Abstract
Changes in population responses to climate are usually studied at broad spatial grains, such as across species ranges. Only a handful of studies have investigated how small-scale variation, for example driven by soil conditions and microtopography, can mediate the responses of population vital rates to climate. Here, we examine responses of Cirsium pitcheri vital rates to climate across five subpopulations occurring in coastal dune locations that range from the foredune to the backdune. We consider five vital rates: seedling survival, adult survival, growth, flowering probability, and the size of new adults. We use generalized linear mixed models to quantify vital rate variation across the dune gradient. Then, we analyze the estimates to quantify the so-called "portfolio effect": the stabilizing effect of spatial asynchrony among the five subpopulations on the population as a whole. Finally, we use a sliding window approach to test whether climate drivers contribute to the portfolio effect. We found evidence of substantial portfolio effects across the dune gradient. These effects arise in part from the opposite responses of vital rates to climatic drivers in our five subpopulations. Strikingly, seedling survival, adult survival, and growth were negatively correlated with precipitation in the subpopulation closer to shore and positively correlated with precipitation in the subpopulation farthest from the shore. We explain this pattern by noting that storms that bring precipitation to the dune system may also cause disturbance and sand burial that negatively affect plants close to the water's edge. Such climate-mediated portfolio effects should have a stabilizing effect on the population abundance over time, with implications for population projections and for assessments of range-wide vulnerability of rare plant species. We call for further scrutiny of variation in small-scale responses to climate and for further empirical tests, especially in coastal dune environments.