This work examines the proposition that positive interactions among neighboring individuals within a population may produce landscape patterns in boundary intensity. The large scale patterns emerge because the interactions favor an aggregated distribution in the face of a potential limiting factor, and the strength of that factor varies over the landscape. The consequences of spatially varying neighborhood processes were explored using cellular automata simulating the structure of mussel beds in 2-dimensional intertidal landscapes, each characterized by a vertical gradient of tidal immersion and a horizontal gradient of wave energy. Running the model with and without the neighborhood processes demonstrated that the facilitating neighborhood processes elevate intensity above that caused by the gradients, and consequently abrupt (high intensity) boundaries emerged in the midst of gradual environmental variation. Trends generated on the 2-D landscape by the model were compared with those in photo-mosaics of intertidal mussel beds, Mytilus californianus on rocky shores of the British Columbia. The analysis involved interpolation of boundary locations using a spatially-constrained cluster algorithm, and then estimation of the corresponding boundary intensities using a landscape index aggregation (CLUMPY). The general similarity between predicted and real trends in intensity over the wave energy gradients suggests that spatially varying neighborhood processes determine much of the landscape scale variation in boundary intensity, while certain discrepancies (e.g. a more rapid rise of observed intensities with increasing wave exposure) suggest modifications of the theory and new empirical work.