Aquatic macrophyte communities have been shown to form sequences along water depth gradients, and water depth has been related to various types of environmental stresses. However, in semi-permanent wetlands, water depth is rarely constant through time; there are large fluctuations in water level between seasons and years, and the relative lengths of wet and dry periods may be as important a stress as the average (or maximum) depth of water. This paper presents a method of quantifying water regimes on the basis of infrequently sampled water level data and of comparing and defining regime-types within and between semi-permanent swamps. Time periods per depth class are calculated from water depth data for each gradient position from the bottom of the swamp to the high-water mark. These data can be represented as a histogram that describes the wetldry pattern of each gradient position. A multivariate clustering is used to compare gradient positions in different swamps on the basis of depth and period of inundation. Water regime types are then defined on the basis of the groups derived in the clustering. The method is illustrated for the Bool Lagoon system in the south-east of South Australia.