PS 20-60 - Shrub encroachment in a south Florida grassland; implications for mammalian inhabitants

Tuesday, August 8, 2017
Exhibit Hall, Oregon Convention Center
Hanna Innocent1, Modeline Celestin1, Nichole Bishop2, Christina M. Romagosa3, Ray Carthy4 and Wesley W. Boone IV5, (1)Doris Duke Conservation Scholars Program, Gainesville, FL, (2)School of Natural Resources & Environment, University of Florida, Gainesville, FL, (3)Wildlife Ecology and Conservation, University of Florida, Gainesville, FL, (4)FL Coop Fish and Wildlife Unit, University of Florida, Gainesville, FL, (5)Department of Wildlife Ecology and Conservation, University of Florida, Gainesville, FL

Shrub encroachment of grassland systems is a global phenomenon known to influence wildlife communities. We investigated the effect of vegetation on marsh rabbit (Sylvilagus palustris), raccoon (Procyon lotor), and armadillo (Dasypus novemcinctus) occurrence on Sanibel Island in southwest Florida. We placed camera-traps at 2 points on each of 36 grids; half in shrub encroached areas and half in grasslands. We camera-trapped from June-August of 2015 and December 2015-February 2016, yielding 1,772 photos of mammals. We measured the percentage of cumulative vegetative groundcover, sand cordgrass (Spartina bakeri), and giant leather fern (Acrostichum danaeifolium) within ½ m plots on each grid. We also measured the percent coverage of buttonwood (Conocarpus erectus) and cabbage palm (Sabal palmetto), 2 species associated with shrub encroachment, in the canopy of each grid at a 4 m scale. We developed 5 a priori single-variable occupancy models using the aforementioned vegetation variables, which were selected because of their perceived importance to wildlife. These models were tested on each wildlife species. We evaluated candidate models using Program Presence and inferred statistical significance for variables with 95% confidence intervals not crossing zero.


We found that cumulative vegetative percent cover and cordgrass percent cover best accounted for the variation in occurrence of raccoons on our grids. Increases in cumulative vegetative percent cover were significantly correlated with decreases in raccoon occurrence. Similarly, increases in cordgrass coverage were negatively associated with raccoon occurrence, although not significantly. We did not find any influence of vegetative variables on marsh rabbit or armadillo occurrence. Models investigating detection did not converge for any species. These findings reveal a potential shift in the mammalian community resulting from shrub encroachment that could negatively influence other wildlife species. Particularly, the tendency of raccoons to raid the nests of sea turtles (594 Caretta caretta nesting events occurred on Sanibel Island in 2016), diamondback terrapins (Malaclemys terrapin; native to Sanibel’s mangrove estuaries), and various bird species is of concern due to the close proximity of Sanibel’s historic grasslands to coastal and estuarine systems.