Floodplains can be viewed as complex adaptive systems (Levin, 1998) because they are comprised of many different biophysical components, such as morphological features, soil groups and vegetation communities as well as being sites of key biogeochemical processing (Stanford et al., 2005). Interactions and feedbacks among the biophysical components often result in additional phenomena occuring over a range of scales, often in the absence of any controlling factors (sensu Hallet, 1990). This emergence of new biophysical features and rates of processing can lead to alternative stable states which feed back into floodplain adaptive cycles (cf. Hughes, 1997; Stanford et al., 2005). Interactions between different biophysical components, feedbacks, self emergence and scale are all key properties of complex adaptive systems (Levin, 1998; Phillips, 2003; Murray et al., 2014) and therefore will influence the manner in which we study and view spatial patterns. Measuring the spatial patterns of floodplain biophysical components is a prerequisite to examining and understanding these ecosystems as complex adaptive systems. Elucidating relationships between pattern and process, which are intrinsically linked within floodplains (Ward et al., 2002), is dependent upon an understanding of spatial pattern. This knowledge can help river scientists determine the major drivers, controllers and responses of floodplain structure and function, as well as the consequences of altering those drivers and controllers (Hughes and Cass, 1997; Whited et al., 2007). Interactions and feedbacks between physical, chemical and biological components of floodplain ecosystems create and maintain a structurally diverse and dynamic template (Stanford et al., 2005). This template influences subsequent interactions between components that consequently affect system trajectories within floodplains (sensu Bak et al., 1988). Constructing and evaluating models used to predict floodplain ecosystem responses to natural and anthropogenic disturbances therefore require quantification of spatial pattern (Asselman and Middelkoop, 1995; Walling and He, 1998). Quantifying these patterns also provides insights into the spatial and temporal domains of structuring processes as well as enabling the detection of self-emergent phenomena, environmental constraints or anthropogenic interference (Turner et al., 1990; Holling, 1992; De Jager and Rohweder, 2012). Thus, quantifying spatial pattern is an important building block on which to examine floodplains as complex adaptive systems (Levin, 1998). Approaches to measuring spatial pattern in floodplains must be cognisant of scale, self-emergent phenomena, spatial organisation, and location. Fundamental problems may arise when patterns observed at a site or transect scale are scaled-up to infer processes and patterns over entire floodplain surfaces (Wiens, 2002; Thorp et al., 2008). Likewise, patterns observed over the entire spatial extent of a landscape can mask important variation and detail at finer scales (Riitters et al., 2002). Indeed, different patterns often emerge at different scales (Turner et al., 1990) because of hierarchical structuring processes (O'Neill et al., 1991). Categorising data into discrete, homogeneous and predefined spatial units at a particular scale (e.g. polygons) creates issues and errors associated with scale and subjective classification (McGarigal et al., 2009; Cushman et al., 2010). These include, loss of information within classified ‘patches’, as well as the ability to detect the emergence of new features that do not fit the original classification scheme. Many of these issues arise because floodplains are highly heterogeneous and have complex spatial organizations (Carbonneau et al., 2012; Legleiter, 2013). As a result, the scale and location at which measurements are made can influence the observed spatial patterns; and patterns may not be scale independent or applicable in different geomorp