Oak savannas are among the most imperiled ecosystems in the United States as a result of habitat degradation and consequently, associated vegetation and wildlife communities have also declined. I evaluated savanna restoration strategies on twelve case studies in Tennessee and Kentucky. These case studies represented a broad range of disturbances and the most advanced savanna restoration sites within the region. I evaluated vegetation and breeding bird responses to landscape and overstory conditions across sites through a meta-analysis. Total grass and forb cover were influenced by overstory metrics but not by topography (P >0.05). Oak regeneration density was influenced by canopy cover, while oak competitor regeneration density was influenced by percent slope and sapling density (P <0.05). With respect to breeding birds, I found forest species persisted within case studies despite substantial disturbance; shrub/scrub birds were common on disturbed sites. Only three obligate grassland bird species, Tyrannus tyrannus, Aimophila aestivalis, Spiza americana, were observed on my sites. Relative abundance of Passerina cyanea was positively related to the groundlayer development; whereas that of Melanerpes erythrocephalus was positively related to basal area of dead trees (P <0.05). Based on my results, canopy reduction and growing-season burns may both be critical for the restoration of savannas within the region. Drum-chopping is a tool that may expedite oak savanna restoration through improved woody competition control, however, its effectiveness has not been investigated. Therefore, I evaluated drum-chopping effects on vegetative structure at Catoosa Wildlife Management Area, Tennessee, during 2008 and 2009 using two adjacent sites with similar fire and overstory removal histories. One site was subjected to drum-chopping in September of 2007, while an adjacent site (control) was not chopped. Drum-chopping reduced grass and forb cover, and oak seedling density, but increased bare ground and density of vines and shrubs versus the control (P <0.05). Except for bare ground, differences were no longer apparent in the second year. Based on my results, drum chopping may reduce midstory vegetation too thick to be effectively controlled by fire, but otherwise has limited utility as a restoration tool. Although wildlife managers have tried to restore savannas using prescribed fire and overstory canopy removal, use of other tools may be warranted. One such method is drum-chopping, which has been used elsewhere to reduce woody competition. However, the effectiveness of this method in restoring oak savannas has not been evaluated. Therefore, I evaluated drum-chopping effects on plant composition at Catoosa Wildlife Management Area on the Cumberland Plateau in Tennessee. Two adjacent sites with similar fire and overstory removal histories were selected for this study. One of these sites was subjected to drum-chopping (CHOP) in September of 2007, the adjacent site was not chopped (NOCHOP). Grass cover differed by treatment (P <0.01) and year*treatment (P = 0.03). Forb cover differed by treatment (P <0.01) and legume cover differed by year (P <0.01), treatment (P <0.01), and year*treatment (P = 0.01). Exposed bare ground differed by year (P <0.01) and treatment (P <0.01). Exposed leaf litter differed by year (P <0.01). Vines and shrubs (<1.37m tall) differed by treatment (P <0.01). Oak seedling (0-30.48 cm tall) densities differed by treatment (P = 0.05). Based on my results, drum chopping may be a valuable tool where woody encroachment has become too thick for fire to be effective or herbicides are not a viable option, but otherwise has limited utility as a tool for oak savanna restoration.
Identifer | oai:union.ndltd.org:UTENN_/oai:trace.tennessee.edu:utk_gradthes-1041 |
Date | 01 May 2010 |
Creators | Barrioz, Seth A |
Publisher | Trace: Tennessee Research and Creative Exchange |
Source Sets | University of Tennessee Libraries |
Detected Language | English |
Type | text |
Format | application/pdf |
Source | Masters Theses |
Page generated in 0.0012 seconds