This paper is aimed at showing how cellular automata can be convenient
ly employed to simulate dynamic phenomena, typically involving transpo
rtation, diffusion, or propagation problems. A cellular automaton can
be viewed as made of two parts: a computational engine based on a prop
er discretization of the domain and charged with correctness and consi
stency controls, and a dynamic model constituted by transition functio
ns that express cell behaviour. The adoption of cellular automata intr
oduces a new means of spatial data modelling, in addition to those tra
ditionally provided by GIS packages, resulting in the possibility of s
toring elements of dynamic knowledge in cellular maps: each cell is pr
ovided with the attributes that constitute its state, and groups of ce
lls with the functions that describe their mutual interaction. The bas
ic characteristics of cellular automata are discussed with reference t
o a significant application case, the study of tide propagation over a
lagoon.