The understanding and prediction of coastal behaviour over a time span of decades and alongshore stretches of tens of kilometres is important in coastal management. The controlling variables for coastal behaviour on these scales are yet unknown. It is hypothesised that regions exhibiting different large-scale coastal behaviour (LSCB-regions) are controlled by different hydrodynamic forcings, or by different sedimentological or morphological constraints, or by some combination of these factors. The objective of this study is to elaborate a method that effectively quantifies large-scale morphological changes from a large bathymetric data set (large temporal and spatial extent combined with a high resolution), such to define LSCB-regions. The method of analysis consists of a combination of the empirical eigenfunction technique and a moving window approach. The method has been applied on a data set of over 14,000 nearshore profiles. The data set consists of yearly soundings of the nearshore profile along a coastal stretch of 115 km length (the Holland coast). The alongshore profile spacing is generally 250 m. All profiles surveyed in the period 1963-1990 are included in the analysis. Application of the method on the Dutch data set reveals that LSCB-regions indeed can be identified. The sizes of the LSCB-regions are 5 km, 15 km, 20 km, 32 km, and 42 km. Differences in behaviour of the 'secondary' morphology (i.e. morphology other than the mean profile shape, e.g. nearshore bars) are most pronounced. The change in behaviour between LSCB-regions generally occurs over distances as short as about 2 km. The boundaries between the regions are therefore surprisingly sharp. In general, no clear long-term trends in the development of the shape of the sub-aqueous profile have been observed along the Holland coast. Along certain stretches, trends in cross-shore shifting of the sub-aqueous profile occurred. The average shape of the sub-aqueous profile seems to have no relation with the trend in cross-shore shifting of the shoreline (the +1 m-contour). The (sub)decadal shoreline fluctuations, however, are possibly related to the largescale dynamics of multiple bar systems.