Advancing ecotones, such as treelines and frontiers of human settlement, may share some characteristic dynamics because both include feedbacks between spatial pattern and process. Both might be examined as complex, self-organizing systems in terms of complexity theory and thus be usefully compared. A cellular automaton of advancing alpine treeline in Montana shows attractors in power-law frequency distributions of spatial and temporal pattern. Frontiers of study areas in the Amazonian region of Ecuador, analyzed using change detection of Landsat Thematic Mapper imagery, have power-law distributions of advancing deforestation. Alternative approaches in self-organized complexity, including self-organized percolation, and the inverse cascade model, and an approach to complexity involving optimization, highly optimized tolerance, are considered. Some combination of these, based on their common ancestry in percolation theory (with its ties to geocomputation), might provide insights into population-environment interactions at settlement frontiers and ecotones together, given comparisons drawn between the spatial feedbacks at alpine treeline and in Ecuador. GIScience and landscape ecology can develop synergies by building on this area of geocomputation and complexity theory, as in analysis of attractors in state spaces of spatial metrics from spatially explicit simulations and representing their uncertainty.