Activator-inhibitor systems of reaction-diffusion equations have been used to describe pattern formation in numerous applications in biology, chemistry, and physics. The rate of diffusion in these applications is manifest in the single parameter of the diffusion constant, and stationary Turing patterns occur above a critical value of $d$ representing the ratio of the diffusion constants of the inhibitor to the activator. Here we consider activator-inhibitor systems in which the diffusion is anomalous subdiffusion; the diffusion rates are manifest in both a diffusion constant and a diffusion exponent. A consideration of this problem in terms of continuous-time random walks with sources and sinks leads to a reaction-diffusion system with fractional order temporal derivatives operating on the spatial Laplacian. We have carried out an algebraic stability analysis of the homogeneous steady-state solution in fractional activator-inhibitor systems, with Gierer-Meinhardt reaction kinetics and with Brusselator reaction kinetics. For each class of reaction kinetics we identify a Turing instability bifurcation curve in the two-dimensional diffusion parameter space. The critical value of $d$, for Turing instabilities, decreases monotonically with the anomalous diffusion exponent between unity (standard diffusion) and zero (extreme subdiffusion). We have also carried out numerical simulations of the governing fractional activator-inhibitor equations and we show that the Turing instability precipitates the formation of complex spatiotemporal patterns. If the diffusion of the activator and inhibitor have the same anomalous scaling properties, then the surface profiles of these patterns for values of $d$ slightly above the critical value varies from smooth stationary patterns to increasingly rough and nonstationary patterns as the anomalous diffusion exponent varies from unity towards zero. If the diffusion of the activator is anomalous subdiffusion but the diffusion of the inhibitor is standard diffusion, we find stable stationary Turing patterns for values of $d$ well below the threshold values for pattern formation in standard activator-inhibitor systems.

• Received 31 January 2005

DOI:https://doi.org/10.1103/PhysRevE.72.026101