Developing a spatial framework involves the systematic identification and labelling of distinctive spatial units and the mapping of their geographic distributions. There is no perfect method for developing spatial frameworks. Different spatial frameworks possess different properties and the application of the framework defines which of these properties are important. Thus, it is important to understand the way in which a spatial framework has been developed.
Spatial frameworks can be broadly subdivided into those based on classification and those based on regionalisation. For classifications, the systemic guiding principles establish a strict procedure for measuring difference and similarity between spatial units and for defining class structure. Classifications can be distinguished by the characteristics (or variables) that are used to determine difference and similarity between spatial units and the procedure that is used to develop a structure of classes. Classifications may be defined using biological or environmental characteristics of the spatial units being classified. Importantly, in classification the geographic location of spatial units is not generally taken into account by the definition procedure. Classes are generally defined by grouping spatial units that are similar with respect to their biological or environmental attributes. The proximity of spatial units is therefore measured in a multi-dimensional space (sensu Austin and Smith 1989) where each dimension is represented by one of the biological or environmental characteristics. Because the classes are defined independently of geographic location, the spatial framework generally shows a mosaic of patches of similar biological or environmental characteristics that recur across the classified area.
Regionalisation differs from classification in that the geographic location, as well as similarity in environmental and/or biotic character, is used to define a structure of distinctive spatial units. Regionalisation and classification have much in common. However, regions are generally singular geographic units that cover a contiguous area. The definition procedure for regionalisations is carried out in geographic space (sensu Austin and Smith 1989) by experts who use maps of biotic and/or environmental attributes. Subjective judgement is used to delineate regions within which there is a certain degree of homogeneity with respect to the defining attributes (deBlij 1978; Wicken 1986). In the past, regionalisation has been a common approach to developing spatial frameworks (e.g. Knox 1995). However, classifications that are defined using quantitative analyses are becoming increasingly feasible with the continued growth in computing power, and the wider availability of spatially explicit descriptions of both the environment and biota. For reasons that are discussed later, quantitative classifications have a variety of advantages over regionalisations and, in New Zealand at least, this is becoming the more common method for defining spatial frameworks.