There are growing concerns about biodiversity decline and species extinctions due to habitat fragmentation, invasive species, and climate change. Habitat restoration is increasingly used to reverse degradation of rare ecosystems and maintain biological diversity, species interactions, and ecosystem function. The effects of restoration activities on these properties, however, are often not measured. The Midwestern U.S. has a relatively high density of globally rare prairie fen wetlands which support high plant diversity and support a number of rare and endangered plants, insects, and vertebrates. The most common exotic invasive species in Michigan prairie fens is ^{Frangula alnus} (glossy buckthorn). Frangula alnus invasion is known to alter soil and plant community conditions in fens but changes following restoration have never been quantified. The goal of my research was to assess changes in abiotic conditions, species diversity, and species interactions after restoration, and to assess whether they are on a trajectory toward those in uninvaded fen. Toward that end, I quantified differences in invaded and uninvaded prairie fen before restoration began, as well as tracking shifts in resource availability, plant community, pollinator diversity, plant-pollinator networks, and pollinator function in the two growing seasons following restoration of prairie fen. I found a range of responses to restoration across resources and plant community metrics. Light availability, herbaceous plant cover, and relative graminoid abundance increased in restored plots, while soil pH, surface spatial heterogeneity and floristic quality index were initially greater in uninvaded areas versus invaded fen and remained so in the first two years after restoration. These results indicate that some soil and plant community factors change rapidly while others may remain altered for years.

A diverse community of mobile generalist pollinators rapidly re-colonized restored areas. Bee and butterfly communities were nearly absent in invaded prairie fen and responded rapidly to restoration in abundance, diversity, and composition. However, plant species diversity and composition in restored plots remained significantly different than in reference plots, suggesting the plant community may take longer to recover.

I examined plant—pollinator networks and the ecosystem function of pollination using quantitative food webs describing plant-pollinator interactions, and by assessing pollinator function using sentinel Asclepius incarnata plants. Plant-pollinator networks in invaded plots were depauperate, with significantly lower plant and pollinator species richness than restored or reference plots. Network connectance, compartment diversity, generality, and vulnerability did not differ between restored and reference plots, with marginally higher interaction evenness in reference than restored plots. Pollinator function was restored in cleared areas in the first growing season following restoration, with no significant differences in pollinator abundance or diversity between cleared and reference areas. This work indicates that rapid restoration of plant community structure, pollinator diversity, and function are possible in open-structured ecosystems. Combined consideration of species diversity, ecosystem function, and species interactions provides a process-based ecosystem analysis that can inform ecological theory and restoration.