Estimates and tests of per capita interaction strength: Diet, abundance, and impact of intertidally foraging birds

Abstract

Predicting the dynamics of natural food webs requires estimates of the strength of interactions among species. The ability to estimate per capita interaction strength from observational data is desirable because of the logistical difficulty of using experimental manipulations to obtain such measures for all species within complex natural communities. In this paper, I derive observational measures of per capita interaction strength having units matching those of dynamic food web models (per capita consumption and assimilation rates). I also highlight the difference between per capita interaction strength (a parameter used in theoretical models) and species impact (empirical measures of total species effect). I then use behavioral observations and population censuses in a rocky intertidal community to estimate both per capita interaction strengths and species impacts on invertebrate prey of Glaucous-winged Gulls (Larus glaucescens), American Black Oystercatchers (Haematopus bachmani), and Northwestern Crows (Corvus caurinus). Estimated per capita interaction strengths exhibited a skewed distribution with many weak interactions and few strong interactions: mean ± 1 SD of log10(interaction strength) = -1.95 ± 1.40 (bird-day/m of shore)-1. Per capita interaction strength correlated poorly (r2 = 0.152-0.157) and nonlinearly with both consumption rates and percentage contribution of a prey species to the diet. Using my observational estimates of per capita interaction strengths, I predicted the species impact of bird predation on different prey taxa. Predictions included strong effects of birds on goose barnacles (Pollicipes polymerus), limpets (Lottia and Tectura spp.), sea urchins (Strongylocentrotus spp.), and large starfish (Pycnopodia helianthoides and Solaster stimpsoni), but little effect on mussels (Mytilus californianus and M. trossulus), dogwhelk snails (Nucella spp.), and acorn barnacles (Semibalanus cariosus). I compared nine of the predictions with 126 results of experimental manipulations of birds. The predictions agreed both qualitatively and quantitatively with the experimental results. These findings suggest that observational measures of interaction strength that have units matching those of dynamical food web models may be reasonable to use in estimating those found in natural communities.