Effects of introduced herbivore removal on an island rare and endemic plant community

Background/Question/Methods

Oceanic islands both support a disproportionate amount of global biodiversity and are particularly vulnerable to the impacts of invasive species. Efforts to eradicate introduced vertebrate herbivores from islands represent important conservation measures, and also opportunities to better understand how invasions alter communities. However, there is little existing research quantifying the response of island plants to herbivore removal, especially for multiple species within a community. This study focuses on the response of 55 rare and endemic plant species on Santa Rosa Island (SRI) to removal of cattle and reductions in elk and deer abundances. From 1994-96, as herbivore control began, a subset of canyon and coastal bluff habitat on SRI was systematically searched, the locations of 442 rare plant occurrences mapped and their total abundances recorded. In 2010-11, we resurveyed approximately 60% percent of this habitat. Locations of rare plant occurrences were mapped with GPS, and abundances and community composition documented.  We present an analysis of changes in the rare plant community for canyon habitats on SRI, based on a comparison of digitized records from the baseline survey and the 2010-11 data in ArcGIS. We quantified community change using three main measures: species presence/absence, species abundances, and spatial extent of distributions.

Results/Conclusions

In Vieja and Cow Canyons, 33 species from the search list were observed in 2010-11. Sixteen of the 25 species in Cow Canyon and 10 of the 18 in Vieja Canyon represented new occurrences. New occurrences were not necessarily for species common elsewhere on SRI at the time of the baseline survey; 11 of these species were not observed on SRI in the original survey.  44% of the species found in 2011 increased in spatial extent by more than 1000%, and 33% by between 0 and 100%.  Rank increases in abundance for a given species were positively correlated with rank changes in spatial extent (F_1,16=24.8, p<0.001).  However, variability in this relationship (R²=0.63) indicates that patterns of recovery differed substantially among species. New species occurrences were unlikely to all represent colonization events, because of low detectability for heavily browsed plants during the baseline survey.  However, our results demonstrate that even the rarest members of this native plant community have shown rapid positive responses to introduced herbivore control, in spite of persistent habitat degradation. Integration of additional recent survey work into these results will allow for a more detailed assessment of differences in responses depending on plant life history and habitat.