Scale—We All Talk About It; What Do We Do With It?

Abstract

Scale is a fundamental concept in landscape ecology. We characterize scale by grain and extent and apply it to space, time and classification. When conducting experiments, it is important to be mindful of what scale to sample at; the “correct” scale depends on the research question and the organism(s) of interest. Frequently, patterns seen at one scale will be different – even opposite – at another scale, which introduces difficulties with extrapolating results from one scale to another. In this chapter, I discuss how scale can affect experimental design. In particular, I focus on how scaling up from smaller manipulative experiments to larger landscapes is confounded by three challenges. The first is coarse-graining, which refers to techniques to aggregate fine-scale data without risking the introduction of statistical bias. Second, the middle-number problem describes complex landscape systems with too many variables to model easily but too few to make averaging a meaningful solution. Finally, there is the issue of non-stationarity. Non-stationary processes are those where the means, variance and co-variances change over time – often in unpredictable, stochastic ways. I offer suggestions to overcome these challenges, including adoption of the hierarchical patch dynamics paradigm, the use of lacunarity metrics or examination of different “currencies”, such as energy or elements, which are amenable to cross-scale analysis.

What is an ‘appropriate’ scale depends in part on the questions one asks.

–John Wiens. (From Wiens JA (Wiens, 1989) Spatial scaling in ecology. Funct Ecol 3(4): 385–397)