We have developed a knowledge discovery system based on high-order hidden Markov models for analysing spatio-temporal data bases. This system, named \carottage, takes as input an array of discrete data -- the rows represent the spatial sites and the columns the time slots -- and builds a partition together with its \aposteriori probability. \carottage has been developed for studying the cropping patterns of a territory. It uses therefore an agricultural french database, named \teruti, which records every year the land-use category of a set of sites regularly spaced. The results of \carottage are interpreted by agronomists and used in research works linking agricultural land use and water management. Moreover, \carottage can be used to find out and study crop sequencies in large territories, that is a main question for agricultural and environmental research.