of [loc. cit.], since it can be computed as the largest eigenvalue of a real symmetric matrix. An explicit computation is also provided for the case \(n=2\). They also define the amplification envelope of the linearized reaction-diffusion system as being the maximum of the relative amplification over all possible initial perturbations. Unlike the finite-dimensional case, in the infinite dimensional case, one may only obtain an upper bound for the transient. Again, a particularization is provided for the case \(n=2\). The effects of spatial heterogeneity on the transient behavior are illustrated for perturbations with varied degrees of localization, being conjectured that spatially homogeneous perturbations may lead to the largest amplification (that is, spatial heterogeneity may lower the amplitude of perturbations). It is also observed that a larger diffusion coefficient tends to inhibit the growth of perturbations. Possible extensions of the model which deal with different boundary conditions and non-random, directional movements are indicated, along with subsequent caveats.