Accounting for size-specific predation improves our ability to predict the strength of a trophic cascade

Abstract

Predation can influence the magnitude of herbivory that grazers exert on primary producers by altering both grazer abundance and their per capita consumption rates via changes in behavior, density-dependent effects, and size. Therefore, models based solely on changes in abundance may miss key components of grazing pressure. We estimated shifts in grazing pressure associated with changes in the abundance and per capita consumption rates of sea urchins triggered by size-selective predation by sea otters (Enhydra lutris). Field surveys suggest that sea otters dramatically decreased the abundance and median size of sea urchins. Furthermore, laboratory experiments revealed that kelp consumption by sea urchins varied nonlinearly as a function of urchin size such that consumption rates increased to the 0.56 and 0.68 power of biomass for red and green urchins, respectively. This reveals that shifts in urchin size structure due to size-selective predation by sea otters alter sea urchin per capita grazing rates. Comparison of two quantitative models estimating total consumptive capacity revealed that a model incorporating shifts in urchin abundance while neglecting urchin size structure overestimated grazing pressure compared to a model that incorporated size. Consequently, incorporating shifts in urchin size better predicted field estimates of kelp abundance compared to equivalent models based on urchin abundance alone. We provide strong evidence that incorporating size-specific parameters increases our ability to describe and predict trophic interactions.

Keywords: Biomass; body size; herbivory; size‐selective predation; trophic cascade.

Figure 1

Site (n = 20) locations where surveys were conducted on the central coast of BC, Canada. Numbers represent otter occupation time (in years) of each site.

Figure 2

Change in (A) red and (B) green urchin density and (C&D) size with increasing otter occupation time. The solid lines represent parsimonious models with greatest strength of evidence. The dotted lines represent alternative candidate models, as there was not strong support for one over the other. No line represents that there was not strong support for the alternative models over the intercept (null) model. Note differences in y‐axis scale among graphs.

Figure 3

Red (A–C) and green (D–F) urchin test diameter size frequency distribution separated by low (0–2 years), intermediate (3–8 years), and high (17–33 years) categories of otter occupation time. Dotted line represents median size of red and green urchins within each sea otter occupation time category and is noted in graph legend. Bin size = 10. Note differences in y‐axis scale among graph panels.

Figure 4

(A) Red urchin and (B) green urchin size‐specific grazing rates. (Grazing_Red = 0.02 × Biomass_Red ^0.56). (Grazing_Green = 0.01 × Biomass_Green ^0.68). Solid line represents power functions fit to the data. Gray area represents 95% confidence intervals.

Figure 5

Estimated total consumptive capacity (TCC) for urchin populations with no impact on size distributions versus estimated total consumptive capacity for urchin populations resulting from size‐selective predation. Only sites with otters present were compared (n = 13).

Figure 6

Comparative models of kelp density as a function of (A) urchin density (B) urchin biomass, (C) urchin TCC, and (D) urchin metabolic biomass. Solid line represents the exponential models fit to each set of data. Darker shading of data points indicates higher frequency of overlapping points. N = 358 1 m × 1 m quadrats.