Spatially lumped models may fail to take into account the effects of spatial heterogeneity and local interactions. These properties sometimes are crucial to the dynamics and evolutions of ecosystems. This paper started from the fundamental aspects of CA and focused on the development and application of the approach to ecological and ecohydraulics modelling. Application cases include modelling of prey–predator dynamics by stochastic CA and simulation of riparian vegetation successions in a regulated river by rule-based CA. The results indicated that spatially explicit paradigms such as cellular automata (CA) have a strong capability to bridge the local processes and global patterns.