The role of boundary constraints in simulating a nonlocal Gray-Scott model

Abstract: We present second-order algorithms to approximate the solution of a nonlocal Gray-Scott model that is known to generate interesting spatio-temporal structures such as pulse and stripes solutions. Our algorithms rely on a quadrature method for the spatial discretization and the method of lines using a second-order Adams-Bashforth for the time marching. We focus on studying the impact of the type of boundary constraints, e.g. nonlocal Dirichlet/Neumann or local periodic, and the type of nonlocal diffusion, i.e. integral operator with thin- or fat-tailed kernels, on the generation of pulse solutions. Our numerical investigations show that when the spread of the kernel is large, i.e. when the model is nonlocal, both the type of kernels and type of boundary constraints have a strong impact on the solutions profiles.