Pattern and scale influence demographic response to climate change in an annual grassland invader

Background/Question/Methods

Climate change will likely influence the spread and impact of biological invasions, and these invaders will have a dynamic influence on their host systems. It is important to predict which invasive species will benefit from future changes in climate, and therefore will need particular attention and prioritization of management efforts. This type of prediction requires detailed information about demographic response. Although many manipulative climate change experiments have been conducted, none have explicitly investigated the impact of local-scale (5-50 cm) distribution pattern on demographic response. To address this issue, we developed a field experiment to understand the interaction of climate change and local-scale patterning on a grassland weed in the Western USA: Aegilops triuncialis(barb goatgrass). In the northern California Coast Range, we manipulated rainfall, seed density, and local-scale seeding pattern in a full factorial experiment. Seeding pattern was either evenly distributed or aggregated in a 5x5 cell ‘checkerboard’ grid scheme (12 seeded cells=aggregated; 25 seeded cells=even). Demographic and environmental data were collected for three years following initial establishment.   

Results/Conclusions

Pattern and scale figure prominently in the demographic response of A. triuncialis to climate change. Aggregated planting of seed led to a multi-year increase in per-plant seed output, especially in the final, and driest, of the three study years. This result indicates that local-scale spatial pattern is an important driver of individual demographics. Pattern also interacts with rainfall and seeding density in its influence on per-plant seed output. There is a contrast between per-plant seed output and plot-scale seed output. When adjusted mathematically for initial seeding pattern, total plot seed output is higher in the evenly distributed seeding treatment. Over three years, this difference became smaller, indicating an increasing compensatory response of plants in aggregated treatments, accounting for areas with no A. triuncialis added. This suggests that over a longer time scale, an aggregated stand of A. triuncialis might have higher population growth rates than an evenly distributed stand given the same number of plants and covering the same total area.