A cautionary tale: The Goldilocks' Effect in measuring ecological complication in food webs

• The Goldilocks Effect is a warning of too little and too much simplification in measuring ecological complication (EC).
• EC is defined as a set of structural and functional measures for a set of biologically reasonable food webs (BRFW).
• The 'just right' level of EC is the minimum level of variable aggregation and number of links needed to compute BRFW.
• Loop analysis can help identify BRFWs, also termed ecological skeletons, and measure EC.
• Ecologists tend to underestimate the 'just right' level of EC needed for ecosystem understanding and management.

The purpose of this paper is to apply the Goldilocks' Effect to an experimental marine plankton community to explore the ‘just right’ level of ecological complication (EC) using loop analysis. First, the ‘just right’ level of variable aggregation (N) as well as the numbers and types of links need to be determined to then compute a minimal set of alternative descriptions of biologically reasonable food webs for a given ecosystem. A food web is biologically reasonable if it can structurally and functionally exist in nature, and encompass the main causal relationships and feedbacks that are necessary to explain how species densities change over time. An adequate food web model is more complicated that a single static representation. For example, there are several universes or sets of food webs that can be constructed. The macro universe is the set of mathematically possible networks of N variables, but relatively few are biologically reasonable. The micro universe represents all variables and links observed in a set of data fitted loop models and their combinations. It contains trillions of biologically-reasonable networks that can be summarized as Ecological Skeletons. The size of the micro universe is both an overestimate since some networks are unstable and include matrix redundancies, and it is an underestimate because there are inherent limitations on ecosystem observation in identifying all variables and links in a real world food web. The meso universe is the micro universe corrected for these over and under estimations or the ‘just right’ measure of EC. To date, its exact size remains unknown, but loop analysis provides a way to approach a realistic estimation. These results serve as a cautionary tale, however, that EC is more expansive than is commonly understood. The Goldilocks' Effect has been previously used with food webs. Loop analysis, however, contributes new insights to these discussions, and a useful methodology for identifying biologically reasonable food webs, and several associated structural and functional measures of EC.