COS 63-10 - Modeling invasive plant species in Florida to inform statewide management practices

Tuesday, August 7, 2012: 4:40 PM
E144, Oregon Convention Center
Jessica E. Steele1, Cameron J. Browne2, Timothy J. Fullman1, Michael A. Hyman3 and Kristen E. Sauby4, (1)Geography, University of Florida, Gainesville, FL, (2)Mathematics, University of Florida, (3)School of Natural Resources and Environment, University of Florida, (4)Department of Biology, University of Florida, Gainesville, FL
Background/Question/Methods

Exotic plant invasions are a biological and ecological threat to the integrity and diversity of native ecosystems.  The Nature Conservancy of Florida, together with other state institutions, monitors invasive plants annually and implements treatment and eradication strategies.  Four terrestrial species are particularly threatening in Florida, especially in Everglades National Park, an internationally important wetland.  We used census data for Schinus terebinthifolius, Melaleuca quinquenervia, Lygodium microphyllum, and Casuarina equisetifolia to model species distributions and predict suitable habitat for the Everglades Cooperative Invasive Species Management Area (ECISMA) and for public and private lands in central Florida.  A general linear model (GLM, binomial family) and maximum entropy (MaxEnt) model were developed for these regions.  Comparisons were made between the utility of presence/absence (GLM) and presence-only (MaxEnt) models to assess the best approach for determining habitat suitability and inform future sampling efforts of the Nature Conservancy.

Results/Conclusions

Both presence/absence and presence-only models yielded similar results regarding significant covariate predictors of habitat suitability. Analysis using Fragstats showed high levels of spatial clustering on the landscape for all species (Clumpiness > 0.7). Species showed marked differences, however, in their response to environmental covariates. Landcover classification was a strong explanatory variable in both GLM and MaxEnt models. Casuarina, Melaleuca, and Schinus all utilized evergreen forest more than expected due to chance, while Lygodium tended to concentrate in wetland habitats. Climate appears not to strongly influence Schinus distribution but may limit Casuarina. These findings are important to management agencies and stakeholders such as the Nature Conservancy as they can inform likely areas for invasive species presence, improving the efficiency of management and removal efforts.