Effects of natural selection from seed predation on serotiny: The role of red squirrels in determining forest structure in lodgepole pine

Background/Question/Methods

Serotiny in lodgepole pine is a trait that influences postfire stand density, community composition, and ecosystem function. Most studies of selection on serotiny have focused on the frequency of crown fires, which is positively related to serotiny; however, our recent and ongoing work suggests that seed predation by red squirrels may select against serotiny, producing a mosaic of contrasting selective pressures and resultant frequencies of serotiny. Thus, an important step in describing selection on serotiny is understanding what drives variation in the intensity of predation from squirrels and whether serotiny itself influences squirrel abundance, or is instead determined by the strength of selection from squirrels. We quantified squirrel abundance, the frequency of serotiny, and several measures of forest structure at 25 sites in the Greater Yellowstone Ecosystem. We used high-elevation sites, where serotiny is rare due to low fire frequencies, to develop a model describing squirrel density independent of variation in serotiny. We then applied this model to low elevation sites, where serotiny is variable, to test whether the frequency of serotiny described additional variance in squirrel density.

Results/Conclusions

We found a strong correlation between squirrel abundance and percent serotiny at low elevation sites, but not at high elevation sites, suggesting that squirrels either prefer sites with less serotiny, or that selection exerted by squirrels has favored a lower frequency of serotiny and the strength of selection varies with squirrel abundance. A habitat model that excluded percent serotiny explained 78% of the variation in squirrel abundance at high-elevation sites. A model using the same predictors accounted for 33% of the variation at low-elevation sites, and all parameter estimates were of the same sign and similar magnitude as at high elevations. Adding percent serotiny did not significantly improve this model (ANOVA F=2.136, P=0.18). We conclude that squirrel density is driven by similar factors at high and low elevations, and that serotiny does not influence squirrel density, supporting the hypothesis that serotiny varies due to a balance of selection from fire and squirrels. These results show that selection resulting from an ecological interaction can mediate the response of an ecosystem to disturbance and are an important demonstration that evolutionary pressures on key traits in foundation species can have profound ecological consequences.