The sandflat habitat: Scaling from experiments to conclusions
Ecological systems are characterised by spatial and temporal variations in the density of organisms and resources, and in the intensity of processes which affect them (e.g. Watt, 1947, Dayton, 1971, Addicot et al., 1987, Kolasa and Pickett, 1991 and Giller et al., 1994). This heterogeneity represents both a difficulty for field study design and statistical testing, and a challenge to describe the spatial structuring of populations, communities and ecosystems (Legendre, 1993). Patterns in ecological and environmental variables are fundamental to developing hypotheses and in designing subsequent field studies, because they determine the spatial and temporal scales of study. Spatial and/or temporal patterns also provide useful stepping stones for addressing issues of scale. Patterns apparent at one scale can collapse to noise when viewed from other scales, indicating that patterns, processes and our perceptions vary in a scale-dependent manner. Perhaps the most important but rarely discussed issue is the question of how to scale-up from small-scale surveys and experiments to conclusions relevant at larger spatial and temporal scales.