Patterns and Drivers of Long-term Plant Community Change in Wisconsin Remnant Prairies.

摘要

Plant communities respond to a number of different ecological drivers, and understanding the relative effects of such drivers is critical to informing effective conservation and management in an era of pronounced global change. However, documenting patterns of change in communities is often hindered by the scarcity of historical data. In this dissertation, I use a unique, historical dataset on plant community composition in Wisconsin prairie remnants to document conspicuous shifts in these plant communities. I replicate survey methods used by the original researchers in 1950 and 1987 to survey 47 remnant grasslands for a third time, in 2012. To assess the role of ecological drivers, I contacted landowners and land managers to obtain records of fire at each site, and used historical and contemporary landscape imagery to quantify landscape connectivity and patch size. I relate these drivers to various plant community metrics to examine the relative impacts of each driver, and whether these roles have shifted over time. I use data on plant functional traits to test if these characteristics relate to gains or losses in species occurrences. I also conducted a fully randomized field experiment to ask whether seed size and site preparation method relate to establishment success.;This historical dataset reveals substantial changes in community composition, and also demonstrates that the pace of change has increased. Annual rates of local colonization and extinction accelerated by 129% and 214% respectively between 1950--1987 and 1987--2012, despite the fact that the second interval is >30% longer. Two anthropogenic drivers, patch area and fire history, increased in importance between these periods, whereas soil moisture declined in importance over the same period. Frequently burned sites were more stable (i.e., diverged less) over time with respect to both functional and taxonomic dissimilarity.;Although total species richness at each site remains similar, their floristic composition continues to diverge, reflecting local extirpations and colonizations of particular groups of species. Most colonization events represent non-native species, which have increased in relative proportion across all sites by more than 500% between 1950 and 2012. Gains in non-native species were accompanied by losses of native species, with short-statured and small-seeded species disappearing particularly quickly.;Contemporary species richness in these remnant grasslands is positively related to patch size. I expected to find a similarly positive relationship between species richness and connectivity, but this prediction is only supported by frequently burned sites. Unburned sites demonstrate no relationship between connectivity and species richness. This pattern may be the result of leaf litter, which accumulates in unburned grasslands, and serves as a major barrier to seedling establishment. Indeed, in the seed addition experiment, rates of successful seedling establishment were 2.6 times higher in plots that were burned prior to planting compared to plots that were unburned. Fewer than one percent all of the seeds sown into our experimental plots were detected as established seedlings in the two seasons following seed addition, demonstrating that the seed to seedling transition is a major bottleneck for plants in this system. Overall, this dissertation contributes to our knowledge of patterns and drivers of long-term change in remnant plant communities, and helps inform effective management and conservation strategy.