FERAL HOG (SUS SCROFA) DISTURB A NCE IN SEEPAGE SLOPE WETLANDS By MEGAN E. BROWN A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2014
2014 Megan E. Brown
To my Dad
4 ACKNOWLEDGMENTS First, I would like to thank my parents, Larry Brown and Patricia Maddox, who early on fostered my interest in the natural world and have always encouraged me to pursue my education. I owe much of my accomplishments to their love and support. I would like to expr ess my sincere gratitude towards those who helped me throughout my doctoral experience. Mark Brown gave me the opportunity to be a part of the Adaptive Management of Water, Wetlands, and Watershed IGERT group. My advisor, Debbie Miller, has been great ment or and friend over the past six years. My growth as a scientist has benefited from her experience and guidance. The other members of my committee, Carrie Reinhardt Adams, Doria Gordon, Kevin Heirs, and Wiley Kitchens have been instrumental in providing wis dom, inspiration, and encouragement along the way. A big thank you to Tyler Barry, Sean Claypool, Gina Duke, Hanna Hunsberger, and Allix North for spending many long, hot, humid days in the field, trudging through snake infested woods and wetlands, lugging a meter square that never stayed together, and collecting hundreds of plant specimens....all in the name of data collection. I also give much thanks to Brett Williams with Jackson Guard at Eglin AFB and John Allen with the USDA APHIS Wildlife Services for their support of this project. I started this doctoral journey with seven other amazing women and I am thankful for having had the opportunity to create such wonderful friendships. Finally, I am grateful for the tireless support of my husband, Joshua Horn barger, who uprooted his life to move with me to Gator Country. I also am eternally grateful for my daughter, Odelia (3), who provides me with more joy than I ever knew possible. I hope to instill in her an admiration of the natural world so that she too c an know and experience its beauty. This project was funded in part by a NSF IGERT fellowship, a USDA NIFA pre doctoral fellowship, a McKnight Dissertation Fellowship, a Society of Wetland Scientists Student Grant,
5 the Doris Lowe and Earl and Verna Lowe Sch olarship (UF), and the William C. and Bertha M. Cornett Fellowship (UF).
6 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ ............... 4 LIST OF TABLES ................................ ................................ ................................ ........................... 8 LIST OF FIGURE S ................................ ................................ ................................ ......................... 9 ABSTRACT ................................ ................................ ................................ ................................ ... 12 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .................. 14 Background ................................ ................................ ................................ ............................. 14 Research Goals ................................ ................................ ................................ ....................... 15 2 INDIRECT EFFECTS OF FERAL HOG DISTURBANCE ON FIRE SPREAD ................. 16 Background ................................ ................................ ................................ ............................. 16 Methods ................................ ................................ ................................ ................................ .. 18 Study Area ................................ ................................ ................................ ....................... 18 Data Collection ................................ ................................ ................................ ................ 20 Data Analyses ................................ ................................ ................................ .................. 20 Results ................................ ................................ ................................ ................................ ..... 21 Discussion ................................ ................................ ................................ ............................... 22 Summary ................................ ................................ ................................ ................................ 23 3 LONG TERM MONITORING OF FERAL HOG DISTURBANCE IN SEEPAGE SLOPES ................................ ................................ ................................ ................................ .. 26 Background ................................ ................................ ................................ ............................. 26 Methods ................................ ................................ ................................ ................................ .. 27 Study Area ................................ ................................ ................................ ....................... 27 Data Collection ................................ ................................ ................................ ................ 29 Data Analyses ................................ ................................ ................................ .................. 31 Results ................................ ................................ ................................ ................................ ..... 32 Discussion ................................ ................................ ................................ ............................... 35 Summary ................................ ................................ ................................ ................................ 41 4 EFFECTS OF INTENSITY OF FERAL HOG DISUTRBANCE ON VEGETATION DYNAMICS ................................ ................................ ................................ ........................... 49 Background ................................ ................................ ................................ ............................. 49 Methods ................................ ................................ ................................ ................................ .. 51 Study Area ................................ ................................ ................................ ....................... 51
7 Data Collection ................................ ................................ ................................ ................ 52 Data Analyses ................................ ................................ ................................ .................. 54 Results ................................ ................................ ................................ ................................ ..... 55 Foliar Cover ................................ ................................ ................................ ..................... 55 Richness ................................ ................................ ................................ ........................... 57 Camera Detection of Hog Activity ................................ ................................ .................. 57 Vegetation Composition ................................ ................................ ................................ .. 57 Discussion ................................ ................................ ................................ ............................... 58 Summary ................................ ................................ ................................ ................................ 63 5 EFFECTS OF EXPERIMENTAL SOIL DISTURBANCE ON VEGETATION DYNAMICS ................................ ................................ ................................ ........................... 72 Background ................................ ................................ ................................ ............................. 72 Methods ................................ ................................ ................................ ................................ .. 74 Study Area ................................ ................................ ................................ ....................... 74 Data Collection ................................ ................................ ................................ ................ 75 Data Analyse s ................................ ................................ ................................ .................. 76 Results ................................ ................................ ................................ ................................ ..... 77 Foliar Cover ................................ ................................ ................................ ..................... 77 Camera Detected Hog Frequency ................................ ................................ .................... 80 Richness ................................ ................................ ................................ ........................... 80 Vegetation Compositi on ................................ ................................ ................................ .. 81 Discussion ................................ ................................ ................................ ............................... 82 Summary ................................ ................................ ................................ ................................ 85 6 CONCLUSIONS ................................ ................................ ................................ .................... 98 LIST OF REFERENCES ................................ ................................ ................................ ............... 99 BIOGRAPHICAL SKETCH ................................ ................................ ................................ ....... 113
8 LIST OF TABLES Table page 3 1 Repeated measures ANOVA results for the percentage of hog disturbed plots over time by zone on Eglin AFB from 2002 to 2012. ................................ ................................ 43 3 2 Repeated measures ANOVA results for the percentage of hog disturbed plots over time by hunting access on Eglin AFB from 2002 to 2012. ................................ ................ 43 3 3 The number of hogs removed from Eglin AFB from October 1 of the previous year to September 30 of that calendar year. The percentage of hog disturbed plots and rainfall totals are also indicated. ................................ ................................ ........................ 43 3 4 Repeated measures ANOVA results for the cover of total vegetation, functional groups, and A. stricta by zone and access on Eglin AFB in 2002, 2010, and 2012. ......... 44 4 1 ANOVA results for the foliar cover of total vegetation, functional groups, and A. stricta and species richness by disturbance intensity and exclosure on Eglin AFB. ......... 64 5 1 Repeated measures ANOVA results for the foliar cover of total vegetation, functional groups, A. stricta Dichanthelium spp., Sarracenia spp., and species richness by disturbance frequency on Eglin AFB. ................................ ............................. 86 5 2 Repeated measures ANOVA results for the foliar cover of total vegetation, functional groups, A. stricta Dichanthelium spp., and Sarracenia spp., and species richness by disturbance frequency on Eglin AFB with controls. ................................ ....... 87
9 LIST OF FIGURES Figure page 2 1 The location of seepage slope sites on Eglin AFB. ................................ ........................... 25 2 2 Topographic relationship between seepage slope, upland pine, and baygall communities including position of zones on seepage slopes. ................................ ............ 25 3 1 Topographic relationship between seepage slope, upland pine, and baygall communities including position of zones on seepage slopes. ................................ ............ 45 3 2 The location of seepage slope sites on Eglin AFB. ................................ ........................... 45 3 3 Percent hog disturbed plots from 2002 to 2012. Analysis included both UF and Eglin AFB data. ................................ ................................ ................................ ........................... 46 3 4 Total foliar cover by year ................................ ................................ ................................ .. 46 3 5 Foliar cover of functi onal groups and A. stricta by z one (i.e., position on slope) ............ 47 3 6 Foliar cove r of woody species through time ................................ ................................ ..... 47 3 7 Foliar cover of Aristida stricta through time. ................................ ................................ .... 48 4 1 Topographic relationship between seepage slope, upla nd pine, and baygall communities including position of zones on seepage slopes. ................................ ............ 65 4 2 The location of seepage slope sites on Eglin AFB. ................................ ........................... 65 4 3 Exclosure and open plot pairs ................................ ................................ ........................... 66 4 4 The location of seepage slope sites with motion detecting wildlife cameras. ................... 66 4 5 Total foliar cover by intensity by year inside and outside exclosures on Eglin AFB ....... 67 4 6 Total foliar cover by exclosure by ye ar on Eglin AFB ................................ ..................... 67 4 7 Foliar cover of woody species by i ntensity by year inside and outside exclosures on Eglin AFB ................................ ................................ ................................ ......................... 68 4 8 Foliar cover of graminoid species by intensity by year inside and outside exclosures on Eglin AFB ................................ ................................ ................................ .................... 68 4 9 Foliar cover of graminoid species by exclosure by year on Eglin AFB. ........................... 69 4 10 Average number of hog occurrences per month in seepage slopes with cameras on Eglin AFB from June 2010 through December 2 013. ................................ ....................... 69
10 4 11 Two dimensional nonmetric multidimensional scaling ordination showing cover by species in high disturbance intensity and control plots on Eglin AFB ............................. 70 4 12 Three dimensional nonmetric multidimensional ordination showing cover by species in high disturbance and control plots on Eglin AFB in A) 2010 and B) 2012. ................. 71 5 1 Topographic relationship between seepage slope, upland p ine, and baygall communities including position of zones on seepage slopes. ................................ ............ 88 5 2 The location of seepage slope sites on Eglin A FB. ................................ ........................... 88 5 3 Total foliar cover by disturbance treatment over time on Eglin AFB. .............................. 89 5 4 Foliar cover of forbs by disturbance treatment over time on Eglin AFB. ......................... 90 5 5 Foliar cover of grasses by disturbance treatment over time on Eglin AFB. ...................... 91 5 6 Foliar cover of woody species by disturbance treatment over time on Eglin AFB. .......... 92 5 7 Foliar cover of A. stricta by disturbance treatment over time on Eglin AFB. ................... 93 5 8 Foliar cover of Dichanthelium spp. by disturbance treatment over time on Eglin AFB. ................................ ................................ ................................ ................................ ... 94 5 9 Foliar cover of Sarracenia spp. by disturbance treatment over time on Eglin AFB. ........ 95 5 10 Species richness by disturbance treatment over time on Eglin AFB. ................................ 96 5 11 Three dimensional nonmetric multidimensional ordination showing cover by species by disturbance treatment on Eglin AFB in June 2010 and Sept 2013.. ............................. 97
11 LIST OF ABBREVIATIONS ANOVA E GLIN A FB Analysis of Variance Eglin Air Force Base MRPP Multi response permutation procedure NMS Nonmetric multidimensional scaling UF University of Florida
12 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy FERAL HOG (SUS SCROFA) DISTURBANCE IN SEEPAGE SLOPE WETLANDS By Megan E. Brown May 2014 Chair: Debbie Miller Major: Interdisciplinary Ecology Florida is home to one of North America's most unique and diverse natural ecosystems, the seepage slope wetland. Unusual hydrology and frequent fires have resulted in a habitat that supports a variety of insectivorous and other endemic, helophytic herbaceo us plants. Feral hog ( Sus scrofa ) foraging has resulted in widespread soil disturbances in seepage slope wetlands in the Florida Panhandle. It is proposed that feral hog rooting is a serious threat to this community because rooting set s back succession, ca use s changes in species composition and plant population structure reduces unique species, and inhibits fire spread The se potential changes in the seepage slope plant community are particularly important because these wetlands provide increasingly rare h abitat for several threatened and endangered plant species. We conducted plant surveys to investigate both the long and short term vegetation dynamics in response to hog disturbance. We demonstrate that hog disturbance has not declined since the USDA APHI S Wildlife Services began trapping hogs on Eglin AFB in 2003 and theorize rainfall and other potential drivers of hog disturbance The short term exclosure studies provide evidence that the intensity and frequency of disturbance alter vegetation compositio n even when foliar cover is partially or fully regained. These results also indicate that Aristida stricta cover is reduced by soil disturbance, while woody cover is maintained A positive
13 feedback is likely to result with reduced grass cover and fire spre ad which further contributes to increased woody cover. In addition, we used GPS collected data to compare areas damaged by hogs and areas burned during prescribed fires to quantify the indirect effects of hog rooting on ecosystem function. We found that h og rooting reduced vegetation cover and decreased fire spread. Through this work, we gain insight to the mechanisms that affect the rate and trajectory of vegetation development in this plant community, with significance for both theoretical and management purposes.
14 CHAPTER 1 INTRODUCTION Background Seepage slope wetlands and bogs of the southeastern United States are of critical conservation concern due to their high levels of species richness and endemism (Johnson 2001). Unusual hydrology oligotrophic conditions, and frequent fires have resulted in a habitat that supports a variety of carnivorous and oth er heliophytic, herbaceous plants The s outheastern coastal plain con tains the highest number of carnivorous plan t species in North America (Hermann 1995). Only one percent of seepage slopes that originally existed along the Gulf Coast al Plain still remain (FNAI 1990). These wetland features have been degraded by encroachment of woody species in the absence of fire, conversion into ponds, or drainage As a result, remnant seepage slopes communities provide increasingly critical but diminishing habitat for threaten ed and endangered plant and amphibian species (F NAI 1990; Hermann 1995). The Spanish explorer, Hernando d e Soto introduced domestic hogs ( Sus scrofa ) in the 16 th Century, (Mayer & Brisbin 1991) and they are now established in 3 5 states and Canadian provinces (Giuliano 2009). Feral hogs pose a serious threat to seepage slopes (FNAI 2010) because they overturn soil when foraging (Howe & Bratton 1976) Thi s behavior creates hummocks and tussocks that alter edaphic properties ( Singer et al. 1984; Jensen 1985) and hydrology (Arrington et al. 1999), accelerates decomposition, alters soil nutrient concentrat ions (Singer et al. 1984; Wirthner et al. 2012; Krull et al. 2013), and facilitates soil erosion (Bratton 1975; Howe & Bratton 1976 ). In addition to these changes in edaphic conditions, the mechanical disturbance of hog foraging deracinates vegetation ( En geman et al. 2007), reduces vegetation cover ( Bratton 1974, 1975; Engeman et al. 2007; Siemann et al. 2009 ; Wirthner et al. 2012 ) and impacts regeneration
15 (Lipscomb 1989; Sweitzer & Van Vuren 2002; Siemann et al. 2009 ) Hog rooting can result in reduced s pecies diversity (Bratton 1975; Kotanen 1995; Tierney & Cushman 2006 ; Siemann et al. 2009 ) and altered species composition (Bratton 1974; Aplet et al. 1991; Siemann et al. 2009; Cuevas et al. 2010). These changes in species composition are due to both dire ct species removal, as well as indirect effects on the ecosystem processes including the disruption of the fine fuel component that facilitates fire spread. Vegetation may (Arrington et al. 1999 ; Tierney & Cushman 2006 ) or may not be able to recover from r ooting disturbances (Bratton 1975). Hog rooting and forag ing could lead to the extirpation of species ( Bratton 1974; Recher & Clark 1974; Challies 1975; Spatz & Mueller Dombois 1975) Deleterious impact from extensive, repetitive rooting of nati ve plants is documented as a critical threat to the conservation of pitcherplant wetland communities in Florida (Johnson 2001). Research Goals It is important to understand how disturbances affect natural communities for both theoretical and management purposes. This research explore d how feral hog disturbance affects the rate and trajectory of vegetation development in seepage slopes of the Florida Panhandle in four interrelated studies. Chapter 2 quantifies the effect of this novel disturbance on the historic fi re regime Chapter 3 investigat es vegetation dynamics in seepage slope s over a 10 year period. Chapter 4 evaluates vegetation development in exclosures erected in areas of varying disturbance intensity. Chapter 5 assesses the role of frequency of disturban ce and time since last disturbance in determining vegetation composition and species richness This research will provide better understanding of: 1) long term patterns of vegetation dynamics; and 2) m echanisms driving shifts in plant composition in response to feral hog disturbance The results of this work can be used with continued monitoring data to direct management actions to prevent degradation of seepage s lopes on E glin AFB.
16 CHAPTER 2 INDIRECT EFFECTS OF FERAL HOG DISTURBANCE ON FIRE SPREAD Ba ckground Ecosystems adapted to disturbance by fire are distributed a ll over the globe (Drewa et al. 2002; Bond & Keeley 2005). These systems include northern latitude boreal forests, lower latitude shrublands, grasslands, and savannas and eucalyptus woodla nds in Australia. Different fire regimes have selected for and resulted in different plant attributes and adaptations. Worldwide these fire maintained systems have different biota, but recent decades of fire suppression in these systems all demonstrate: 1) risks associated with increasing fue l accumulation (Bond & Keeley 2005); 2) a loss of fire dependent species (Frost et al. 1986; Hermann 1995; Frost 2000; Lett & K napp 2003; Glitzenstein et al. 2003; Bond & Keeley 2005; FNAI 2010); and 3) management cha llenges in establishing habitat needs (e. g., determining fire return intervals, season of burn, and use of fire surrogates) (Slapcinsky et al. 2010). These threats and challenges are especially significant in Florida where 70% of the terrestrial ecosystems are fire dep endent or adapted (FNAI 1990). Seepage slopes, shrub bogs, and baygalls are wetlands that occur downslope from Pinus palustris Mill. (Longleaf pine) upland communities. These communities have a successional relationship dependent on fire frequ ency Frequent fire upslope from shrub bogs and b aygalls maintain the herbaceous component of the seepage slope ( Eleuterius & Jones 1969; FNAI 2010). Fire return intervals are estimated to have been one to three years historically in the surrounding uplands ( Frost 1995; Frost 2000 ; Frost 2006), while somewhat lon ger in wet seepage slopes (FNAI 2010). Further downslope, shrub bogs are estimated to have longer burn return intervals of 20 to 50 years (Wharton et al. 1976 ). In the absence of fire, shrub bogs replace the herbaceous wet prairie species and eventually, these shrub bogs succeed to baygalls dominated by hardwood
17 species (Means & Moler 1979). While fire is a disturbance in the classical sense, it is essenti al to the stability of the seepage slope community (Frost et al. 1986). Beginning in the 1920's, suppression of wildfires became public policy in the southeast ern United States Realizing the importance of fire in the maintenance of natural communities, la nd managers now use prescribed fire to mimic natural, lightning ignited fire. Most prescribed fire is applied in the dormant season rather than the summer growing season (Knapp et al. 2009). Alteration of natural fire regimes can lead to degradation of th e remaining seepage slopes (FNAI 2010). H istorically lightning ignited fires during the spring or the beginni ng of the growing season (Olson 1992). This seasonal timing of spring drought and lightning producing thunderstorms generate d upland fires that wo uld carry down into seepage slope s (Olson & Platt 1995). It is well documented that fire suppression and in some cases, dormant season prescribed fires have allowed the encroachment of woody species ( Frost et al. 1986; Frost 1995; Glitzenstein et al. 1995 ; Olson & Platt 1995; Drewa et al. 2002; Slocum et al. 2003) and loss of herbaceous species in seepage slopes (Eleuterius & Jones 1969; Folkerts 1982; Frost et al. 1986 ; Hermann 1995) Fire is influenced by multiple factors including fuel conditions, weather and wind conditions (Cheney et al. 1993; Turner et al. 1994 ; Brooks et al. 2004) and topogra phy (Weber 1990; Turner et al. 1994; Brooks et al. 2004). This chapter focus es primarily on the relationship between fuel continuity and fire spread. The h orizontal continuity of fuels can determine the extent of fire s pread (Brooks et al. 2004; Nader et al. 2007). For example, wider gap distance between fuel beds can result in decreased flame contact a nd thus decreased spread (Weber 1990; Finney et al. 2010 ). Modeling has shown that fuel gaps are more likely to inhibit fire spread
18 when fuels are heterogeneous ( Kerby et al. 2007), like fuels of bunch grass systems (Finney et al. 2010) such as the P palustris (L ongleaf pine) / A ristida stricta Michx. ( W iregra ss ) ecosystem of the Southeastern United States. Herbivores can alter a fire regime by ingesting vegetation that is critical for fire spread ( Savage & Swetnam 1990; McNaughton 1992; Knapp et al. 1999 ; Waldram et al. 2008; Hierro et al. 2011 ) Feral hog foraging that results in uprooted vegetation ( Bratton 1974; Engeman 2007) and increased cover of bare ground ( Bratton 1975; Kotanen 1995; Engeman 2007 ; Doup et al. 2010) effectively reduces the load and continuity of fine fuels This reduction of fine fue ls can potentially : 1) reduce the extent of fire sp read (Harrington & Hodgkinson 1986; McNaughton 1992 ; Knapp et al. 1999; Kerby et al. 2007; Waldram et al. 2008); and 2) alter the frequency ( Savage & Swetnam 1990; Van Auken 2000 ; Hierro et al. 2011 ) and i ntensity of future fires ( Hobbs et al. 1991; V an Langevelde et al. 2003; Briggs et al. 2005; Kerby et al. 2007; Hierro et al. 2011 ). This chapter investigates the indirect effects of hog rooting on ecosystem function. We sought to determine whether soil di sturbances associated with feral hog foraging reduce fire spread in seepage slopes on E glin AFB. Methods Study Area Sites for this study were located on Eglin AFB in Walton County in the western Florida Panhandle (Figure 2 1) This area has a mean annual precipitation of 158cm (DoD Air Force 1995). Precipitation tends to peak in the summer months and is lowest in May and October Lightning producing t hunderstorms are common in the spring and summer mo nths ( NOAA 2002 ). Suppression of wildfires and limited use of prescribed fire resulted in infrequent fire in seepage slopes on Eglin AFB until about 2002 Subsequently, Eglin land
19 managers increased the burning frequency to annual and biennial prescribed fires in both the growing and dor mant seasons (pers. comm. 2014 Eglin AFB land manager ). The soils on the eastern portion of EAFB are loamy sands. Creeks have created rolling erosional hills with relief of up to about 30 meters (NRCS 1989). The mesic, u pland pine community (FNAI 1990) fou nd on these loamy sands is speckled with wet prairies on hillsides called seepage slopes (Figure 2 2). These seepag e slopes result from groundwater seeping downslope above an impermeable layer of clay that intersects the hillside. A unique variety of herba ceous vegetation can be found in seepage slopes including carnivorous plants like Sarracenia spp. ( P itcherplants ) Drosera spp. (S undews) Pinguicula spp. ( B utterworts) and Utricularia spp. (B ladderworts). Upslope, in the drier portions, characteristic species include Aristida stricta Ctenium aromaticum (Walter) A. Wood (Toothachegrass) and Rhexia spp. (M eadowbeauties) while downslope the longer hydroperiod accommodates Eriocaulon spp (P ipeworts) Lophi ola aurea Ker Gawl. (Golden C rest) Pleea tenuifolia Michx. (Rush F eatherling) Rhy n chospora spp. (B eaksedges) and Xyris spp (Yelloweyed G rasses) These wetlands are dependent on ground water and dry out in times of drought. It is during dry spells that seepage slopes are able to carry fire Frequent fire is necessary to prevent shrubs, like Cliftonia monophylla (Lam.) Sarg (Black T iti) from invadi ng from the shrub bogs and baygalls alon g the seepage creek and shading out helophytic herbaceous species. Other scattered shrubs that are common in seepage slopes include Gaylussacia mosieri Small (Woolly huckleberry) and Hypericum spp. (St. John's wort) (FNAI 2010) The baygalls located at the base of the slope are saturated, peat filled seepage depressions that support a dense forest of evergreen hardwoods. This forest is dominated by Magnolia virginiana L. (Sweetbay) Persea palustris (Raf.) Sarg. (Swamp bay) and Gordonia lasianthus
20 (L.) J. Ellis (Loblolly bay) Sphagnum spp. (Sphagnum moss) can form mats on the ground (FNAI 1990). A mature canopy of these fire intolerant hardwoods indicates the lack of destructive fire for many years (Clewell 198 6). These areas may burn catastrophically in times of drought (FNAI 2010). Data Collection Research q uestions 1 and 2. Did fire carry through a grea ter percentage of seepage slope area with or without hog disturbance ? D id hog disturbed area s burn? Prescrib ed burns were used to examine the effect of hog disturbance and fuel continuity on the extent of fire spread We analyzed fire spread at 21 seepage slopes, some burned more than once (n=37), on Eglin AFB in 2010, 2012, and 2013 A reas where hog foraging resulted in discontinuous fuels were captured as polygons using a handheld Trimble GPS. These sites ranging in size from 2,549m 2 to 16,316m 2 were then burned by fire crews for management purposes, as instructed by Jackson Guard After each burn wa s exti nguished, any areas of seepage slope vegetation not consumed in the fire were captured as polygons using the Trimble GPS Research q uestion 3. Was total foliar cover prior to burning greater in plots that carried fire compared to those that did not? B urn histories for 38 control exc losure plots (no recent hog disturbance) used in the intensity experiment and 61 experimental disturbance treatment plots used in the frequency experiment (Chapters 4 & 5 ) were also examined to determine if prior to burning unbu rned plots had less foliar cover than burned plots. T otal foliar cover prior to the burn was recorded at sites burned between 2010 and 2013. Plots were categorized as burned or unburned after prescribed fires Data Analyses Research q uestion s 1 and 2. Did fire carry through a grea ter percentage of seepage slope area with or without hog disturbance? Did hog disturbed areas burn? ArcGIS 9.2 software
21 ( ESRI 2009) was used to determine if there was a difference in: 1) the percentage of hog d isturbed seepage slope that burned an d the percentage of undisturbed seepage slope that burned ; and 2) the percentage of hog disturbed area that burned and did not burn Areas of hog disturbance, areas left unburned after the fire and total seepage slope areas were calculated The analysis t ool "Union" was used to compute the geometric intersection of the three input layers From this output layer, the percentage of the hog disturbed seepage slope that did and did not burn and the percentage of undisturbed seepage slope that did and did not b urn was calculated. Student s t test s ( PROC TTEST ; SAS 9.3 ; SAS Institute 2011) was used to reveal any differences in these percentages Research q uestion 3. Was total foliar cover prior to burning greater in plots that carried fire compared to t hose that did not? Total cover from plots used in the intensity and frequency experiments at sites that burned was also analyzed using a Student s t test ( PROC TTEST; SAS 9.3 ; SAS Institute 2011) to discern any difference in total foliar cover prior to burning in burned and unburned plots. Results Researc h question s 1 and 2. Did fire carry through a grea ter percentage of seepage slope area with or without hog disturbance? Did hog disturbed areas burn? S oil disturbance caused by hog rooting is associated with reduced fire spread. P rescribed fires burn ed significantly more area in undisturbed portions (75.80%) than hog disturbed portions (21.04%) of the slopes (t<0.0001). Also, significantly more hog disturbed area was unburned than burned after prescribed fire (t<0.0001). This study found that 77.32% of hog disturbed portions of seepage slopes remained unburned after prescribed burns. Research q uestion 3. Was total foliar cover prior to burning greater in plots that carried fire compared to those that did not? Pre burn total foliar cover was significantly lower in plots
22 that did not burn than plots that burned (t<0.0001). Prior to the burn, average total foliar cover of burned plots was 71.31%, while unburned plots averaged 39.15% cover This indicates that there is a threshold between approxi mately 40% and 70% total foliar cover below which fire spread is reduced Discussion The literature lacks research concerning indirect effects of introduced herbivores on ecosystem processes (Stritar et al. 2010), whereas most studies quantify the effects of native herbivores on fire regime in African ecosystems (Van Langevelde et al. 2003; Waldram et al. 2008; Holdo et al. 2009). As we hypothesized, vegetation in areas of hog disturbance remains unburned and is associated with reduced fire spread in seepag e slopes. Non indigenous hog foraging exposes bare ground ( Bratton 1975; Kotanen 1995; Engeman 2007 ; Doup et al. 2010) and reduces fine fuel loads, ultimately reducing fire spread. This research indicates that this novel disturbance affects ecosystem proc esses in this wet prairie community. We categorized hog disturbance with less than 40% total foliar cover as high intensity hog disturbance in Chapter 3. After three years of recovery, areas of high intensity hog disturbance failed to have average total fo liar cover of 70% or greater. This study found that there is a threshold of between 40% and 70% total cover, below which fire is likely not to spread. It can be interpreted from these data that areas of high intensity hog disturbance are less likely to car ry fire Future research may be necessary to investigate the extent of high intensity hog disturbance in seepage slopes to quantify the degree of this fuel reduction problem. Reduced fire spread has sig n ificant implications for the physiognomy of the seepage slope. In the absence of fire, the longleaf pine ecosystem experience s increased hardwood density and decreased herbaceous cover and diversity ( Frost et al. 1986; Hermann, 1995; Frost 2000; Lett & Knapp 2003; Glitzenstein et al. 2003 ; Bond & Keeley 2005; FNAI 2010). As a
23 result, most indicator species groups for bogs and wet savannas (i.e. forbs, sedges, insectivorous plants) are extremely sensitive to reductions in fire frequency (Glitzenstein et al. 2003). Woody plants may set up positive feedbac k that facilitate further woody invasion (Schlesinger et al. 1990; Callaway & Davis 1 998 ), making restoring herbaceous dominance in shrub encroached areas unlikely even with the reestablishment of the historic fire regime (Archer 1989; Lett & Knapp 2003; B riggs et al. 2005). Losses of this community due to changes in physiognomy in the past have not been well documented. Like other wet prairies of the southeastern United States that occur on seasonally saturated soils, seepage slopes can be spatially narrow Much of the acreage of this wet community is limited to narrow areas that are less than 50 meters wide between the pyrogenic upland savanna and deep wetlands. Wet prairies, including seepage slopes, also share a considerable number of species with the ad jacent communities. For these reasons, many authors have regarded wet prairies as ecotones rather than distinct communities. Thus, losses in this community due to woody encroachment and other anthropogenic activities have been overlooked by conservationist s (Clewell et al. 2009). We acknowledge that f ire behavior is dynamic by nature, influenced by fuel conditions, weather or climate, and wind conditions (Cheney et al. 1993; Turner et al. 1994; Brooks et al. 2004) and topography (Weber 1990; Turner et al. 1 994; Brooks et al. 2004). While we did not control for environmental conditions, we found that fuel continuity is a major predictor of fire spread in seepage slopes. Summary Feral hog foraging creates a complex network of novel ecosystem linkages and feedbacks that have both direct and indirect effects on the biological and physical components of the ecosystem. Yet, r esearch on indirect effects of this exotic species on community structure and
2 4 ecosystem function is limited (Barrios & Ballari 2012) Ou r research indicates that feral hogs alter fuel continuity in the landscape, indirectly altering the fire regime, and potentially resulting in changes in community structure. Implications of this finding necessitate management plans to remove hogs to lower densities or eradicate when possible in an attempt to reduce impacts and preserve community structure Further investigations, involving both direct and indirect effects of feral hog disturbance, should involve a whole ecosystem approach (Barrios & Ballar i 2012)
25 Figure 2 1. The location of seepage slope sites on Eglin AFB. Figure 2 2. Topographic relationship between seepage slope, upland pine, and baygall communities including position of zones on seepage slopes.
26 CHAP TER 3 LONG TERM MONITORING OF FERAL HOG DISTURBANCE IN SEEPAGE SLOPES Background Management options to address feral hog damage are limited, with trapping/hunting used as the primary means to address the threat to n atural ecosystems. The effectiveness of hog removal at reducing hog densities and density dependent damage is mixed. A study in 2007 assessed the effect of three years of hog trapping on hog d isturbance in seepage s lopes on Eglin AFB in northwest Florida The findings suggest th at r emoval of hogs from the reservation has dramatically reduced the extent of hog disturbance in seepage slopes (Engeman et al. 2007 ). In contrast, recent research shows compensatory reproduction and immigration to increase in response to hog removal util izing traps and ground shooting (Hanson et al. 2009; Sparklin et al. 2009). Thus these common management techniques may see limited success in reducing hog densities due to a density dependent increase in recruitment (likely immigration) that can exceed re moval rates (Hanson et al. 2009). Hogs also have a high reproductive capacity ( Dzieciolowski et al. 1992; Choquenot et al. 1996; Taylor et al. 1998 ) Models indicate that pre removal densities can be reached in a short period of time if removal techniques do not remove a high proportion of the population (i.e. greater than 70% instantaneous kill) ( Dzieciolowski et al. 1992). Another concern is that trapping success is reduced when hogs become "trap shy" after previous experience with traps (Saunders et al. 1993; Richardson et al. 1997). Populations of hogs can rebound after the initial reduction. This failure to reduce hog numbers can then correspond to failure to reduce hog disturbance (Hone 1995; Sweitzer 1998; Hone 2002; Sweitzer & Van Vuren 2002). Monit oring data can be used by managers to determine : 1) if a system is departing from the desired state ; 2) the success of management actions; and 3) the effects of disturbances (Legg
27 & Nagy 2006). At Eglin AFB, seepage slope monitoring was initiated in 2003 to evaluate the extent of hog disturbance and the success of hog trapping at reducing that extent (Engeman et al. 2007). As with most monitoring programs, there were no true control sites as half the sites were trapped and half the sites were sport hunted This design can preclude conclusive results. The observed change in hog disturbance after hog trapping began in seepage slopes on Eglin AFB was attributed to the trapping (Engeman et al. 2007). Yet change is inherent in all ecosystems, so the observed ch ange may be unrelated to the management action (Legg & Nagy 2006) (i.e., hog trapping). Without repeated and contemporaneous sampling before and after the initiation of the management action at control and treatment sites, causality can only be inferred wh en a plausible mechanism is identified and reasonable alternative mechanisms are investigated and rejected (Schroeter et al. 1993). Also, monitoring data involving hog population dynamics should be analyzed over a long term time scale because hog populatio ns can experience great interannual fluctuation in relation to food availability (Peine & Farmer 1990; Brisbin & Sturek 2009 ). The objectives of Chapter 3 are to investigate the long term fluctuations in extent of hog disturbance and corresponding florist ic changes in seepage slopes on Eglin AFB. We analyze environmental factors and hog removal data that may affect variability in interannual patterns of hog disturbance. Finally, we explore the spatial patterns of hog damage within the complex hydrological gradient within seepage slope wetlands and assess implications of damage across that gradient Methods Study A rea This research was conducted on E glin AFB in the western Florida Panhandle where the mean annual temperature is 18.3 o C and mean annual precipitation is 158cm (DoD Air Force 1995). Precipitation is greatest in the summer months
28 (NOAA 2002). Eglin AFB is a 187, 548 ha reservation that represents the largest remnant of Pinus palustris Mill. ( longleaf pine ) ecosystems (Compto n et al. 2006; Mitchell et al. 2009). A variety of birds and mammals, including feral hogs, can be hunted for sport on 105 218 ha of Eglin AFB during hunting season, from mid October until mid February ( United States Air Force 2002). Sites were located on the eastern portion of E glin AFB where soils are of the Citronelle Formation, the loamy sand remains of a Pliocene Pleistocene age delta plain. The gently rolling erosional hills have relief of up to about 30 meters and areas of steeper slope along creeks (NRCS 1989). The clay component in these sandy soils creates more mesic conditions than the similar, nearby xeric P palustris sandhills. These mesic longleaf pine woodlands are classified as U pland P ine Forest (FNAI 1990). Embedded in the upland pine comm unity are seepage slopes Seepage slope s are wet prairies located on hillsides (Folkerts 1982). Soil saturation is sustained by the downslope seepage of groundwater above a relatively impermeable layer o f clay, organic matter, or rock. The impermeable layer prohibits groundwater penetration and results in seeps along the hillsides of dendritic systems of streams. The characterist ic dense and diverse herbaceous vegetation includes Rhy n chospora spp. (B eakse dges) Aristida stricta Michx. (Wiregrass) Ctenium aromaticum (Walter) A.W. Wood (Toothachegrass) Dic h anthelium spp. ( W itchgrasses) Xyris spp. (Y elloweyed G rass es ) Rhexia spp. (M eadowbeauties) Eriocaulon spp. (P ipeworts) and Lycopodiella spp. (C lub M osses) (FNAI 2010). Carnivorous plants, including Sarracenia spp. (P itcherplants) Drosera spp. (S undews) Pinguicula spp. (B utterworts) and Utricularia spp. (B ladderworts) have evolved the ability to trap insects in order to acquire nutrients in these a naerobic and acidic conditions that limit nutrient availability (Schnell 1976). Frequent fire within the seepage slope maintains s hrubs at a reduced height In the absence of frequent fire,
29 Cliftonia monophyll a (Lam.) Sarg. (Black T iti) and other shrubs invade upslope from the shrub bogs or baygalls found a long the seepage creek (Figure 3 1) and outcompete the heliophytic, herbaceous vegetation. These wetlands dry out in times of drought due to their dependence on groundwater (FNAI 2010). Data Collection Research q uestion s 1 and 2. Has the extent of hog damage and its effect on foliar cover changed from 2002 2012 ? Are hogs preferentially disturbing a particular zone of the slope? In 2002 222 permanent plots were established at 24 randomly chosen indepen dent seepage slope sites on E glin AFB by UF (Figure 3 2) Half of the se 24 sites were located in areas open to sport hunting and half of the sites were located in areas closed to sport hunting. Each seepage slope site was stratified by position on slope in to three zones (i.e., lower, middle, upper) based on the approximate distance from the seepage creek Zonation represented a moisture gradient and was delineated to account for differences in soil moisture and direct sunlight. G enerally the low er zone near the creek wa s equivalent to the toe slope, the mid zone wa s equivalent to the foot slope and the upper zone, furthest upslope, wa s equivalent to the mid slope (Figure 3 1) The low er zone wa s characterized by saturated, organic mud and seeping w ater. The mid dle zone, although not present at every site due to size and t opograph ic differences, h ad saturated organic sands, but wa s drier than the low er zone. The upper zone consisted of sandy soils with low soil moisture and organic content T he upper boundary of the slope wa s delineated by dry, sandy soil and the dominance of upland species including a canopy of pine trees Depending on the size of the slope, up to five 1m 2 permanent plots were randomly placed in each zone within each slope. In the s ummer of 2002, 2010, and 2012 each permanent p lot w as characterized as disturbed or und isturbe d, depending on whether hog d isturbance was present. A modified Daubenmire scale was used to estimate cover of total vegetation, forbs, grasses, and woody
30 vegeta tion Th e modified Daubenmire scale (0= 0%, 0.5 5 = 3%, 5 10 = 8%, 10 15 = 13%, 15 25 = 21%, 25 50 = 38%, 50 75 = 68%, 75 95 = 88%, 95 100 = 98 % ) reflects the precision of ocular estimation of foliar cover that can be expected with multiple observers and allows for data comparison (Daubenmire 1959 ). Although there can be between observer error, this study used pairs of observers and observers trained by the same person in attempt to reduce observational error to the greatest extent possible ( Vittoz & Guisan 2007) Foliar cover of the keystone species A. stricta was also estimated because it dominates the undisturbed upland pine system and carries fire throug h dryer portions of the seepage slope (Noss 1989; Outcalt et al. 1999). Land managers at Eglin AFB monitored additional seepage slopes, intermingled with those established by UF Both groups monitored ten of the same slopes. For the internal Eglin monitori ng project, twenty randomly selected, 1m 2 plots were permanently marked with rebar at each of 28 randomly chosen seepage slope sites on E glin AFB in 2003. The sites were located half in areas open to hog hunting and half in areas closed to hog hunting. C ov er of hog disturbance was estimated during surveys in 2003, 2004, 2005, 2007, and 2011. By 2010, seven sites were excluded because access had become restricted. Data from the remaining 21 sites was used for this study, bringing the total number of sites fo r the study to 34. Plots at the 21 Eglin sites were also surveyed in 2010 by UF researchers. Research question 3. Do es hog removal or do environmental variables account for changes in hog disturbance? T he USDA APHIS Wildlife Services began removing hogs on Eglin AFB in 2003. Hog removal primarily involved trapping with pen traps for euthanization with some control hunting (not sport hunting) Hog removal was conducted in areas closed to sport hunting from 2003 to 2005 (Engeman et al. 2007) In more rece nt years, hog removal has been extended to include areas open to sport hunting. T he rainfall on Range C52N, approximately 4
31 kilometers north of the seepage slope sites, and number of hogs removed by trapping each year w ere recorded by Eglin AFB The occurre nce of prescribed burns in seepage slopes evaluated by UF and Eglin AFB were also documented by Eglin AFB Data A nalyses Research q uestions 1 and 2. Has the extent of hog damage and its effect on foliar cover changed from 2002 2012? Are hogs preferentiall y disturbing a particular zone of the slope? This study consisted of a completely randomized design with repeated measures. Each seepage slope site represented a replicate within which there were three fixed zones along a topographic gradient. Main effects were zone (lower, middle, and upper) year (2002, 2010, and 2012) and hunting access (open or closed) Significance of main effects and interactions were evaluated using a repeated measures one way ANOVA ( ) in a generalized linear mixed mo del (GLIMMIX; SAS 9.3 ; SAS Institute 2011) Dependent variables were total foliar cover, forb cover, grass cover, woody vegetation cover, A. stricta cover and percent hog disturbed plots. Seepage slope sites were considered a random effect. Data was squar e root transformed to meet assumptions of normality. Analysis of Eglin AFB data from 2003 2005 was previously published by E n gema n et al. (2007). Because the additional plots established by Eglin AFB were not stratified by zones, analysis of Eglin AFB plot s evaluated main effect s of year (2007, 2010, and 2011 ) and hunting access (open or closed) on percent hog disturbed plots Tukey Kramer tests were used to analyze d ifferences between means (Zar 2010 ). Research q uestion 3. Does hog removal or do environmental variables account for changes in hog disturbance? Correlations and a multiple regression were used to determine relationship between precipitation, hog removal, vegetation variables, and percent hog disturbed plots. UF and Eglin AFB data were combined for all P earson correlation s. The first correlation w as run (CORR; SAS 9.3; SAS Institute 2011) to detect correlation between percent of hog
32 disturbed plots and number of hogs removed annually by trapping efforts. In add ition, Pearson correlation analyzed the relationship between total vegetation cover and: 1) rainfall during the previous 11 months; and 2) time since last burn. Pearson correlation was then used to analyze the relationship between percent of disturbed plot s and total rainfall. The rainfall totals were from the following time periods: 1) the 1 1 months prior to vegetation surveys ; 2) Jan May of each vegetation survey year ; 3 ) March June of the year prior to vegetation surveys ; and 4) August through November of the year prior to vegetation surveys Rainfall totals from the previous spring (March June) and fall (August November) were investigated because hog breeding peaks in the spring and fall in Florida (Gi uliano 2009). Pe ars on correlation was used to evaluate the relationship between number of hogs removed annually and rainfall totals in the previous spring and fall. Pearson correlation was also used to identify correlation between time since burn and: 1) percent cover hog da mage; 2 ) woody species cover; and 3 ) A. stricta cover. A multiple regression was used to determine the relative importance of independent variables (including percent hog disturbed plots, time since burn, and rainfall totals from the previous August Novemb er) in determining total vegetation cover. Results Has the extent of hog damage and its effect on foliar cover changed from 2002 2012 ? No clear long term pattern in the percentage of plots disturbed by hogs was observed (Figure 3 3) For UF study slopes, there was no difference ( p=0.6 474 ) in the percent disturbed between 2002 and 2010 ( Table 3 1 ). However, there was a decline in disturbed plots in 2012. As a result, there was a significant difference in the percent disturbed from 2002 to 2012 (p=0.0387) an d 2010 to 2012 (p=0.0034). There were significantly more hog disturbed plots in areas open to hog hunting (p=0.0222) compared to area s closed to hog hunting (Table 3 2). Although not stratified by slope position, plot s established and monitored by Eglin AF B showed an increase in the
33 percent disturbed from 2007 to 201 0 (p<0.0001) and from 2007 to 2011 (p<0.0001) ( Table 3 1). P ercent disturbed did not differ between 2010 and 2011 (p=0. 7938 ) nor did it differ between areas open and closed to hog hunting (p=0. 6195). For the most part, total vegetation cover reflected intensity and frequency of feral hog damage and recovery rate. There was a significant interaction of year with zone (Table 3 4). For UF slopes, vegetation cover declined from 2002 to 2012, but no t in all zones (Figure 3 4 ). There was significantly less total cover in the upper zone in 2010 (p=0. 00 20) and 2012 (p=0. 00 44) compared to 2002. There was, however, no difference by year in lower or middle zone s. In 2002, total vegetation cover was significantly less in the lower zone than the upper zone (p=0.0443) but indistinguishable from the middle zone (p=1.00). The total cover in the three zones did not differ significantly in 2010 or 2012 There was no dif ference in total vegetation cover at sites in areas open and closed to sport hunting (Table 3 2) The mean cover of functional groups and A. stricta var ied depending on zone and was consistent across years (Figure 3 5 ; Table 3 4 ). Forb cover in the upper zone was significantly less than cover in the lower zone ( p=0.000 6) and the middle zone ( p=0.0002 ). Grass cover was significantly less in the lower zone than the middle (p<0.0001) or upper zone (p<0.0001). The same relationship was observed for A. stricta with less cover in the lower zone than the middle (p<0.0001) and upper zones (p<0.0001). In contrast, woody species cover was highest in the lower zone and differed significantly (p=0.0010) from that of the middle zone. Woody cover in the upper zone was a lso greater than the middle zone, although this difference was marginally insignificant (p=0.06 21). Only cover of woody species and A. stricta differed during the ten year s of this study. Cover of forbs and grasses did not change significantly through time or by sport hunting access
34 (Table 3 4). Woody species cover increased significantly from 2002 to 2012 (p=0.0567) (Figure 3 6 ). A. stricta cover decreased significantly from 2002 to 2012 (p=0.0310) ( Figure 3 7 ). Are hogs preferentially disturbing a particu lar zone of the slope? Feral hog disturbance was more extensive in the lower, wetter zone of seepage slopes. The percentage disturbed differed significantl y among the three zones (Table 3 1) and consistently across years. Significantly more plots in the lower zone (50.92%) were disturbed when compared to the middle (11.02%; p<0.0001) and upper zones (5.07%; p<0.0001). The percentage disturbed in the middle and upper zo nes did not differ (p=0.1971). Does hog removal or do environmental variables account fo r changes in hog disturbance? When data from UF and Eglin AFB were combined, there was no correlation between percentage of plots disturbed by hogs and number of hogs removed annually (R 2 =0.0008; p=0.9111) or with the number of hogs removed the previous ye ar (R 2 =0.0004; p=0.8149) (Table 3 3). Total vegetation cover at sites, including those monitored by Eglin managers, was found to have a weak, but significant negative correlation (R 2 = 0.0558; p=0.0004) with total rainfall from the previous 11 months. Total vegetation was weakly correlated (R 2 =0.0473; p=0.0034) to time since burn. The relationship between the percent of hog disturbed plots and total rainfall varied for the four time periods that were investigated. There was no correlation (R 2 =0.0061; p=0.250 1) between the percent of hog disturbed plots and total rainfall from the previous 11 months, from July through M ay of the sampling year (Table 3 3). Similarly, there was no relationship ( R 2 =0 0036 ; p=0.3816) between the percent disturbed and total rainfal l during the five months prior to sampling, January through May. The percentage of hog disturbed plots also had no relationship (R 2 =0. 0007; p=0.6951 ) with total rainfall from March to June of the previous year, around the time of spring breeding. There was however, a
35 significant relationship (R 2 =0. 0364; p=0.0047 ) observed with total rainfall from August through November of the previous year, around the time of fall breeding The number of hogs removed was not correlated (R 2 =0.0335; p=0.9498) with rainfall during this fall time period. The number of hogs removed from Eglin AFB annually was more strongly correlated (R 2 =0.7865; p=0.0635) to rainfall during the previous March through June. Percent cover of hog damage in plots monitored by Eglin AFB showed no si gnificant relationship with time since burn (R 2 =0.0043; p=0.4316). The foliar cover of woody species also had a significant positive relationship (R 2 =0.1484; p<0.0001) with time since burn. A. stricta cover was not related (R 2 =0.0081; p=0.2530) to time sin ce burn. The multiple regression revealed percent hog disturbed plots had the most significant influence (p=0.0003) on total vegetation cover followed by rainfall totals from the previous fall months (p=0.0054) and time since burn (p=0.0208). Discussion Ha s the extent of hog damage and its effect on foliar cover changed from 2002 2012? This chapter offers insight concerning hog disturbance and vegetation dynamics in seepage slopes on Eglin AFB. There was fluctuation in the percentage of disturbed plots over this ten year period. Engeman et al. (2007) attributed the steady decline in hog disturbance from 2003 to 2005 to hog trapping. However, our data indicates hog disturbance was lower in 2002 and increased the year hog trapping was initiated In fact since hog trapping began, there were periods of increased hog disturbance. Engeman et al. (2007) also indicated that there was more hog disturbance in areas closed to hog hunting. On the contrary, our results indicate a higher percentage of hog disturbed plots i n areas open to hunting. This discrepancy may be due to difference in methods and/or sampling plots between the UF and internal Eglin monitoring efforts. Our results do however indicate that more extensive monitoring is necessary to draw conclusions on the effects of hog trapping on hog disturbance
36 There is potential for hog disturbance to facilitate an increase in woody vegetation cover. There was an increase in woody cover in the lower and upper zones where shrubs can encroach from the bordering baygall and uplands respectively in the absence of fire. We found that hog disturbance that disrupts the continuity of the herbaceous ground cover has the potential to inhibit fire spread (Chapter 2) because t he horizontal continuity of fuels can determine the ext ent of fire spread (Brooks et al 2004; Nader et al. 2007). If hogs migrate following available food, recently disturbed areas are left to recover in terms of foliar cover T here may be more potential for hog disturbance to reduce fire spread in places whe re prescribed fire is infrequent. If vegetation has time to recover pre disturbance foliar cover prior to prescribed fire, then it is more likely that the fire will spread through the slope. While if hog disturbance occurs too frequent ly for foliar cover t o recover prior to a prescribed burn, then fire spread would be less likely, providing opportunity for increased woody vegetation cover. Although there was no change in total foliar cover of grasses, there was a decline in foliar cover of A. stricta A. st ric t a does not recover from intense soil disturbance due to a shallow root system (Outcalt & Lewis 1990). As a disturbance sensitive species, it may be unable to recover with continued hog disturbance. Chapter 5 indicates frequent experimental disturbance results in a decline in A. stricta cover. In seepage slopes on Eglin AFB, more ruderal species, including Dichanthelium spp., may recolonize disturbed areas more quickly and contribute to the maintenance of grass cover (Chapter 5 ). This decline is important because A. stricta acts as a f ine fuel source to carry fire (Noss 1989; Haywood & Harris 1999). Further monitoring is necessary to determine if this decline is continuing. There was a significant difference in total forb and gras s cover by position on slope. We found less forb cover in the upper zone than the other two zones, while grass cover in the lower
37 zone was less than the other two zones. Thi s represent s the hydrologic gradient (Clewell et al. 2009) inherent to the slopes The upper zone s of seepage slope are generally drier and dominated by characteristic grass species including A stricta C aromaticum and Muhlenbergia expansa (L am.) Trin (Cutover muhly) T he lower zones have longer hydroperiods and are typically dominat ed by forb species including Sarracenia spp., D r osera spp., Xyris spp., Eriocaulon spp., Pleea tenuifolia Michx. ( Rush F eatherling) and Lophiola aurea Ker Gawl (Golden C rest) (FNAI 2010). Are hogs preferentially disturbing a particular zone of the slope? Hogs commonly forage in and disturb wet or mesic habitats ( Bratton 1975 ; Wood & Brenneman 1980; Saunders & Kay, 1991, Mitchell & Mayer 1997; Welander 2001 ; Chavarria et al. 2007 ) where it is easier to root and food is abundant ( Mitchell et al. 2007a). Our research indicates that h ogs are preferentially disturbing areas in the lower, wetter portions of the seepage slopes closer to the seepage streams Th is result suggest s that degradation associated with hog disturbance, including erosion (Bratton 1975; Howe & Bratton 1976), plant mortality (Bratton 1974; Bratton 1975; Johnson 2001), and increased plant available nitrogen resulting from excrement deposition (Ruess & McNaug hton 1987), is not uniform across the landscape. Preference of hogs to root in this portion of seepage slopes has serious implications for the persistence of listed plants (e.g. Sarracenia spp. ) that inhabit the lower zone Engeman et al. (2007) found hog disturbance to be negatively correlated with the presence of Sarracenia leucophylla Raf. and Sarracenia flava L. among other species of interest This lower zone is also important refugia for rare and endemic amphibians (e.g. Hyla andersonii (Pine Barren s tree frog) Eumeces anthracinus pluvialis (Southern Coal Skink), and Rana okaloosae (Florida bog frog) ) (Enge 1997). If hog impacts are
38 to be managed in seepage slopes on Eglin AFB, it is important to focus on protecting the lower zone microhabitat ( Mitc hell et al. 2007a). Does hog removal or do environmental variables account for changes in hog disturbance? Although the number of hogs remov ed each year from Eglin AFB fluctuated, there was no correlation with the percent of disturbed plots surveyed that y ear or the subsequent year. The number of hogs trapped would not affect the amount of hog disturbance if there were a density dependent increase in recruitment ( Hanson et al. 2009 ) but this was not quantified in our study. It appears that other f actors in addition to hog trapping may be driving the changes in the amount of hog disturbance E xtrinsic variations in food availability (i.e., variation in food availability due to environmental factors not related to the effects of density dependent animal fo raging) have been shown to influence hog population dynamics. Rainfall, an extrinsic factor, can drive forage biomass and thus hog abundance (Choquenot 1998; Sweitzer et al. 2000). In our study, rainfall totals from the months of the previous fall were sig nificantly correlated with the number of hog disturbed plots. In Florida, f eral hogs can breed throughout the year, but breeding peaks in the spring and fall (Gi uliano 2009). Greater precipitation during peak breeding months may contribute to greater repro ductive success and correspond to an increase in hog disturbance the following year, after sows farrow. The correlation between rainfall during the previous spring months and the number of hogs removed from Eglin AFB by trapping efforts also supports this hypothesis These results are consistent with research that found higher reproductive rates in hog populations in response to abundant forage (Coblentz & Barber 1987), while drought and limited forage delayed breeding, reduced reproductive rates (Giles 198 0), and increased mortality (Woodall 1983; Massei et al. 1997). Lower densities of hogs then corresponded to less hog disturbance (Hone 1995; Sweitzer 1998; Hone 2002; Sweitzer & Van Vuren 2002).
39 Patterns in hog foraging and disturbance can also reflect m ovement between communit y type s following mast crop (Howe & Bratton 1976, Bratton et al. 1982). Hogs prefer to feed on hard mast when it is available (Henry & Conley 1972). Heavy rooting for tubers and roots may increase during years of poor mast because p referred acorns are not available (Bratton et al., 1982). After extensive searching, w e concluded that information concerning the timing of hard mast ( Quercus spp.) on Eglin AFB has not been recorded. It is known that P palustris experienced the first large mast event since 1996 in the winter of 2011 ( pers. comm. 2012 Eglin AFB land manager). Roots of longleaf pine seedlings are a preferred food of hogs ( Wood & Brenneman 1980) This mast event in 2011 increased availability of see dlings in the upland habitat and may have contributed to the decline in hog disturbance observed in the seepage slopes in 2012. Bratton et al. (1982) indicate that v egetation may have time to recover from heavy rooting during good mast years when acorns ar e plentiful, but will be rooted again when mast crops fail Longer trend data including data on mast crops, is needed in order to identify a stronger trend in the amount of hog disturbance and its potential drivers. The percent cover of hog disturbance w as unrelated to time since burn, although a variety of research indicates that herbivores (Vinton et al. 1993; Coppedge & Shaw 1998; Vermeire et al. 2004) as well as omnivores (e.g. Chen caerulescens ( Snow G eese )) (Brennan et al. 2005) may selectively forage in recently burned vegetation This could be due to preference for plants with increased nutritional value (Lemon 1946; Pearson et al. 1972; Shindler et al. 2004) and/or new growth resulting from the reduction of shade (Copeland et al. 2002). Lack of correlation between hog rooting and seepage slope burning is potentially a result of one of four scenarios: 1) hogs show no preference to forage in areas that burned, possibly because the ground is drier and more difficult for rooting; 2) ho gs are foraging for a food source (e.g., invertebrate) that is
40 unaffected by fire; 3) fire spread in seepage slopes is not extensive enough for differences to be significant; or conversely 4 ) fire spread in seepage slopes is frequent, uniform, and complete leaving no unbur ned areas from which to differ. Our results indicated that woody vegetation cover had a significant positive correlation with time since burn W hile fire halts woody recruitment, it is not always effective at reducing the cover of woody s pecies once they become dominate in seepage slopes. Perennial shrubs resprout and persist in communities adapted to frequent fire. Mortality from fire only occurs in the very small shrubs (Olson & Platt 1995). Low mortality and resilient resprouting of est ablished shrubs after burning may necessitate further management actions to reduce shrub densities (Olson & Platt 1995; Drewa et al. 2002). T hese data also indicate that t otal vegetation cover is a function of intensity and frequency of feral hog disturban ce and to a lesser extent, rainfall, and time since burn. Over the past five years, Eglin AFB land managers have focused on increasing fire frequency in seepage slope areas. Yet even with almost annual prescribed fire ( pers. comm. 2013 Eglin AFB land manag er) there was only weak correlation between total vegetation cover and time since burn. While it is likely that many factors affect total vegetation cover, pe rcent of disturbed plots was relatively the most important factor in our analysis. Recovery rate of vegetation was not quantified, but is none the less important to total vegetation cover. Recovery via perennial regrowth, vegetative spread, seedling establishment, and germination requires favorable environmental conditions, such as adequate soil moist ure (Outcalt et al. 1999). Even though there is less hog disturbance in th e drier portion of the slopes, there is lower soil moisture and thus less recruitment, recovery, and ultimately less cover. This region experienced exceptional drought conditions (D4 ) in June of 2011 and was abnormally dry in June of 2012 (D0) ( National Drought
41 Mitigation Center 2013 ). The middle and lower zones of the slope experience longer hydroperiods and may have had an advantage recovering due to increased soil moisture. No diff erence in total cover by year in the lower or middle zones supports this idea. Significantly less total vegetation cover in the lower zone reflects hog preference for the wetter portions of the slopes. The total cover in the three zones did not differ sig nificantly some years One explanation could be that in years of heavier hog use, disturbance is more uniform across the slope, while in years of light hog use, there is not enough disturbance for the difference to be significant. Although hog use as measu red by percent plot disturbed was greater in areas open to hunting, this was not reflected in average total vegetation cover. Summary Hog management must involve long term strategies As with any ecological data with inter annual variation, long term data are necessary (Krebs 1991; Troeng & Rankin 2005) because there is the risk of drawing erroneous conclusions from short term data Our results indicate that monitoring seepage slope vegetation and other environmental variables must continue in order to unde rstand the variability of feral hog disturbance and long term consequences. Annual adjustments should be made to management plans depending on environmental factors so that hog trapping is increased in years of: 1) mast failure when hogs are forced to for age for other food sources (Bratton et al., 1982); or 2) high rainfall when hog densities have the potential to increase (Coblentz & Barber 1987). We recommend that the lower portions of seepage slopes be protected (e.g., using fence exclosures) because fe ral hogs target this area where rare, endemic flora and fauna are found D ue to the increase in woody vegetation cover we observed, t here should also be a concerted effort to ensure fire spreads downslope into seepage slope areas where fire may not spread unassisted Larger, more established woody vegetation may necessitate cut ting or clear ing by hand. A ctive management must continue and
42 address multiple facets in order to prevent further degradation of ecosystem structure and function.
43 Table 3 1. Repeated measures ANOVA results for the percentage of hog disturbed plots over time by zone on Eglin AFB from 2002 to 2012. Plots Percentage disturbed (%) 2002 2010 2012 Variable Num df df F P UF 27.88 30.40 15.98 Year 2 122 5.54 p=0.0050 Zone 2 122 47.63 p<0.0001 Year x Zone 4 122 0.46 p=0.7659 2007 2010 2011 Eglin AFB 31.13 59.60 57.51 Year 2 38 32.76 p<0.0001 Table 3 2. Repeated measures ANOVA results for the percentage of hog disturbed plots over time by hunting access on Eglin AFB from 2002 to 2012. Plots Open Hunting Closed Hunting Variable Num df df F P UF 38.25 16.67 Access 1 122 5.37 p=0.0222 Year x Access 2 122 1.98 p=0.1430 Eglin AFB 51.77 47.30 Access 1 38 0.45 p=0.5051 Year x Access 2 38 1.00 p=0.3781 Table 3 3. The number of hogs removed from Eglin AFB from October 1 of the previous year to September 30 of that calendar year. The percentage of hog disturbed plots and rainfall totals are also indicated. Year Hogs Removed by Trapping Percentage Hog Disturbed Plot s (%) Rainfall Previous July May (in) Rainfall January May (in) Rainfall Previous March June (in) Rainfall Previous August November (in) 2002 0 27.88 UF 32.92 16.37 18.86 11.58 2003 0 47.62 Eglin 59.63 26.04 29.22 20.56 2004 439 36.11 Eglin 50.05 13.66 35.13 15.47 2005 189 34.92 Eglin 80.37 45.07 26.34 26.72 2007 68 31.13 Eglin 35.77 10.21 19.95 19.93 2010 261 44.75 Both 76.25 27.74 18.35 29.54 2011 106 57.51 Eglin 38.58 15.38 17.22 18.26 2012 110 15.98 UF 54.88 20.65 10.6 14.15
44 Table 3 4. Repeated measures ANOVA results for the cover of total vegetation, functional groups, and A. stricta by zone and access on Eglin AFB in 2002, 2010, and 2012. Percent Mean Cover (%) 2002 2010 2012 Variable Num df df F P Total vegetation 78.75 74.57 77.48 Year 2 122 2.59 p=0.0794 Zone 2 122 5.17 p=0.0070 Year x Zone 4 122 3.76 p=0.0064 Access 1 122 0.04 p=0.8405 Year x Access 2 122 0.08 p=0.9263 Forb 23.69 18.90 22.97 Year 2 122 1.86 p=0.1593 Zone 2 122 11.25 p<0.0001 Year x Zone 4 122 0.69 p=0.6017 Access 1 122 0.33 p=0.5652 Year x Access 2 122 0.32 p=0.7288 Grass 44.09 38.39 39.98 Year 2 122 1.56 p=0.2144 Zone 2 122 21.34 p<0.0001 Year x Zone 4 122 0.50 p=0.7366 Access 1 122 0.42 p=0.5169 Year x Access 2 122 0.78 p=0.4592 Woody 16.53 19.31 21.03 Year 2 122 2.94 p=0.0567 Zone 2 122 6.85 p=0.0015 Year x Zone 4 122 1.09 p=0.3637 Access 1 122 3.35 p=0.0696 Year x Access 2 122 0.26 p=0.7685 Aristida stricta 25.91 16.89 21.63 Year 2 122 3.58 p=0.0310 Zone 2 122 56.34 p<0.0001 Year x Zone 4 122 1.84 p=0.1249 Access 1 122 0.39 p=0.5335 Year x Access 2 122 0.53 p=0.5908
45 Figure 3 1. Topographic relationship between seepage slope, upland pine, and baygall communities including position of zones on seepage slopes. Figure 3 2 The location of seepage slope sites on Eglin AFB.
46 Figure 3 3. Percent hog disturbed plots from 2002 to 2012. Analysis included both UF and Eglin AFB data. Figure 3 4 Total foliar cover by year. Bars are mean foliar cover 1 SE the P values can be found in Table 3 4, and different letters above bars indicate significant difference ( = 0.05) in foliar cover following Tukey Kramer post hoc tests.
47 Figure 3 5 Foliar cover of functional groups and A. stricta by zone (i.e., position on slope). Bars are mean foliar cover 1 SE, the P values can be found in Table 3 4 and different letters above bars indicate significant difference ( = 0.05) between zones following Tukey Kramer post hoc tests. Figure 3 6 Fo liar cover of woody species through time. Bars are mean foliar cover 1 SE and the P values can be found in Table 3 4
48 Figure 3 7 Foliar cover of Aristida stricta through time. Bars are mean foliar cover 1 SE and the P value s can be found in Table 3 4
49 CHAPTER 4 EFFECTS OF INTENSITY OF FERAL HOG DISUTRBANCE ON VEGETATION DYNAMICS Backgrou n d Herbivores have the potential to greatly modify their habitat (Bratton 1975; Tardiff & Stanford 1998) in a manner that can be either beneficial or detrimental to ecosystem function and structure depending on the resident plant species Herbivory can benefit some plant species when dominant plants are targeted and the competitive interaction between plants is reduced. Thus, competitive dominance is pr evented and species diversity enhanced (Lubchenco 1978) Herbivores can also increase spatial and temporal variation in the resources with localized excrement deposition and soil disturbances resulting from digging (Tardiff & Stanford 1998), trampling, and wallowing. These activities create competition reduced gaps that can also have a positive effect on seedling survival (Silvertown & Bullock 2003). On the contrary, herbivory targeting competitively inferior species or that is particularly intense can redu ce species diversity. This disturbance may result in the dominance of only a few tolerant species (Lubchenco 1978). Selective herbivory can also reduce preferred plant species (Bratton 1974; Recher & Clark 1974 ; Challies 1975; Howe & Bratton 1976 ). Dependi ng on the system, species, and herbivory regime, herbivore effects on plant community composition and diversity can vary greatly ( Maschinski & Whitham 1989 ) The modern, Pinus p alustris Mill. ( longleaf pine ) savanna ecosystem of the southeastern United States came into existence in the early Pleistocene, approximately 2 million Y BP. Plant species evolved under grazing and browsing disturbance from megaherbivore including horses, tapirs, peccaries, proboscidea ns, ground sloth, and various ruminants. Modern characteristics of the P p alustris (i.e., masting, strong taproot and wood) can be interpreted as adaptations to defend against these megaherbivore s (Means 2006). Other modern vegetative characteristics of
50 this system still reflect coevolution with megafauna, such as seeds of herbaceous species being viable after being digested by livestock (Janzen 1984). However, disturbance resulting from large herbivores ended after the mass extinction of these animals at the end of the Pleistocene, approximately 11,000 Y BP (Grayson & Meltzer 2002). The Bison bison (American buffalo) migrate d into the Gulf Coastal Plain in the late sixteenth century, but populations were extirpated in the early ninet eenth century (Rostlund 1960). After the Pleistocene extinction of the megafauna, it is likely that vegetation experienced an e cological release greater than the influence extant megafauna had on savanna plant characteristics or distributions (Means 2006). One thing that has no t changed through time in this system is the dependence of the P. palustris grasslands on fire (Watts 1971; Komarek 1965, 1968, 1974; Myers 1990; Stout & Mari on 1993; Means 2006). Lightning ignited fire was a major component of the climate, enough so to fr equently burn large expanses and to shape the evolution of pyrophytic vegetation long before aboriginal peoples (Komarek 1965). Paleoindians were not widespread in the southeast until 8,500 5,000 BP. It can be in ferred that by that time, human ignited fire contributed to an increased frequency of fires preventing the pine savanna from succeeding to broad leaved forest (Watts 1971). Native Americans burned the forests frequently for cultivation, hunting reducing insects and defense (Williams 1992 ; Kay 2007 ). Vegetation has evolved to recover rapidly after the disturbance of high frequency (1 3 years), low intens ity fire. Some examples of fire evolved traits include: the early growth of longleaf pine in a grass stage (Keeley & Zedler 1998); fire induced germ ination (Wiggers 2011), vigorous resprouting of shrubs following fire (Olson & Platt 1995); and increased flowering, seed production, and germination of dominant matrix grasses after growing season fire (Shepherd et al. 2012). It seems that fire has
51 been t he principal selective agent in the evolution of longleaf pine savannas of the southeast (Frost et al. 1986). Introduced herbivores, including feral hogs, present substantial potential for degradation to native plant communities resulting from herbivory an d altered disturbance regimes (Singer et al. 1984; Van Vuren & Coblentz 1987 ; Hobbs & Huennek e 1992; Kotanen 1995). We propose hogs could either replace the role of now extirpated megaherbivores or re pre sent a novel disturbance that results in ecosystem change In this chapter, we investigate the short term floristic response after different intensities of hog disturbance in seepage slopes We examine the effect of dist u r bance intensity on cover, richness, and species composition over a three year period We hypothesize that over the short term high hog disturbance intensity reduces cover, species richness and alters species composition when compared to undisturbed vegetation. Methods Study A rea Eglin AFB is located in the western Florida Panhandle, Walton County, USA S eepage slope s ites are restricted to the loamy sand soils of the eastern portion of the AFB It is here where g ently rolling erosional hills have been cut into the remains of a Pliocene Pleistocene age delta plane. T hese dendritic drainage systems have relief of up to about 30 m and areas of steeper slope along creeks (NRCS 1989 ). The relatively higher clay content of these soils compared to the deep sands of xeric P palustris sandhills results in more mesic conditions. This mesic upland community is classified as Upland Pine Forest (FNAI 1990). Seepage slopes are grass and sedge dominated wet prairies locate d on hillsides in the upland pine community (Figure 4 1) ( FNAI 1990 ). The nutrie nt poor soils are kept saturated by the downslope seepage of groundwater above a n impermeable layer of clay, organic matter, or
52 rock. Seeps occur along hillsides of these dendritic systems of streams where the impermeable layer intersects the hillside. Dur ing times of drought, seepage slopes may dry out due to this dependence on groundwater (FNAI 2010). A diverse suite of herbaceous species inhabit this wet prairie community. Aristida stricta Michx. (Wiregrass) Ctenium aromaticum (Walter) A. Wood (Toothac hegrass) Muhlenbergia expansa (Lam.) Trin. (Cutover M uhly ) and Rhexia spp. (M eadowbea u ties) can be found upslope in the drier soils while Rhy n chospora sp p. (B eaksedges) Xyris spp. (Y elloweyed G rasses) Eriocaulon spp. (P ipeworts) and Lycopodiella spp. (C lub M osses) inhabit the wetter portions of the slopes (FNAI 2010). Carnivorous plants that acquire nutrients by trapping insects (Schnell 1976) can be abundant, including Sarracenia spp. (P itcherplants) Drosera spp. (S undews) Pinguicula spp. (B utterworts) and Utricularia spp. (B ladderworts). Shrubs are maintain ed at a reduced height by frequent fire that burns down slope from the uplands Cliftonia monophylla (Lam.) Sarg. (Black T iti) and other shrubs invade from the shrub bogs or baygalls alon g the seepage creek, outcompeting the heliophytic, herbaceous vegetation in the absence of frequent fire (FNAI 2010). Data Col l ection Research q uestions 1, 2, and 3. Does the difference in foliar cover between disturbance intensities persist over time? Doe s species richness differ in areas of varying disturbance intensity? Does species composition differ in areas of varying disturbance intensity? The intensity of feral hog foraging and wallowing activity varies within seepage slopes on Eglin AFB. In 2010, 20 independent, replicate seepage slope s ites with the varying levels of hog disturbance were randomly chosen on Eglin AFB (Figure 4 2) Three hog disturbance intensi ties were identified based on percent of uprooted vegetation and total ve getation cover. T he intensities were: 1) undisturbed (control), with less than 3% uprooted and total cover of 75
53 100%; 2) moderate, with <50% uprooted and total cover of 50 75%; and 3) high, with >50% uprooted and total cover of 0 50%. Potential plot locations were restric ted to flat, open slopes to minimize the influence of aspect. Plots were placed within the middle zone of the slope to minimize the effect of the moisture gradient and restricted to the area of the slope with different intensities of hog damage within clos e proximity. At each seepage slope site, two locations were randomly selected for placement of wire fence exclosures ( 6m in circumference and 1m in height) w ithin each disturbance intensity (control /undisturbed moderate, and high ) Exclosures prevent ed su bsequent hog damage. Within each exclosure one 1 m 2 permanent plot was established. Adjacent to each exclosure (approximately .5 m downslope) an open 1 m 2 plot was permanently marked (Figure 4 3) This resulted in approximately six open and six exclosure plots at each site. The two subsamples accounted for within site variability. Two of the 20 sites only contained two disturbance intensities due to the nature of hog disturbance at the time of plot initiation Between 2010 and 2013, eight exclosure plots w ere compromised by hog disturbance and removed from the study. In August of 2010 and 2012, total foliar cover and cover of functional groups was determined usi ng a modified Daubenmire scale (0= 0%, 0.5 5 = 3%, 5 10 = 8%, 10 15 = 13%, 15 25 = 21%, 25 50 = 3 8%, 50 75 = 68%, 75 95 = 88%, 95 100 = 98 % ) within 1 m 2 plots inside and outside exclosures. T otal number o f species present was recorded. Due to high species richness, percent cover by species was estimated within of each 1 m 2 plots. Species nomenclatur e was based on Wunderlin & Hansen (2008). In August of 2013, total foliar cover only was estimated in 1 m 2 plots inside and outside exclosures. Wildlife cameras with motion sensors recorded hog activity at 21 seepage slopes (Figure 4 4) Hog observations were tallied independent of size of group or length of time spent foraging
54 at site. Cameras were checked every two to three months to download images and replace batteries. Over the course of the four year study, cameras were either stole n (7), burned in prescribed fire (1), or eventually became inoperable after prolonged exposure to the elements (4). These losses, in combination with inaccessibility of areas due to military operations and removal of cameras for prescribed burns, resulted in limited and intermittent data acquisition. This data represents a conservative estimate of hog activity. Data A nalyses Research q uestions 1 and 2. Does the difference in foliar cover between disturbance intensities persist over time? Does species richne ss differ in areas of varying disturbance intensity? The experimental followed a randomized, split plot repeated measure design. Within each seepage slope site initial intensity of hog disturbance (control, moderate, and h igh ) represented whole plots and exposure to subsequent hog disturbance (exclosure vs open) were split plots. Significance of main effects and interactions ( = 0.10) were evaluated using a repeated measure one way ANOVA ( PROC GLIMMIX ; SAS 9.3; SAS Institute 2011 ) to determine significance of recovery with and without potenti al for further hog disturbance. Means were arcsine square root transformed when necessary to meet assumptions of ANOVA (Zar 1998). Year, intensity, and exclosure were the main effects. Site w as considered a random effect. Tukey Kramer tests were used for post hoc comparisons ( = 0.05) to analyze differences between means. The average number of hog occurrences per month was calculated from the camera data. Research q uestion 3. Does species composition differ in areas of varying disturbance intensity? NMS was used to better understand temporal changes in species of interest with respect to disturbance intensity and edaphic conditions We removed all species that occurred in less than 5% of the plots and grouped all And ropogon spp. (Bluestems) Asteraceae spp. (Asters)
55 Eriocaulon spp ., Rhexia spp., Rhynchospora spp., Sarracenia spp ., Scleria spp. (Nutrushes) and Xyris spp. Moderate disturbance intensity plots were removed to reduce overlap with high intensity and control plots. A Bray Curtis distance measure was used for the ordination in PC ORD. Monte Carlo tests were used to determine if the ordination results were significantly different from random data. Then we used MRPP to test for differences within and betw een disturbance intensity groups ( McCune & Grace 2002). Results Foliar Cover Does the difference in foliar cover between disturbance intensities persist over time? Hog disturbance intensity significantly altered v egetation cover over time (Figure 4 5 ; Tabl e 4 1). Foliar cover in high disturbance plots (inside and outside exclosures) increased significantly (p<0.0001) from 2010 to 2013. There was no significant change in cover in control (p=0.8 185 ) or moderate (p=0.999 4 ) disturbance plots. Total foliar cover differed among all intensities in 2010 (Year 1): high intensity plots had less cover than control plots (p<0.0001) and moderate disturbance plots (p<0.0001); while control plots had greater total foliar cover than moderate (p<0.0001). By 2013, moderate di sturbance plots still differed significantly from control (p=0.0 670 ) and high (p=.0609) disturbance plots. C over in high disturbance plots also continued to be less than control (p<0.0001) plots. Initially in 2010, there was no difference (p=0.51 ) in foliar cover inside ( 68.5 5 %) and outside (64.22%) exc losu res (Figure 4 6) By 2012 foliar cover increased both inside (p<0.0001 ; 83.02% ) and outside (p=0.0011 ; 73.04% ) exclosures. However by 2013, continued, albeit less intense and frequent, hog activity resulted in significantly greater (p<0.0001) total foliar cover inside ( 77.19 %) compared to outside ( 68.39 %) exclosures Cover of woody species increased in plots inside and outside exclosures in all disturbance intensi ti es from 2010 to 2012 (Figure 4 7 ; T able 4 1). Initially, high disturbance plots did not
56 differ from moderate disturbance plots (p=0.9952), but had less woody cover than control plots (p=0.0553). Over time, the only significant change (p<0.0001) took place in high disturbance plots, where co ver of woody species more than doubled. There were no significant differences between the three disturbance intensities by 2012: high and control (p=0.9995); high and moderate (p=0.9060); and control and moderate (p=0.9781). There was also no significant d ifference between plots inside and outside exc losures. Intensity of disturbance was a determining factor i n foliar cover of forbs (Table 4 1). Control plots had higher forb cover (34.01%) than high (19.34%; p<0.0001) and moderate (25.04%; p=0.0207) disturbance plots. There was no difference (p=0.1573) between high and moderate disturbance plots. Forb cover inside and outside exclosures did not differ significantly. Foliar cover of grass species changed over time with d isturbance intensity (Figure 4 8 ; Table 4 1). High disturbance plots (inside and outside exclosures combined) had less grass cover than control (p<0.0001) and moderate ( p=0.0002) disturbance plots in 2010. Control plots did not differ from moderate (p=0.7272) disturbance plots By 2012, high disturbance plots still had significantly less grass cover than control (p=0.0404) and moderate (p=0.0139) disturbance plots There was still no difference (p=0.9991) between control and moderate disturbance plots. Initially in 2010, there was no diff erence (p=0.4491) in grass cover inside (35.71 %) and outside ( 32.08 %) exc losures (Figure 4 9). By 2012, foliar cover increased both inside (p<0.0001) and outside (p<0.0001) exclosures. This increase in grass cover resulted in significantly greater (p=0.000 6 ) grass cover inside ( 53.27% ) compared to outside ( 43.97% ) exclosures. The cover of A. stricta differed by disturbance intensity with no effect of time (Table 4 1 ). High disturbance plots had the lowest cover (7.58%) of A. stricta The high disturbance plots had significantly less cover than the control (22.42%; p<0.0001) and moderate (16.51%;
57 p=0.0224) disturbance plots, while control and moderate disturbance plots did not differ (p=0.1602). Exclosures had greater A. stricta cover (16.27%; Table 3 1) th an outside exclosures (14.74%) throughout the study Richness Does species richness differ in areas of varying disturbance intensity? Species richness was dependent on time since disturbance, but not disturbance intensity (Table 4 1). There was no differen ce in species richness among disturbance intensities at the start of the study (high and control p=0.1617; high and moderate, p=0.7780; control and moderate, p=0.8711). The number of species per plot increased significantly from 2010 to 2012 for all inten sities (high, p<0.0001; moderate, p<0.0001; control p=0.0017). Still at the end of the study, no significant difference in richness was detected between the intensities (high and control p=1.0; high and moderate, p=0.9975; control and moderate, p=0.9925) There also was no difference between the species richness inside (15.16) and outside (15.36) exclosures. Camera Detection of Hog A ctivity Hog detection by the cameras fluctuated a great deal, but there was a decreasing trend from 2010 until the end of 2011 (Figure 4 10 ). The average number of hog occurrences detected by cameras was lower in the summer months. Overall, hog occurrences increas ed from 2012 to the highest numbers in the fall of 2013. Vegetation C omposition Does species composition differ in areas of varying disturbance intensity? MRPP indicated that there were significant differences in vegetation composition depending on soil ty pe when all seepage slope sites were included in the analysis (T= 8.1775; A= 0.0349; p<0.0001 ). The subsequent multivariate analyses focused on 10 sites located on the Leefield Stilson Loamy
58 Sands soil type. This soil type was selected in an attempt to redu ce variability and retain the largest sample size possible. NMS ordination and MRPP revealed differences in species composition. Vegetation composition at sites located on the Leefield Stilson Loamy Sands soil type differed between high disturbance and control plots when years were co mbined (T= 56.9861; A=0.3296; p< 0.000 1 ; F igure 4 11 ). A two dimensional ordination was optimal (stress = 6.1094) with axis 1 explaining 82.7% and axis 2 explaining 15.9% of the variation. Monte Carlo test determined that the ordination differed from random data (p=0.0164). Comparable results were observed when years were analyzed individually Both of these analyses had three dimensional solutions. When 2010 data was analyzed separately (stress=17.8216), axis 1 explained 11.9 %, axis 2 explained 24.4%, and axis 3 explained 37.6% of the variation observed. When 2012 data was analyzed separately (stress=19.0722), axis 1 explained 20.0%, axis 2 explained 20.6%, and axis 3 explained 31.5% of the variation observed. This analysis re vealed that the difference between disturbance intensities and separation between groups decreased over time as the high intensity disturbance recovered (2010: T= 16.6659, A=0.0616, p<0.0001 ; 2012: T = 8.8013, A=0.0284, p<0.0001 Figure 4 1 2 a and 4 1 2 b). D iscussion Does the difference in foliar cover between disturbance intensities persist over time? Three years of field observations suggest that heavily hog disturbed areas differ from undisturbed areas in both cover and composition. The most obvious change resulting from hog disturbance is loss of total foliar cover (Bratton 1974, 1975; Challies 1975; Cuevas et al. 2010). Even after three years of recovery inside exclosures, high intensity disturbance plots still have significantly less cover than moderatel y disturbed and control plots. Although theoretically successional communities maintained by frequent disturbance should recover more quickly than
59 "climax equivalents" (Baron 1982), these results indicate that more time is necessary before heavily disturbe d seepage slopes areas are indistinguishable in term s of total cover Total cover increased over time both inside and outside exclosures due to reduced hog foraging This region experienced moderate to exceptional drought conditions ( D1 D4) from October 20 10 to June 2012 ( National Drought Mitigation Center 2013 ). We discussed previously in Chapter 3 that rainfall drives forage biomass and can influence hog abundance (Choquenot 1998; Sweitzer et al. 2000). Drought and limited forage delays breeding, reduces reproductive rates (Giles 1980), and increases mortality (Woodall 1983; Massei et al. 1997). A drought and subsequent decline in hog numbers corresponds to less hog disturbance outside exclosures and likely explains the observed increase in cover outside exclosures. Our camera data also substantiates this idea with fewer hog occurrences (i.e. image s) from 2011 through 2012. However with an increase in hog activity and disturbance in 2013, o ur exclosure study showed that continued hog disturbance outside exclosures was enough to result in significantly lower total foliar cover compared to inside exlo sures. H og foraging has the potential to have different effects on woody cover at different scales. Uprooting vegetation, including woody species, reduces cover at the small scale. I n the absence of intense and/or frequent hog disturbance during our study, the cover of woody species increased significantly over time in heavily disturbed plots as vegetation recovered. A t the large scale, hog rooting reduces herbaceous cover that acts as fuel to carry fire (Chapter 2) Woody species then have an opportunity t o become established with reduced fire spread from the uplands, increasing woody cover at the large scale. This idea is discussed in further detail in Chapter 2 The effect of exclosure s on woody cover would likely have been significant given
60 more time or if hog disturbance had been heavier in seepage slopes in 2012 (Mitchell et al. 2007b; Taylor et al. 2011). Although hogs are omnivorous ( Taylor & Hellgren 1997 ), previous studies have found herbaceous vegetation can account for the greatest proportion of t he diet of feral hogs (Challies 1975; Everitt & Alaniz 1980; Coblentz & Baber 1987; Cuevas et al. 2010). While little is known about the diet of feral hogs in this system, our findings indicate that either direct consumption or indirect soil disturbance as sociated with rooting results in greater forb and grass cover in undisturbed areas. The observed increase in graminoid cover was greater inside the protection of exclosures where ruderal species (e.g. Dichanthelium spp.) quickly recolonize disturbed soil (Rodgers & Provencher 1999). In the absence of hog foraging, other studies have observed similar increases in herbaceous cover (Siemann et al. 2009; Cole et al. 2012). A ristida stricta is a dominant perennial bunchgrass in the longleaf pine system and cons idered a keystone species (Noss 1989). This native understory species is critical for carrying frequent, low intensity fire (Clewell 1989; Haywood & Harris 1999). Hog rooting has been found to have little impact on native perennial bunchgrasses in coastal meadows of California because these species are deep rooted and difficult to uproot ( Kotenan 1995; Cushman et al. 2004). However, A. stricta has a shallow root syste m and is notorious for not recolonizing after intense soil perturbations (Outcalt & Lewis 1 990; Huffman & Judd 1998). It was no surprise that A. stricta cover was significantly lower in high intensity than control plots. Although it may seem concerning that A. stricta cover was reduced in plots outside exclosures, th e difference between inside and outside exclosures was within one cover class and present from the beginning of the study. Again with 2012 being a particularly dry y ear, hog disturbance was negligi ble compared to years with greater rainfall (Chapter 2). Fut ure periods of heavy rainfall, like that of
61 2013, could potentially be conducive to hog rooting with greater effect on A. stricta cover and thus fire spread. Does species richness differ in areas of varying disturbance intensity? T he effects of hog rooting on diversity varies depending on plant community type. Hog disturbance results in increased plant species richness in coastal grasslands in California (K otanen 1995, 1997; Cushman et al. 2004), while it can reduce plant species richness in Grey Beach For ests of the Great Smoky Mountains National Park (Bratton 1974, 1975 ; Howe & Bratton 1976 ; Bratton et al. 1982), floodplain marshes of Central Florida (Arrington et al. 1999), and mixed pine hardwood forests in Texas (Siemann et al. 2009 ). The vegetati ve response depends on whether the community is adapted to that particular disturbance regime (Hobbs & Huenneke 1992 ; Doup et al, 2010 ). Similar to studies in other systems (Doup et al. 2010; Taylor et al. 2011; Cole et al. 2012), our study did not fi nd that intensity of hog rooting affected plant species richness. It is hypothesized that because seepage slope vegetation is subject to stress from periods of prolonged soil saturation, there is less potential for competitive exclusion ; thus soil disturba nce has a less pronounced effect on diversity (Huston 1979). Another possibility could be that t his study was too short term to detect this community level change (Taylor et al. 2011). The change we observed in richness was an increase over time with recov ery and increase in cover. More time may be necessary to detect an exclosure effect. Does species composition differ in areas of varying disturbance intensity? Our results from the NMS ordination analyses indicate significant difference s between: 1) plant assemblages on differing soil types within the same community; and 2) undisturbed (controls) assemblages and those that are recovering from high intensity hog disturbance. Although the difference between disturbance intensities and similarity within groups decreased over time as vegetation
62 recovered in high hog disturbance, the composition of these groups remained significantly different even when only one soil type was included. It may take more than three years without disturbance by hogs for heavily hog disturbed areas to succeed from an early successional state (Kirkman et al. 2004) It could be contested whether or not it is beyond the scope of this study to conclusively state that the difference in plant composition (and abiotic components) of disturbed and undisturbed areas is a direct result of hog disturbance. Siemann et al. (2009) suggest that hog disturbance is not random, but targets a particular biotic or abiotic component. T hus observational studies cannot quantify effects of hogs on vegetation because the area was likely inherently different from undisturbed areas. On the contrary, Arrington et al. (1999) indicate that years of observational data of vegetation composition sh owed areas of hog rooting were similar to undisturbed areas before rooting occurred. We followed similar protocol to Cuevas et al. (2010) and set up rooted plots in close proximity to undisturbed plots in an attempt to reduce abiotic and biotic variability Field observations over the course of this study validate the observations of Arrington et al. (1999) as undisturbed areas were converted to heavily rooted areas overnight, indicating little difference between the pre disturbance states of undisturbed an d heavily dist u r bed plots. Although domestic hogs ( Sus scrofa ) were first introduced over 400 years ago, n o record of large numbers of hogs on Eglin AFB exists prior to domestic herds that were free range until 1960 (Mayer & Brisbin 1991 ). Considering t hat hogs have been in this area for a substantial amount of time, there is no definitive way to determine if undisturbed areas were different from disturbed areas prior to the presence of hogs. Hogs could be acting as ecological engineers, as rooting i s kn own to exacerbate abiotic differences (e.g. increase depth of depression and water
63 retention), perpetuating biotic differences ( Cuevas et al. 2010). However this study did not discover that hog rooting created unique plant associations (e.g. Lachnanthes carol iana (Lam.) prair i es) as other research in peninsular Florida has indicated (Huffman & Judd 1998). Summary The resilience of a system describes the amount of disturbance that system can withstand and still remain in the same state ( Holling 1973 ). Ther e is not enough data to indicate that hog disturbance has pushed s eepage slopes on Eglin AFB past a threshold to an alternative stable state. Feral hogs could be ecologically equivalent to extinct me gaherbi vores in regards to soil disturbance just as it is hypothesized that an intermediate level of feral hog disturbance replaces activities of the extirpated Ursus arctos L. ( grizzly bear ) in California (Work 1993) Still, it is dangerous to interpret these results as if feral hog disturbance is ecologically benign. State changes in ecosystems occur when there are changes in the key variables that affect the stability of the system. The popular he u r istic of a ball in a cup can be used to illustrate this change, where changes in the shape of the cu p correspond to alteration of both stability (return time) and resilience (width of stability domain) (Gunderson 2000). Hog disturbance has likely reduced the resilience of the seepage slope community on Eglin AFB and thus it may be approaching a threshold even though a threshold has not yet been crossed.
64 Table 4 1. ANOVA results for the foliar cover of total vegetation, functional groups, and A. stricta and species richness by disturbance intensity and exclosure on Eglin AFB. Variable Num df df F P Total Vegetation Year 2 287.7 17.34 <0.0001 Intensity 2 55.54 75.12 <0.0001 Year x Intensity 4 304.5 14.60 <0.0001 Exclosure 1 80.51 27.02 <0.0001 Year x Exclosure 2 287.4 2.61 0.0800 Inten. x Exclos. 2 64.35 0.89 0.4143 Yr. x Inten. x Exclos. 4 201.8 0.75 0.5592 Woody Year 1 157.5 20.76 <0.0001 Intensity 2 54.30 1.75 0.1836 Year x Intensity 2 157.5 6.65 0.0017 Exclosure 1 152.7 1.47 0.2276 Year x Exclosure 1 152.7 1.26 0.2625 Inten. x Exclos. 2 152.7 1.39 0.2519 Yr. x Inten. x Exclos. 2 152.7 0.15 0.8571 Forb Year 1 156.9 0.68 0.4101 Intensity 2 53.88 10.61 0.0001 Year x Intensity 2 156.9 1.05 0.3522 Exclosure 1 152.2 0.0 0.9919 Year x Exclosure 1 152.2 1.04 0.3097 Inten. x Exclos. 2 152.2 0.21 0.8118 Yr. x Inten. x Exclos. 2 152.2 0.17 0.8442 Grass Year 1 159.8 72.53 <0.0001 Intensity 2 55.22 15.43 <0.0001 Year x Intensity 2 159.8 4.70 0.0103 Exclosure 1 153.9 14.5 0.0002 Year x Exclosure 1 153.9 2.73 0.1006 Inten. x Exclos. 2 153.9 0.42 0.6583 Yr. x Intens. x Exclos. 2 153.9 0.26 0.7735 Aristida stricta Year 1 153.0 0.22 0.6408 Intensity 2 53.06 10.44 0.0002 Year x Intensity 2 153.0 2.11 0.1250 Exclosure 1 151.0 4.54 0.0347 Year x Exclosure 1 151.0 0.24 0.6262 Inten. x Exclos. 2 151 1.4 0.2506 Yr. x Intens. x Exclos. 2 151 0.72 0.4891 Richness Year 1 158.0 89.34 <0.0001 Intensity 2 54.19 1.08 0.3471 Year x Intensity 2 158.0 3.47 0.0336 Exclosure 1 152.7 0.55 0.4582 Year x Exclosure 1 152.7 1.05 0.3063 Inten. x Exclos. 2 152.7 0.58 0.559 0 Yr. x Intens. x Exclos. 2 152.7 2.16 0.1193
65 Figure 4 1. Topographic relationship between seepage slope, upland pine, and baygall communities including position of zones on seepage slopes. Figure 4 2 The location of seepage slope sites on Eglin AFB
66 Figure 4 3. E xclosure and open plot pairs were located in the middle zone of the slope Figure 4 4. The location of seepage slope sites with motion detecting wildlife cameras.
67 Figure 4 5 Total foliar cover by intensity by year inside and outside exclosures on Eglin AFB. Bars are mean foliar cover 1 SE and different letters above bars indicate significant difference ( = 0.05) in foliar cover following Tukey Kramer post hoc tests. Figure 4 6. Total foliar cover by exclosure by year on Eglin AFB. Bars are mean foliar cover 1 SE and different letters above bars indicate significant difference ( = 0.05) in foliar cover following Tukey Kramer post hoc tests.
68 Figure 4 7 Foliar cover of woody species by intensity by year inside and outside exclosures on Eglin AFB. Bars are mean foliar cover 1 SE and different letters above bars indicate significant difference ( = 0.05) in foliar cover following Tukey Kramer post hoc tests Fig ure 4 8 Foliar cover of graminoid species by intensity by year inside and outside exclosures on Eglin AFB. Bars are mean foliar cover 1 SE and different letters above bars indicate significant difference ( = 0.05) in foliar cover following Tukey Kramer post hoc tests.
69 Figure 4 9. Foliar cover of graminoid species by exclosure by year on Eglin AFB. Bars are mean foliar cover 1 SE and different letters above bars indicate significant difference ( = 0.05) in foliar cover following Tukey Kramer post hoc tests. Figure 4 10 Average number of hog occurrences per month in seepage slopes with cameras on Eglin AFB from June 2010 through December 2013.
70 Figure 4 1 1 Two dimensional nonmetric multidimensional scaling ordination showing cover by species in high disturbance intensity and control plots on Eglin AFB. Samples from 2010 and 2012 are combined. Each point represents a plot and the distance between points is proportional to the difference in species composition Control High
71 A. B. Figure 4 1 2 Three dimensional nonmetric multidimensional ordination showing cover by species in high disturbance and control plots on Eglin AFB in A ) 2010 and B ) 2012. Each point represents a plot and the distance between points is proportional to the difference in species composition. By 2012, high disturbance intensity plots are recovering and separation between groups has decreased. Control High Control High
72 CHAPTER 5 EFFECTS OF EXPERIMENTAL SOIL DISTURBANCE ON VEGETATION DYNAMICS Backgrou n d Disturbances are environmental fluctuations or destructive events that cause abrupt changes in the community or population structure of natural communities. They can be biological or physical in nature and diff er in magnitude and distribution over time. The irregular nature of these events can produce heterogeneous or patchy effects (Sousa 1984; Pickett & White 1985; Moloney & Levin 1996). It is generally accepted that disturbances keep natural communities in a state of nonequilibrium, maintaining diversity by preventing competitive equilibrium. A dynamic balance is reached between the frequency of disturbance and rate of competitive displacem ent when competitive equilibrium is prevented and thus a stable level of diversity is maintained (Connell 1978; Huston 1979) The intermediate disturbance hypothesis (IDH) proposes that high diversity results from an intermediate disturbance regime. Longer lived species are lost when disturbance is too great or too often, while ruderal species are lost when disturbance is insignificant or too infrequent (Connell 1978; Petraitis et al. 1989; Reynolds 1993; Shea et al. 2004). Therefore the diversity patter n observed in disturbed environments is unimodal (Shea et al. 2004). An alternative hypothesis to the IDH is that diversity and richness are independent of dist urbance (Mackey & Currie 2001). Glitzenstein et al. (2003) hypothesized that in the longleaf pin e ecosystem the open, herbaceous ground cover is maintained with maximum species diversity when burning occurs as frequently as fuels allow The authors referred to this as the "Most Frequent Fire Hypothesis" (MFFH). This idea is supported by other studies that found frequent fire (i.e. one to two years interval) necessary to achieve reduced woody cover and dominance of diverse herbaceous
73 vegetation ( Walker & Peet 1983; Waldrop et al. 1992; Brockway & Lewis 1997; Haywood & Grelen 2000; Glitzenstein et al. 2003; Glitzenstein et al. 2012). Multiple types of disturbances with different characteristics often interact to make up the disturbance regime of a community (Collins 1987) and may result in complex patterns not described by the IDH (Collins & Barber 1985) or other hypothes es. Alternatively it may be the type and characteristics of the combined disturbances that determine effect on post disturbance community trajectory (Collins 1987; Paine et al. 1998; Platt et al. 2002). Grassland d isturbance regime s may include combinations of fire, grazing, and soil disturbance by animals where the effect of each disturbance is variable depending on scale, frequency, and intensity (Collins 1987 ). For example, one disturbance could reduce competitive superiority (e. g. grazing), while another could increase habitat heterogeneity (e.g. soil distu r bance). Still, community response is dependent upon the historical disturbance regime (Collins & Barber 1985). I t is ultimately divergence from the historical disturbance re gime that can leave native species ill adapted for successful recovery (Denslow 1980; Hobbs & Huenneke 1992) and may result in extirpation or even extinction (Connell 1978). All over the world, disturbance regimes are changing due to climate change, anthropogenic modifications, and biological invasions (Shea et al. 2004). Introduced species of both plants and animals have the potential to alter disturbance regimes. These speci es can increase or reduce the frequency and/or intensity of disturbances, like fire or erosion, which can quickly result in mortality of a significant portion of the community. Introduced feral hogs demonstrate that the invader can also be the actual agent of disturbance (Mack & D'Antonio 1998). Regardless of mechanism, these alterations to the disturbance regime can cause dramatic changes in composition and successional trajectory of the invaded community (Bratton 1975;
74 Mack & D'Antonio 1998). The purpose of Chapter 5 i s to assess the short term effect of four experimental soil disturbance regimes, meant to simulate hog disturbance, on community structure and composition in the seepage sl ope community We examine the effect of dist u r bance frequency and time since last disturbance on foliar cover, richness, and species composition over a four year period Methods Study A rea This study was conducted on Eglin AFB located in Walton County, Florida USA On the eastern portion of Eglin AF B dendritic drainage systems have cut into the remains of a Pliocene Pleistocene age delta plane, leaving erosional hills with relief of up to about 30 meters (NRCS 1989). Here the loamy sand soils allow for the formation of seepage slopes. Along hillside s, the Pinus p alustris Mill. canopy opens up into grass and sedge dominated wet prairies (Figure 5 1). Nutrient poor soils are saturated by groundwater that seeps downslope above an impermeable layer of clay, organic matter, or rock. Depending on the size of the watershed, t hese wetland s may dry out during times of drought as soil saturation is dependent on groundwater (FNAI 2010 ). The hydrologic gradient inherent to the slopes is reflected in the diverse herbaceous vegetation. Upslope tends to be drier and dominated by Aristida stricta Michx. (Wiregrass) Ctenium aromaticum (Walter) A. Wood (Toothachegrass) Muhlenbergia expansa (La m.) Trin. (Cutover M uhly ) and Rhexia spp. (M eadowbeauties) The lower slope generally has a longer hydroperiod and is dominated by Eriocaulon spp. (P ipeworts) Lycopodiella spp. (C lub M osses) Rhy n chospora spp. (B eaksedges) Xyris spp. (Y elloweyed G rasses) and a diverse suite of carnivorous plants (FNAI 2010). Sarracenia spp. (P itcherplants) Drosera spp. (S undews) Pinguicula spp. (B utterworts) and Utricularia spp. (B ladderworts) acquire nutrients by trapping
75 insects (Schnell 1976). The heliophyti c, herbaceous vegetation of this community is dependent upon frequent fire because without it, Cliftonia monophylla (Lam.) Sarg (Black T iti) and other shrubs will move in from the shrub bogs or baygalls and dominate (FNAI 2010). Data Collection How does frequency and time since last disturbance effect foliar cover, richness and species composition? In June 2010, two 1m 2 plots were randomly selected and surveyed at 20 independent seepage slope sites on the eastern portion of Eglin AFB, Walton County, Flori da, USA (Figure 5 2 ). Plot locations were chosen in undisturbed vegetation on flat, open slopes in an attempt to reduce the influence of aspect and reduce the variability of the hydrologic gradient. Experimental disturbance treatments were applied using a shovel to turn the soil in each plot to a depth of ~7 15 cm (Kotanen 1997) W ire fence exclosure s (6m in circumference and 1m in height) w ere erected to protect plots from future hog disturbance. Two disturbance treatments were applied at 10 sites while another two disturbance treatments were applied at the remaining 10 sites (e.g., F1 and I1 at 10 sites; F4 and I4 at 10 sites). The disturbance treatments applied were categorized as one of the foll owing four classes. 1. F1 O ne plot at 10 of 20 sites was disturbed frequen tly (quarterly) for one year (2010). 2. F4 O ne plot at 10 of 20 sites was disturbed frequently ( quarterly) for four years. 3. I1 O ne plot at 10 of 20 sites was disturbed infre quently (once), for one year (2010). 4. I4 O ne plot at 10 of 20 sites was di sturbed infrequently (once annually ) for four years ( June, 2010; September 2011; De cember 2012; and June 2013) Prior to application of the four disturbance treatments foliar c over was estimated in June 2010, September 2010 December 2010 and March 2011 before subsequent disturbances. A modified Daubenmire scale (0= 0%, 0.5 5 = 3%, 5 10 = 8%, 10 15 = 13%, 15 25 = 21%, 25 50 = 38%, 50 75 = 68%, 75 95 = 88%, 95 100 = 98 %) was used to estimate foliar cover of total vegetation, forbs, grasses, woody vegetation, Aristida stricta and Dichanthelium spp. (W itchgrasses) in the 1m 2 plot. A. stricta was chosen because it is considered a keystone species
76 (Noss 1989) due to its impor tance in fire spread (Clewell 1989; Haywood & Harris 1999 ). Dichanthelium spp. were chosen because they may be less sensitive to disturbance. Foliar cover for individual species was estimated for one quarter of each plot. Species nomenclature was based on Wunderlin & Hansen (2008). Richness was also estimated. Plots were also surveyed in June 2011 September 2011 March 2012, June 2012, September 2012, March 2013, June 2013 and September 2013 preceding disturbances. Undisturbed, control plots inside exclosu res were established in 2010 and surveyed in June 2010, June 2011 (total cover only), June 2012, and September 2013. Data A nalyses How does frequency and time since last disturbance effect foliar cover and species richness? This experiment followed a bala nced incomplete blocking design E ach seepage slope site was considered a block with a subset (two of four) of all disturbance treatments present in each block. Each disturbance treatment had equal replication within blocks. R epeated measure one way ANOVA (PROC GLIMMIX ; SAS 9.3; SAS Institute 2011 ) was used to evaluate s ignificance of main effects and interactions ( = 0.10). Means were arcsine square root transformed when necessary to meet assump tions of ANOVA (Zar 1998). Survey date and disturbance treatm ent (frequency and time since last disturbance) were the main effects. Site was considered a random effect Tukey Kramer tests were used for post hoc comparisons ( = 0.05) to analyze differences between means. A dditional ANOVAs were run to analyze data fro m controls because t he control plots were surveyed fewer times (once a year) than the experimental plots The data from wildlife cameras with motion sensors used in Chapter 4 was also used to analyze annual frequency of hog occurrences at seepage slopes The number of hog occurrences per site per year was calculated from the camera data and average d across cameras for each year.
77 How does frequency and time since last disturbance effect species composition? NMS ordination was used to depict temporal change s in species in terms of disturbance regime burn events, and soil type. Species that occurred in less than 5% of the plots were excluded. Andropogon spp. (Bluestems) Asteraceae spp. (Asters) Eriocaulon spp., Rhexia spp ., Rhynchospora spp., Sarracenia sp p ., Scleria spp. (Nutrushes) and Xyris spp. were grouped. In PC ORD, a Sorenson (Bray Curtis) distance measure was used in addition to shortest path for data with high beta diversity and partially disjunct outliers. Monte Carlo tests determined if the ordination results were significantly different from random data. MRPP was additionally used to test for differences within and between disturbance treatment groups ( McCune & Grace 2002). This method is effective for testing differences when sample sizes are small or distrib utions of the data are unkno wn. Results Foliar C over How does frequency and time since last disturbance effect foliar cover? Disturbance frequency and time s ince last disturbance aff ected t otal foliar cover across dates (Figure 5 3; Table 5 1; Table 5 2). Pre treatment total foliar cover among plots was indistinguishable ( p=1.0). There was a significant decline in total foliar cover after the initial disturbance treatments (p<0.0001). Frequently disturbed plots (F 1 and F4 ) continued to decline the first year of the study (p<0.0001 ). Total foliar cover was significantly less in all treatment plots (F1, I1, F4, and I4) than controls (p<0.0001) in June 2011. Although infrequently disturbed plots (I1 and I4) began to recover, total foliar cover remained significantly lower after one year than pre treatment (p<0.0001). Plots that were not disturbed after the first year of the study (F 1 and I 1 ) continued to increase in total foliar cover over the next three years. Only I1 plots recovered to pre disturbance cover (p=0.99 91 ) and w ere ind istinguishable from control s (p=0.96 3 2). After
78 three years of recovery, cover for F1 p lots remained significantly lower than cover pre disturbance (p=0.00 12 ) and the control s (p<0.0001) F1 plots were indistinguishable from I1 plots (p =0.8116 ) by September 2013 T otal cover continued to decline after subsequent disturbances when disturbed once a year (I 4 ) for the next three years. After four years of annual disturbance (I4) cover was 16.6%, significantly lower than pre disturbance (p<0.0001 ) and the contro l s (p<0.0001), but not different from F4 plots (p= 1.0 ). Plots disturbed quarterly for four years (F4 ) declined steadily to 8.60% and had t he lowest cover out of all treatments except I4, at the end of the study Frequency of d isturbance affected total cover, however these effects are confounded by time since last disturbance Cover of functional g roup s also changed by disturbance treatment over time (Table 5 1; Table 5 2). As expected, there was a decline in forb (Figure 5 4 ), grass (Figure 5 5 ), and woody species (Figure 5 6 ) cover after the initial disturbances (p<0.0001). I1 plots recovered to pre disturbance cover for forbs (p= 1.0 ), grass (p=1.0); and woody (p= 0 9994 ) by the end of the study and did not differ from control s (forbs, p=0.99 47 ; grass, p=1.0; woody, p=1.0). F1 p lots had less forb cover than pre disturbance plots (p<0.0001) but did not differ from controls (p=0.2 099 ) Grass and woody cover in F1 plots recovered to levels indistinguishable from the pre disturbance levels (p=1. 0) and controls (grass, p=0.99 01 ; woody, p=1.0). Plots disturbed over the entire four years (F 4 and I 4 ) experienced the greatest decline in cover by functional g roup Forb and grass cover was significantly less than pre disturbance cover (p<0. 0001) and controls (p<0.0001) by 2013 for F4 plots while woody vegetation cover recovered to levels comparable to pre disturbance levels (p=0. 6209 ) and controls (p=0. 2428 ) The same pattern was observed for I4 plots Forb and grass cover did not recover to pre disturbance levels (p<0.0001) or to that of controls (p<0.0001). Woody cover in I4 plots remained lower than controls (p=0.04 63 ), but did
79 not differ from pre disturbance levels (p=0.99 96 ). By 2013 cover of forbs did not differ between F4 and I4 plo ts (p=1.0), but both treatments were significantly lower than I1 (p<0.0001 ). F1 plots had forb cover that was indistinguishable from I1 plots (p=1.0), but significantly greater than I4 and F4 plots ( p=0.0766 ; p=0.00 30 respectively ). By 2013 cover of grass e s did not differ between F4 and I4 plots (p=1.0), but both treatments were significantly lower than I1 (p<0.0001 ). F1 plots had grass cover that was indistinguishable from I1 (p=1.0), but significantly greater than I4 plots and F4 plots (p=0.000 2 ; p=0<0.0 001 respectively ). By 2013, cover of woody vegetation was significantly greater in F1 and I1 plots than I4 (p=0.0917; p=0.0964 respectively). The cover of woody species did not differ between the remaining disturbance treatments by the end of study (F1&F4, p=0.4 372 ; F1&I1, p=1.0; I1&F4, p=0. 4500 ; and I4&F4, p=1.0). A ristida stricta ( Figure 5 7 ) and Dichanthelium spp. (Figure 5 8 ) varied in their responses to the experimental disturbance treatments (Table 5 1 ; Table 5 2). A ristida stricta and Dichanthelium spp. experienced a significant decline (p<0.0001 ; p=0.0 310 respectively ) in cover af ter the first disturbance event A. stricta cover continued to decline and was significantly reduced by Sept. 2013 in all disturbance treatments (F 1, p=0.00 33 ; F4, p<0.000 1; I4 p<0.0001) except I1 plots ( p=0.3 859 ) and controls (p=0.999 9 ) At the end of the study, A. stricta cover was indistinguishable between controls and I1 plots (p=0. 4033 ), but A. stricta cover was significantly reduced in F1 plots (p=0.00 44 ), I4 plots (p=0.00 11 ), and F4 plots (p=0.000 8 ) compared to control s All treatments were indistinguishable in terms of A. stricta cover by 2013 (F1& I1, p=0.9 811 ; F1&I4, p=1.0; I4&F4, p=1.0; F1&F4, p=1.0; F4&I1, p=0. 9073 ; I1&I4, p=0. 9278 ). It should be noted that whi le I1 did not significantly differ from the other treatments, A. stricta cover was double that observed in the other treatments. Also, A. stricta cover observed in
80 controls was double that of I1 plots. Dichanthelium spp. were able to recover to pre disturb ance values in all disturbance treatments (I 1 p=1.0; F 1 p=0.999 6 ; F 4 p=0. 7268 ; I 4 p=0. 9253 ) and w as n ot significantly different from controls (I1, p=0. 6118 ; F1, p=0. 9359 ; F4, p=0.7 855 ; I4, p=0.99 79 ) b y the end of the study. F4 plots had significantly l ess Dichanthelium spp. cover than I1 plots (p=0.02 6 3), but no other treatments differed by 2013 (F1& I1, p=1.0 ; F1&I4, p=0. 4896 ; I4&F4 p=1.0 ; F1&F4, p=0.1 128 ; I1&I4, p= 0. 1856 ). Cover of Sarracenia spp. declined significantly over the four year study in treatment plots ( Figure 5 9; Table 5 1; Table 5 2). The effect of disturbance regime was significant. These results indicate that soil disturbance has a negative effect on Sarracenia spp. Camera D etected H og F requency Our estimate for the number of hog occurrences at each camera site each year is conservative considering the number of months that cameras were stolen, broken, or non functional Still, on average there were more than two hog occurrences at each site every year of the study. Cameras were fi rst set up in June 2010 and averaged 3.79 hog occurrences in 2010. In 2011, the average hog occurrences at camera sites were 2.29. Hog occurrences increased to 5.75 in 2012 and 5.13 in 2013. Richness How does frequency and time since last disturbance effec t species richness? Richness was also affected by the experimental disturbance treatments (Figure 5 10 ). Overall, a seasonal pattern in richness was detected, with lower richness in March and higher richness in June. The initial disturbance resulted in a s ignificant decline (p<0.0001) in richness. Richness in I1 plots increased until September 2011 when there was no significant difference ( p=0.990 3 ) com pared to pre disturbance levels Richness in F1 plots were indistinguishable from pre disturbance values ( p=1.0) by June 2012 with values not significantly different from controls and I1
81 (p=0.9999). Until the end of the study, richness continued to increase and did not significantly differ from pre disturbance (I 1 and F 1 p=1.0 ) and controls (I1, p=0.99 1 2; F1, p=1.0) Plots disturbed for the duration of the study (F 4 and I 4 ) exhibited further decline s in richness I4 plots regained pre disturbance richness ( p=0.16 1 1) with values that continued to differ from the control s (p=0.000 5 ) by the end of the study F 4 p lots never regained pre disturbance richness (p<0.0001) and richness was significantly lower than the control (p<0.0001), I1 (p<0.0001), and F1 (p<0.0001), yet indistinguishable from I4 (p=1.0). Vegetation C omposition How does frequency and time since l ast disturbance effect species composition? NM S ordination and MRPP revealed increasing compositional change with increased disturbance frequency and decreased time since last disturbance (T= 14.1016; A=0.0875 ; p<0.0001 ; Figure 5 11 ). Vegetation compositio n in control plots and I1 plots clustered together and change d relatively little over the four years. F1 and I4 treatments changed vegetation composition, but the most extreme change in vegetation composition was observed in the F4 treatment A three dimensional ordination was optimal (stress = 18.8439 ) with axis 1 explaining 9.6%, axis 2 explaining 36.4 %, and axis 3 explaining 25.4 % of the variation. Monte Carlo test determined that the ordination differed from random data (p=0. 0164 ). When d ata w ere grouped by time since disturbance (i.e. I1 & F1 compared to I4 & F4), MRPP showed that time since disturbance is important in the clustering of groups in the NMS ordination ( T= 12 2302 ; A=0.0 3451 ; p<0.0001 ). MRPP also illustrated that soil does i nfluence community composition (T= 7.4020 ; A=0.0 385; p<0.0001 ), although disturbance regime was most reflected in the NMS ordination clusters.
82 Discussion The results from these analyses demonstrate that the effect of disturbance on seepage slope vegetation cover, richness, and composition was a function of frequency of occurrence Recovery of vegetation was, in turn, a function of time since disturbance. Frequent disturbance for four years (F4) resulted in the lowest cover and richness and most altered comp osition, while frequent disturbance for one year (F1) with three years of recovery resulted in cover, richness, and composition that was similar to controls and plots only disturbed once (I1). How does frequency and time since last disturbance effect folia r cover? D ominant flora such as A. stricta was negatively affected by soil disturbance treatments This shallow rooted (Outcalt & Lewis 1990) keystone species (Noss 1989) is important to carry fire (Clewell 1989; Haywood & Harris 1999). Hog foraging resul ts in soil disturbance that is novel in type, intensit y and frequency. A ristida stricta experienced significant declines in all treatments except plots disturbed once (I1 ) and failed to reestablish in plots that were disturbed more than once. The decline in A. stricta cover that occurred in 2013 in all treatments and controls could not be explained by removing from the analysis plots that burned in prescribed fires during the spring of 2013. It is likely that some other unquantified varia ble affected cover. Still t hese observations substantiate those in Chapter 3 where A. stricta cover was significantly lower in high intensity than control s. The loss of this keystone species is significant, especially considering that woody species were l ess negatively impacted by disturbance regime. It seems that a positive feedback is established through soil disturbance where A. stricta cover is lost and woody cover maintains, both of which reduce fire spread ( Chapter 2 ) ultimately leading to increased woody cover. Unlike A. stricta some Dichanthelium species are commonly found in early, secondary successional communities. These grasses require bare soil to recolonize and capitalize on post disturbance conditions with decreased competition and increase d colonization opportunities,
83 increasing dramatically in terms of cover (Estes 2006). In plots that were disturbed during the first year of the study (F1, I1), Dichanthelium species actually increased in cover from pre disturbance cover, although this incr ease was not significant Cover of Dichanthelium species kept grass cover high even while other grasses, such as A. stricta declined. Sarracenia species are sensitive to hog disturbance (Johnson 2001). Trampling and rooting, both the physical action and associated changes in soil moisture, negatively impact survival of juvenile Sarracenia (Hermann 1995). Research has also shown that abundance of w hite top pitcherplants and yellow trumpets are both negatively correlated with areas of hog disturbance (Engeman et al. 2007). Our results also indicate that Sarracenia species are sensitive to soil disturbance as cover declined in all treatments although cover did not decline in controls. Our experimental soil disturbance results indicate that soil disturbance more than once a year will reduce species important to the function and unique composition o f the seepage slope community. M otion sensor wildlife cameras detected a true frequency of hog foraging around two to five hog occurrences at each site every year of the study. This conservative estimate of true hog disturbance frequency in specific portions of particular slopes is greater than our experiment al frequency that resulted in degradation of the community in terms of cover of important/ keystone species. As a result, hogs appear to return to disturb seepage slopes frequently enough to negatively impact the community composition, structure, and functi on. How does frequency and time since last disturbance effect species richness? This analysis did not find richness to vary unimodally across a gradient of disturbance frequencies, but rather found that highest species richness resulted from the lowest exp erimental soil disturbance frequency and longest time since soil disturbance. This richness was equivalent to the observed pre disturbance richness levels. In the P. palustris A. stricta ecosystem, fire is a
84 large scale disturbance that reduces aboveground competition, but does not reduce richness (Kirkman et al. 2001). High diversity is achieved in P. palustris communities when frequent fire limits hardwood species dominance (and competitive exclusion) and perennial life histories allow species to persist once established (Brockway & Lewis 1997; Kirkman et al. 2001). S eepage slope vegetation is also subject to prolonged soil saturation and thus this stress (Huston 1979). This stress combined with frequent disturbance from fire limits the potential for competitive exclusion enhancing diversity (Kirkman et al. 2001). Also, oligotrophic conditions slow the rate of recovery from disturbance due to the lack of available nutrients (Walker & Peet 1983), which explains why recurrent soil disturbance at any freq uency reduces seepage slope plant diversity temporarily Species richness then has the opportunity to recover with increasing time after the disturbance. This recovery of richness should not be the only metric used to interpret recovery as differences in s pecies composition is not taken into account (McLachlan & Bazely 2001). How does frequency and time since last disturbance effect species composition? Deterministic factors of life history characteristics (Kirkman et al. 2004) and competitive relationships of resident plant species influence the vegetation dynamics and recovery trajectory after disturbance events. Many resident species in seepage slopes, like other successional communities maintained by frequent disturbance, have life history traits that al low for resprouting and vegetative expansion after fire. This response facilitates (when given three years to recover) the community returns to the pre disturbance composition (Halpern 1988). The NMS and MRPP analyses revealed that continued disturbance (F 4, I4), even at a low frequency of once a year, alters species composition. Once extirpated from a site, species with limited dispersal take longer to reestablish in disturbed sites then seeds of plants that are wind or animal
85 dispersed (Kirkman et al. 200 4). Further research is necessary to determine if seepage slopes with pro longed disturbance will recover in terms of species composition given more time considering other studies in P. palustris savannas have found that some species are particularly vulner able to disturbance and do not reestablish even after extended periods of time (i.e. 65 years) (Kirkman et al. 2004). Summary A fundamental question in ecology is whether or not communities have an inherent resilience to recover pre disturbance structur e and function, or if historical events create legacies that allow for more th an one alternative state The s uccessional trajectory and development of a community after a disturbance event is dependent upon remaining vegetation and recruitment, which in turn depends on dispersal limitations and differences in mechanisms of establishment (Young et al. 2001). This experiment demonstrate d : 1) development of a legacy effect in plots that had continued disturbance for four years ; and 2) differences among the abili ty of species to recover and recruit after varying levels of disturbance. Frequent disturbance for four years (F4) created a legacy that impaired foliar cover, richness, and altered species composition. Longer term c ontinuation of this study would be necessary to determine if a threshold has been crossed to an alternative stable state.
86 Table 5 1. Repeated measures ANOVA results for the foliar cover of total vegetation, functional groups, A. stricta Dichanthelium spp., Sarracenia spp., and specie s richness by disturbance frequency on Eglin AFB. Variable Num df df F P Total Vegetation Time 11 375.5 90.95 < 0 .0001 Treatment 3 35.99 40.22 < 0 .0001 Time x Treatment 33 375.5 17.71 < 0 .0001 Forb Time 11 374.9 27.46 < 0 .0001 Treatment 3 35.18 15.81 < 0 .0001 Time x Treatment 33 374.8 5.62 < 0 .0001 Grass Time 11 376 33.4 < 0 .0001 Treatment 3 36.09 23.46 < 0 .0001 Time x Treatment 33 375.9 9.82 < 0 .0001 Woody Time 11 375.5 14.07 < 0 .0001 Treatment 3 36.12 6.1 0.00 22 Time x Treatment 33 375.5 2.83 < 0 .0001 Aristida stricta Time 11 376 17.19 < 0 .0001 Treatment 3 36.33 11.3 < 0 .0001 Time x Treatment 33 375.9 4.01 < 0 .0001 Dichanthelium spp. Time 11 375.4 17.46 < 0 .0001 Treatment 3 35.96 7.16 0.000 8 Time x Treatment 33 375.4 3.5 < 0 .0001 Sarracenia spp. Time 11 377.9 3.17 < 0 .0001 Treatment 3 36.44 2.75 0.0 743 Time x Treatment 33 377.9 0.99 0. 5497 Richness Time 11 375.9 68.68 < 0 .0001 Treatment 3 36.34 16.49 < 0 .0001 Time x Treatment 33 375.9 6 < 0 .0001
87 Table 5 2. Repeated measures ANOVA results for the foliar cover of total vegetation, functional groups, A. stricta Dichanthelium spp., and Sarracenia spp., and species richness by disturbance frequency on Eglin AFB with controls in 2010, 2012 and 2013. Total veg etation cover was also surveyed in 2011. Variable Num df df F P Total Vegetation Time 3 161.5 325.6 <.0001 Treatment 4 55.21 142.6 <.0001 Time x Treatment 12 161.4 54.86 <.0001 Forb Time 2 105.3 37.75 <.0001 Treatment 4 52.96 16.97 <.0001 Time x Treatment 8 105.3 6.98 <.0001 Grass Time 2 106.7 25.57 <.0001 Treatment 4 54.12 11.84 <.0001 Time x Treatment 8 106.7 12.95 <.0001 Woody Time 2 106.1 7.24 0.0011 Treatment 4 54.73 5.67 0.000 6 Time x Treatment 8 106.1 1.4 0.2 200 Aristida stricta Time 2 105.8 48.68 <.0001 Treatment 4 54.63 7.44 <.0001 Time x Treatment 8 105.7 5.7 <.0001 Dichanthelium spp. Time 2 106.9 0.68 0. 6041 Treatment 4 55.52 2.54 0.0 441 Time x Treatment 8 106.9 2.9 0.0 111 Sarracenia spp. Time 2 106.3 2.78 0. 4446 Treatment 4 55.35 1.92 0. 3745 Time x Treatment 8 106.3 1.29 0. 5518 Richness Time 2 107.7 5.48 0.0 159 Treatment 4 55.87 12.05 <.0001 Time x Treatment 8 107.7 6.98 <.0001
88 Figure 5 1. Topographic relationship between seepage slope, upland pine, and baygall communities including position of zones on seepage slopes. Figure 5 2. The location of seepage slope sites on Eglin AFB.
89 Figure 5 3 Total foliar cover by disturbance treatment over time on Eglin AFB. D ifferent letters abov e points indicate significant difference ( = 0.05) within each survey date following Tukey Kramer post hoc tests. Dotted lines indicate annual disturbance treatment.
90 Figure 5 4 F oliar cover of forbs by disturbance treatment over time on Eglin AFB. Different letters above points indicate significant difference ( = 0.05) within each survey date following Tukey Kramer post hoc tests. Dotted lines indicate annual disturbance treatment.
91 Figure 5 5 Foliar cover of grasses by disturbance treatment over time on Eglin AFB. Different letters above points indicate significant difference ( = 0.05) within each survey date following Tukey Kramer post hoc tests. Dotted lines indicate annual disturbance tre atment.
92 Figure 5 6. Foliar cover of woody species by disturbance treatment over time on Eglin AFB. Different letters above points indicate significant difference ( = 0.05) within each survey date following Tukey Kramer post hoc tests. Dotted lines indicate annual disturbance treatment.
93 Figure 5 7. Foliar cover of A. stricta by disturbance treatment over time on Eglin AFB. Different letters above points indicate significant difference ( = 0.05) within each survey date following Tukey Kramer p ost hoc tests. Dotted lines indicate annual disturbance treatment.
94 Figure 5 8. Foliar cover of Dichanthelium spp. by disturbance treatment over time on Eglin AFB. Different letters above points indicate significant difference ( = 0.05) within each survey date following Tukey Kramer post hoc tests. Dotted lines indicate annual disturbance treatment.
95 Figure 5 9. Foliar cover of Sarracenia spp. by disturbance treatment over time on Eglin AFB. Dotted lines indicate annual dis turbance treatment.
96 Figure 5 10. Species richness by disturbance treatment over time on Eglin AFB. Different letters above points indicate significant difference ( = 0.05) within each survey date following Tukey Kramer post hoc tests. Dotted lines indicate annual disturbance treatment.
97 Figure 5 11. Three dimensional nonmetric multidimensional ordination showing cover by species by disturbance treatme nt on Eglin AFB in June 2010 and Sept 2013 Each point represents a plot and the distance between points is proportional to the dif ference in species composition
98 C HAPTER 6 CONCLUSIONS The seepage slope community is for the most part, poorly understood despite its intrins ic beauty and scientific value (Hermann 1995) This research addresses a data gap by quantifying hog disturbance and its impact on this unique wetland habitat. Our results strongly suggest that hog management should be viewed as a long term management strategy that may need to be adjusted on an annual basis dependi ng on environmental factors (Bieber & Ruf 2005). Managers should protect the lower portions of seepage slopes (e.g., using fence exclosures) where feral hogs prefer to forage and where rare, endemic flora and fauna are found. Additionally, it is evident that this novel soil disturbance reduces cover of A. stricta and alters vegetation composition, likely indicating that ruderal species recolonize hog created gaps where competit ion is decreased (Cushman et al. 2004). This research indicates that the exclusion of feral hogs from seepage slopes is the first step towards the conservation and/or restoration of this endangered plant community In some cases and where practical, it may be necessary to actively restore portions of a slope with appropriate combinations of soil stabilization, direct seeding, and/or outplanting of seedlings of native species. Considering that so little of the original extent of this community is still intac t (FNAI 1990), encroachment of woody species into remaining seepage slopes is a serious threat. It may be necessary for prescribed fire crews to push fire downslope into areas where fire may not spread unassisted or cut and clear these areas by hand. Conti nued active management, including hog control and monitoring, is essential to mitigating hog damage and preserving the ecological i ntegrity of this rare resource.
99 LIST OF REFERENCES Aplet, G.H., Anderson, S.J., & Stone, C.P. 1 991. Association between feral pig disturbance and the composition of some alien plant assemblages in Hawaii Volcanoes National Park. Vegetatio 95(1): 55 62. Archer, S. 1989. Have southern Texas savannas been converted to woodlands in recent history? Ameri can Naturalist 134: 545 561. Arrington, D.A., Toth, L.A., & Koebel, J.W. 1999. Effects of rooting by feral hogs Sus scrofa L. on the structure of a floodplain vegetation assemblage. Wetlands 19 (3): 535 544. Barrios Garcia, M.N., & Ballari, S.A. 2012. Impa ct of wild boar ( Sus scrofa ) in its introduced and native range: a review. Biological Invasions 14 (11 ): 2283 2300. Bieber, C. & Ruf, T. 2005. Population dynamics in wild boar Sus scrofa : ecology, elasticity of growth rate and implications for the management of pulsed resource consumers. Journal of Applied Ecology 42: 1203 1213. Bond, J.B. & Keeley J.E 2005. Fire as a global 'herbivore': the ecology and evolution of flammable ecosystems Trends in Ecology and Evolution 20(7): 387 394. Bratton, S.P. 1974. Effect of European wild boar ( Sus scrofa ) on high elevation vernal flora in Great Smoky Mountains National Park. Bulletin of the Torrey Botanical Club 101: 198 206. Bratton, S.P. 1975. Effect of European Wild Boar, Sus Scrofa on Gray Beech Forest in Great Smoky Mountains. Ecology 56 (6): 1356 1366. Bratton, S.P., Harmon, M.E., & White, P.S. 1982. Patterns of European wild boar rooting in the western Great Smoky Mounta ins. Castanea 47: 230 242. Brennan, E.K., Smith, L.M., Haukos, D.A., & LaGrange, T.G. 2005 Short term response of wetland birds to prescribed burning in Rainwater Basin wetlands Wetlands 25 (3): 667 674. Briggs, J.M., Knap, A.K., Blair, J.M., Heisler, J.L., Hoch, G.A., Lett, M.S., & McCarron, J.K. 2005. An Ecosystem in Transition: Causes and Consequences of the Conversion of Mesic Grassland to Shrubland. BioScience 55(3): 243 254. Brisbin, Jr., I.L. & Sturek, M. S. 2009. The Pigs of Ossabaw Island: A Ca se Study of the Application of Long term Data in Management Plan Development. In: Ma yer, J.J. & Brisbin Jr., I.L. (eds.) Wild Pigs: Biology, Damage, Control Techniques and Management SRNL RP 2009 00869, pp. 365 378. Savannah River National Laboratory, Ai ken, SC, US.
100 Brockway, D.G. & Lewis, C.E. 1997. Long term effects of dormant season prescribed fire on plant community diversity, structure and productivity in a longleaf pine wiregrass ecosystem. Forest Ecology and Management 96: 167 183. Brooks, M.L., D' Antonio, C.M., Richardson, D.M., Grace, J.B., Keeley, J.E., DiTomaso, J.M., Hobbs, R.J., Pellant, M., & Pyke, D. 2004. Effects of invasive alien plants on fire regimes. BioScience 54 : 677 688. Callaway, R.M. & Davis, F.W. 1998. Recruitment of Quercus agrif olia in central California: the importance of shrub dominated patches. Journal of Vegetation Science 9 (5): 647 656. Challies, C.N. 1975. Feral pigs ( Sus scrofa ) on Auckland Island: status, and effects on vegetation and nesting sea birds. New Zealand Journal of Zoology 2: 479 490. Chavarria, P.M., Lopez, R.R., Bowser, G., & Silvy, N.J. 2007. A landscape level survey of feral hog impacts to natural resources of the Big Thicket National Preserve. Human Wildlife Conflicts 1: 199 204. Cheney, N.P., Gould, J.S., & Catchpole, W.R. 1993. The Influence of Fuel, Weather and Fire Shape Variables on Fire Spread in Grasslands. International Journal of Wildland Fire 3(1): 31 44. Choquenot, D. 1998. Testing the relative influence of intrinsic and extrinsic variation in food Journal of Animal Ecology 67: 887 907. Choquenot, D., McIlroy, J., & Korn, T. 1996. Managing Vertebrate Pests: Feral Pigs Australian Government Publishing Service, Canberra, AU. Clew ell, A.F. 1986. Natural setting and vegetation of the Florida Panhandle An account of the environments and plant communities of northern Florida west of the Suwannee River. Report No. COESAM/PDEI 86/001. United States Army Corps of Engineers, Mobile Dist rict, AL, US. Clewell, A.F. 1989. Natural history of wiregrass ( Aristida stricta Michx., Gramineae). Natural Areas Journal 9: 223 233. Clewell, A.F., Raymond, C., Coultas, C.L., Dennis, W.M., & Kelly, J.P. 2009. Spatially narrow wet prairies. Castanea 74( 2): 146 159. Coblentz, B.E. & Barber, D.W. 1987. Biology and Control of Feral Pigs on Isla Santiago, Galapagos, Ecuador. Journal of Applied Ecology 24(2): 403 418. Cole, R.J., Litton, C.M., Koontz, M.J., & Loh, R.K. 2012. Vegetation recovery 16 years after feral pig removal from a wet Hawaiian forest. Biotropica 44 (4): 463 471. Collins, S.L. 1987. Interaction of disturbances in tallgrass prairie: a field experiment. Ecology 68 (5): 1243 1250.
101 Collins, S.L., & Barber, S.C. 1986. Effects of disturbance on div ersity in mixed grass prairie. Vegetatio 64 (2 3): 87 94. Compton, V., Bachant Brown, J., Hicks, M., & Penniman, P. 2006. Role of Public Private Partnerships in Restoration. In: Jose, S., Jokela, E.J., & Miller, D.L. (eds.) The longleaf pine ecosystem: eco logy, silviculture, and restoration pp. 413 429. Springer Science+Business Media, LLC, New York, NY, US. Connell, J.H. 1978. Diversity in tropical rain forests and coral reefs. Science 199 (4335): 1302 1310. Copeland, T.E., Sluis, W., & Howe, H.F. 2002. F ire season and dominance in an Illinois tallgrass prairie restoration. Restoration Ecology 10 (2): 315 323. Coppedge, B.R., & Shaw, J.H. 1998. Bison grazing patterns on seasonally burned tallgrass prairie. Journal of Range Management 51(3): 258 264. Cuevas, M.F., Novillo, A., Campos, C., Dacar, M.A., & Ojeda, R.A. 2010. Food habits and impact of rooting behaviour of the invasive wild boar, Sus scrofa in a protected area of the Monte Desert, Argentina. Journal of Arid Environ ments 74(11): 1582 1585. Cushman, J.H., Tierney, T.A., & Hinds, J.M. 2004. Variable effects of feral pig disturbances on native and exotic plants in a California grassland. Ecological Applications 14 (6): 1746 1756. Daubenmire, R. 1959. A Canopy Coverage Method of Vegetational Anal ysis. Northwest Science 33(1): 43 64. Denslow, J.S. 1980. Patterns of plant species diversity during succession under different disturbance regimes. Oecologia 46 (1): 18 21. Drewa, P.B., Platt, W.J., & Moser, E.B. 2002. Fire effects on resprouting of shrubs in headwaters of southeastern longleaf pine savannas. Ecology 83 (3): 755 767. DoD Air Force. 1995. Eglin Air Force Base Natural Resource Climatological Report. Natural Resource Management Branch, Ni ceville, FL, US. Doup, R.G., Mitchell, J., Knott, M.J., Davis, A.M., & Lymbery, A.J. 2010. Efficacy of exclusion fencing to protect ephemeral floodplain lagoon habitats from feral pigs ( Sus scrofa ). Wetlands Ecology and Management 18 (1): 69 78. Dzieciolow ski, R.M., Clarke, C.M.H., & Frampton, C.M. 1992. Reproductive characteristics of feral pigs in New Zealand. Acta Theriologica 37: 259 270. Eleuterius, L.N. & Jones, S.B. 1969. A floristic and ecological study of pitcher plant bogs in south Mississippi. Rh odora 71(785): 29 34.
102 Enge, K.M. 1997. Habitat Occurrence of Florida's Native Amphibians and Reptiles Technical Report No. 16. Florida Game and Fresh Water Fish Commission, Quincy, Florida, US. Engeman, R.M., Stevens, A., Allen, J., Dunlap, J., Daniel, M., Teague, D., & Constantin, B. 2007. Feral swine management for conservation of an imperiled wetland habitat: Florida's vanishing seepage slopes. Biological Conservation 134 (3): 440 446. Estes, B.L. 2006. Impact of interacting disturbances on longleaf pine communities. Ph.D. dissertation, Auburn University, Auburn, AL, US. ESRI (Environmental Systems Resource Institute). 2009. ArcMap 9.2. ESRI, Redlands, CA, US. Everitt, J.H., & Alaniz, M.A. 1980. Fall and winter diets of feral pigs in south Texas. Journal o f Range Management 33(2): 126 129. Finney, M.A., Cohen, J.D., Grenfell, I.C., & Yedinak, K.M. 2010. An examination of fire spread thresholds in discontinuous fuel beds. International Journal of Wildland Fire 19: 163 170. FNAI ( Florida Natural Areas Invento ry ) 1990. Guide to the natural communities of Florida Florida Natural Areas Inventory and Florida Department of Natural Resources, Tallahassee, FL, US. FNAI (Florida Natural Areas Inventory). 2010. Guide to the natural communities of Florida: 2010 editi on Florida Natural Areas Inventory, Tallahassee, FL, US. Folkerts, G. W. 1982. The Gulf Coast Pitcher Plant Bogs. American Scientist 70(3): 260 267. Frost, C.C. 1995. Presettlement fire regimes in southeastern marshes, peatlands and swamps. In: Cerulean, S.I. & Engstrom, R.T. (eds.) Fire in Wetlands: A Management Perspective pp. 39 60. Proceedings Tall Timbers Fire Ecology Conference No. 19, Tallahassee, FL, US. Frost, C.C. 2000. Studies in landscape fire ecology and presettlement vegetation of the southe astern United States. Ph.D. dissertation, University of North Carolina, Chapel Hill, NC, US. Frost, C 2006. History and Future of the Longleaf Pine Ecosystem. In : Shibu, J., Jokela, E.J., & Miller, D.L. (eds.). The Longleaf Pine Ecosystem: Ecology, Silviculture, and Restoration pp. 9 42. Springer, New York, NY, US. Frost, C.C., Walker, J., & Peet, R.K. 1986. Fire dependent savannas and prairies of the Southeast: original extent, preservation status and management problems. In: Kulhavy, D.L. & Conne r, R.N. (eds.) Wilderness and Natural Areas in the Eastern United States: A Management Challenge, pp. 348 357. Center for Applied Studies, School of Forestry, Stephen F. Austin State University, Nacogdoches, TX, US. Fujinuma, J. & Harrison, R.D. 2012. Wild p igs ( Sus scrofa ) mediate large scale edge effects in a lowland tropical rainforest in p eninsular Malaysia. PLoS One 7 (5): e37321.
103 Giles, J.R. 1980. The ecology of the feral pig in western New South Wales. Ph.D. thesis, University of Sydney, Sydney, AU. Giuliano, W.M. 2009. Wild hogs in Florida: Ecology and management (WEC 227). Gainesville: University of Florida Institute of Food and Agricultural Sciences. URL: http://edisifasufledu/DLN. Glitzenstein, J.S., Streng, D.R., Masters, R.E., Robertson, K.M., & Hermann, S.M. 2012. Fire frequency effects on vegetation in north Florida pinelands: Another look at the long term Stoddard Fire Research Plots at Tall Timbers Research Station. Forest Ecology and Management 264: 197 209. Glitzenstein, J.S., Streng, D.R., & Wade, D.D. 2003. Fire frequency effects on longleaf pine ( Pinus palustris P. Mill.) vegetation in South Carolina and northeast Florida, USA. Natural Areas Journal 23: 22 37. Grayson, D.K. & Meltzer, D.J. 2002. Clovis hunting and large mammal extinction: a critical review of the evidence. Journal of World Prehistory 16 (4): 313 359. Gunderson, L.H. 2000. Ecological resilience -in theory and application. Annual Review of Ecology and Systematics 31: 425 439. Halpern, C.B. 1988. Early successional pathways and the resistance and resilience of forest communities. Ecology 69 (6): 1703 1715. Hanson, L.B., Mitchell, M.S., Grand, J.B., Jolley, D.B., Sparklin, B.D., & Ditchkoff, S.S. 2009. Effect of experimental manip ulation on survival and recruitment of feral pigs. Wildlife Research 36: 185 191. Harrington, G.N. & Hodgkinson, K.C. 1986. Shrub grass dynamics in mulga c ommunities of Eastern Australia. In: Joss, E J., Lynch, E.W., & Williams, O.B. (eds.) Rangelands: A R esource under Siege, pp. 26 28. Australian Academy of Science, Canberra, AU. Haywood, J.D. & Grelen, H.E. 2000. Twenty years of prescribed burning influence the development of direct seeded longleaf pine on a wet pine site in Louisiana. Southern Journal of Applied Forestry 24: 86 92. Haywood, J.D. & Harris F.L. 1999. Description of vegetation in several periodically burned longleaf pine forests on the Kisatchie National Forest. In: Haywood, J.D. (ed.), Tenth Biennial Southern Silvicultural Research Confere nce pp. 217 222. Shreveport, LA, US. Henry, V.G. & Conley, R.H. 1972. Fall foods of European wild hogs in the Southern Appalachians. Journal of Wildlife Management 36: 854 860. Hermann, S.M. 1995. Status and management of Florida's carnivorous plant com munities Nongame Wildlife Program project no. GFC 84 033. Florida Fish and Wildlife Conservation Commission, Tallahassee, FL, US.
104 Hierro, J.L., Clark, K.L., Branch, L.C., & Villarreal, D. 2011. Native herbivore exerts contrasting effects on fire regime a nd vegetation structure. Oecologia 166(4): 1121 1129. Hobbs, R.J. & Huenneke, L.F. 1992. Disturbance, Diversity, and Invasion Implications for Conservations. Conservation Biology 6 (3): 324 337. Hobbs N.T., Schimel, D.S., Owensby, C.E., & Ojima, D.J. 1991. Fire and grazing in the tallgrass prairie: contingent effects on nitrogen budgets. Ecology 72: 1374 1382. Holdo, R.M., Holt, R.D., & Fryxell, J.M. 2009. Grazers, browsers, and fire influence the extent and spatial pattern of tree cover i n the Serengeti. Ecological Applications 19: 95 109. Holling C.S. 1973. Resilience and stability of ecological systems. Annual Review of Ecology and System atics 4:1 23 Hone, J. 2002. Feral pigs in Namadgi National Park, Australia: dynamics, impacts and m anagement. Biological Conservation 105: 231 242. Hone, J. 1995. Spatial and temporal aspects of vertebrate pest damage with emphasis on feral pigs. Journal of Applied Ecology 32: 311 319. Howe, T.D. & Bratton, S.P. 1976. Winter rooting activity of the Euro pean wild boar in the Great Smoky Mountains National Park. Castanea 41(3): 256 264. Huffman, J.M. & Judd, W.S. 1998. Vascular flora of Myakka River State Park, Sarasota and Manatee Counties, Florida. Castanea 63(1): 25 50. Huston, M. 1979. A general hypot hesis of species diversity. American Naturalist 113(1): 81 101. Janzen, D.H. 1984. Dispersal of small seeds by big herbivores: foliage is the fruit. American Naturalist 123(3): 338 353. Jensen, A. 1985. The effect of cattle and sheep grazing on salt marsh vegetation at Skallingen, Denmark. Vegetatio 60(3): 7 48. Johnson, E.D. 2001. Pitcherplants and their habitats in the Florida State Park system: resource conditions, trends, and manageme nt needs. Resource Management Evaluation. Division of Recreation and Parks, Florida Department of Environmental Protection, Tallahassee, FL, US. Kay, C. E. 2007. Are lightning fires unnatural? A comparison of aboriginal and lightning ignition rates in the United States. In: Masters, R.E. & Galley, K.E.M. (eds.) Fire in grassland and shrubland ecosystems Proceedings Tall Timbers Fire Ecology Conference No. 23, Tallahassee, FL, US. Kirkman, L.K., Coffey, K.L., Mitchell, R.J., & Moser, E.B. 2004. Ground cover recovery patterns and life history traits: implications for restoration obstacles and opportunities in a species rich savanna. Journal of Ecology 92 (3): 409 421.
105 Kirkman, L.K., Mitchell, R.J., Helton, R.C., & Drew, M.B. 2001. Productivity and species richness across an environmental gradient in a fire dependent ecosystem. American Journal of Botany 88 (11): 2119 2128. Keeley, J.E. & Zedler, P.H. 1998. Evolution of life history in Pinus. In: Richardson, D.M. (ed.) Ecology and Biogeography of Pinus pp. 2 19 249. Cambridge University Press, Cambridge, UK. Kerby, J.D., Fuhlendorf, S.D., & Engle, D.M. 2007. Landscape heterogeneity and fire behavior: scale dependent feedback between fire and grazing processes. Landscape Ecology 22: 507 516. Knapp, A.K., Blair J.M., Briggs, J.M., Collins, S.L., Hartnett, D.C., Johnson, L.C., & Towne, E.G. 1999. The keystone role of bison in North American tallgrass prairie. BioScience 49: 39 50. Knapp, E.E., Estes, B.L., & Skinner, C.N. 2009. Ecological Effects of Prescribed F ire Season: A Literature Review and Synthesis for Managers. USDA Forest Service, Gen. Tech. Rep. PSW GTR 224, Albany, CA, US. Komarek, E.V. 1965. Fire ecology Grasslands and man. Proceedings of Tall Timbers Fire Ecology Conference 4: 169 220. Komarek, E.V 1968. Lightning and lightning fires as ecological forces. Proceedings of Tall Timbers Fire Ecology Conference 8: 169 197. Komarek, E.V. 1974. Effects of fire on temperate forests and related ecosystems: southeastern United States. In: Kozlowski, T.T. & A hlgren, C. (eds.) Fire and Ecosystems, pp. 251 277. Academic Press, New York NY, US Kotanen, P.M. 1995. Responses of vegetation to a changing regime of disturbance Effects of feral pigs in a Californian coastal prairie. Ecography 18 (2): 190 199. Kotanen P.M. 1997. Effects of experimental soil disturbance on revegetation by natives and exotics in coastal California meadows. Journal of Applied Ecology 34(3): 631 644. Kraaij, T. & Milton, S.J. 2006. Vegetation changes (1995 2004) in semi arid Karoo shrubland, South Africa: effects of rainfall, wild herbivores and change in land use. Journal of Arid Environments 64 (1): 174 192. Krebs, C.J. 1991. The experimental paradigm and long te rm population studies. Ibis 133: 3 8. Krull, C.R., Choquenot, D., Burns, B.R., & Stanley, M.C. 2013. Feral pigs in a temperate rainforest ecosystem: disturbance and ecological impacts. Biological Invasions 15: 2193 2204. Lacki, M.J. & Lancia, R.A. 1986. E ffects of Wild Pigs on Beech Growth in Great Smoky Mountains National Park. The Journal of Wildlife Management 50(4): 655 659.
106 Legg, C.J. & Nagy, L. 2006. Why most conservation monitoring is, but need not be, a waste of time. Journal of Environmental Mana gement 78 (2): 194 199. Lemon, P.C. 1946. Prescribed burning in relation to grazing in the longleaf slash pine type. Journal of Forestry 44 (2): 115 117. Lett, M.S. & Knapp, A.K. 2003. Consequences of shrub expansion in mesic grassland: Resource alterations and graminoid responses. Journal of Vegetation Science 14: 487 496. Lipscomb, D.J. 1989. Impacts of feral hogs on longleaf pine regeneration. Southern Journal of Applied Forestry 13: 177 181. Lubchenco, J. 1978. Plant species diversity in a marine intertidal community: importance of herbivore food preference and algal competitive abilities. American Naturalist 112(983): 23 39. MacDonald, D.W. 1980 The Red Fox, Vulpes vulpes as a Predator upon Earthworms, Lumbricus terrestris Zeitschrift fr Tierpsychologie 52(2): 171 200 Mack, M.C., & D'Antonio, C.M. 1998. Impacts of biological invasions on disturbance regimes. Trends in Ecology & Evolution 13 (5): 195 198. Macke y, R. L. & Currie, D.J. 2001. T he diversity disturbance relationship: is it generally strong and peaked? Ecology 82 (12): 3479 3492. Maschinski, J. & Whitham, T.G. 1989. The continuum of plant responses to herbivory: the influence of plant association, nutrient availability, and timing. American Naturalist 134: 1 19. Massei, G. Genov, P.V., Staines, B.W., & Gorman, M.L. 1997. Mortality of wild boar, Sus scrofa in a Mediterranean area in relation to sex and age. Journal of Zoology (London) 242: 394 400. Mayer, J.J. & Brisbin, I.L. 1991. W ild pigs in the United States: Their life history, morphology and current status University of Georgia Press, Athens, GA, US. McCune, B. & Grace J.B. 2002. Analysis of ecological communities. MjM Software Design, Gleneden Beach, OR, US. McLachlan, S.M. & Bazely, D.R. 2001. Recovery patterns of understory herbs and their use as indicators of deciduous forest regeneration. Conservation Biology 15 (1): 98 110. McNaughton, S.J. 1992. The propagation of disturbance in savannas through food webs. Journal of Veget ation Science 3: 301 314. Means, B.D. 2006. Vertebrate Faunal Diversity of Longleaf Pine Ecosystems. In: Jose, S., Jokela, E.J., & Miller, D.L. (eds.) The longleaf pine ecosystem: ecology, silviculture, and restoration pp. 413 429. Springer Science+Busi ness Media, LLC, New York, NY, US.
107 Means, D.B. & Moler, P.E. 1979. The pine barrens tree frog: fire, seepage bogs, and management implications. In: Odum, R.R. & Landers, L. (eds.) Proceedings of Rare and Endangered Wildlife Symposium. Technical Bulletin WL4 pp. 77 83. Georgia Department of Natural Resources, Game and Fish Division, Athens, Georgia, US. Mitchell, J., Dorney, W., Mayer, R., & McIlroy, J. 2007a. Spatial and temporal patterns of feral pig diggings in rainforests of north Queensland. Wildlife Research 34: 597 602. Mitchell, J., Dorney, W., Mayer, R., & McIlroy, J. 2007b. Ecological impacts of feral pig diggings in north Queensland rainforests. Wildlife Research 34 (8): 603 608. Mitchell, R., Engstrom, T., Sharitz, R. R., De Steven, D., Hiers, K ., Cooper, R., & Kirkman, L. 2009. Old forests and endangered woodpeckers: old growth in the southern Coastal Plain. Natural Areas Journal 29 (3): 301 310. Mitchell, J.L. & Mayer, R. 1997. Digging by feral pigs within the wet tropics World Heritage Area of north Queensland. Wildlife Research 24: 591 603. Moloney, K.A. & S.A. Levin 1996. The effects of disturbance architecture on landscape level population dynamics. Ecology 77 (2): 375 394. Myers, R.L. 1990. Scrub and high pine. In: Myers, R.L. & Ewel, J.J. (e ds.) Ecosystems of Florida University of Central Florida Press, Orlando, FL, US. Nader, G., Henkin, Z., Smith, E., Ingram, R., & Narvaez, N. 2007. Planned herbivory in the management of wildfire fuels. Rangelands 29(5): 18 24. National Drought Mitigation Center. 2013. Drought monitor archives. Lincoln, Nebraska, US. URL: http://droughtmonitor.unl.edu/archive.html. Natural Resource Conservation Service (NRCS). 1989. Soil Survey of Walton County, Florida. National Cooperative Soil Survey. United States Depar tment of Agriculture, Washington, DC, US. NOAA (National Oceanic and Atmospheric Administration). 2002. Monthly station normals of temperature, precipitation, and heating and cooling degree days 1971 2000. Climatography of the United States No. 81, U.S. Department of Commerce, National Oceanic and Atmospheric Administration, Washington, D.C., US. Noss, R.F. 1989. Longleaf pine and wiregrass: keystone components of an endangered ecosystem. Natural Areas Journal 9: 211 213. O'connor, T. G. & Roux, P.W. 1995. Vegetation changes (1949 71) in a semi arid, grassy dwarf shrubland in the Karoo, South Africa: influence of rainfall variability and grazing by sheep. Journal of Applied Ecology 32(3): 612 626.
108 Olson, M.S. 1992. Effects of early and late growing season fires on resprouting of shrubs in upland longleaf pine savannas and embedded seepage savannas M.S. Thesis, Louisiana State University, Baton Rouge, LA, US. Olson, M.S. & Platt, W.J. 1995. Effects of habitat and growing season fires on resp routing of shrubs in longleaf pine savannas. Vegetatio 119: 101 118. Outcalt, K.W. & Lewis, C.E. 1990. Response of wiregrass ( Aristida stricta ) to mechanical site preparation. In: Duever, L.C. & Noss, R.F. (eds.) Wiregrass biology and management: maintaini ng groundcover integrity in longleaf pine ecosystems Symposium Proceedings pp. 1 12. KBN Engineering and Applied Sciences, Gainesville, FL, US Outcalt, K.W., Williams, M.E., & Onokpise, O. 1999. Restoring Aristida stricta to Pinus palustris ecosystems o n the Atlantic Coastal Plain, U.S.A. Restoration Ecology 7 : 262 270. Paine, R.T., Tegner, M.J., & Johnson, E.A. 1998. Compounded perturbations yield ecological surprises. Ecosystems 1(6): 535 545. Pearson, H.A., Davis, J.R., & Schubert, G.H. 1972. Effects of wildfire on timber and forage production in Arizona. Journal of Range Management 25(4): 250 253. Peart, D., Patten, D.T., & Lohr, S.L. 1994. Feral pig disturbance and woody species seedling regeneration and abundance beneath coast live oaks ( Quercus agrifolia ) on Santa Cruz Island, California. In: Halverson, W.L. & Maender, G.J. (eds.) The fourth California Islands symposium: update on the status of resources pp. 313 322. Santa Barbara Museum of Natural History, Santa Barbara, CA, US. Peine, J.D. & Farmer, J.A. 1990. Wild hog management programme at Great Smoky Mountains National Park. Proceedings of the 14th Vertebrate Pest Conference 14: 221 227. Petraitis, P.S., Latham, R.E., & Niesenbaum, R.A. 1989. The maintenance of species diversity by distur bance. Quarterly Review of Biology 64(4): 393 418. Pickett, S.T.A., Kolasa, J., Armesto, J.J., & Collins, S.L. 1989. The ecological concept of disturbance and its expression at various hierarchical levels. Oikos 54(2): 129 136. Pickett, S.T. & White, P.S. 1985. The Ecology of natural disturbance and patch dynamics Orlando, Florida, Academic Press. Platt, W.J., Beckage, B., Doren, R.F., & Slater, H.H. 2002. Interactions of large scale disturbances: prior fire regimes and hurricane mortality of savanna pin es. Ecology 83: 1566 1572. Recher, H.F. & Clark, S.S. 1974. A biological survey of Lord Howe Island with Biological Conservation 6: 263 273.
109 Reynolds, C.S. 1993. Scales of disturbance and their role in plankton ecology. Hydrobiologia 249: 157 171. Richardson, C.D., Gipson, P.S., Jones, D.P., & Luchsinger, J.C. 1997. Extirpation of a recently established feral pig population in Kansas. Proceedings of the Eastern Wildlife Damage Management Confere nce 7: 100 103. Rodgers, H.L. & Provencher, L. 1999. Analysis of longleaf pine sandhill vegetation in northwest Florida. Castanea 64(2): 138 162. Rostlund, E. 1960. The geographic range of the historic bison in the southeast. Annals of the Association of A merican Geographers 50 (4): 395 407. Ruess, R.W. & McNaughton, S.J. 1987. Grazing and the dynamics of nutrient and energy regulated microbial processes in the Serengeti grasslands. Oikos 49: 101 110. SAS Institute, 2011. Base SAS 9.3 Procedures Guide. SAS Institute Inc. Cary, NC, US. Saunders, G.R. & Kay, B. 1991. Movements of feral pigs ( Sus scrofa ) at Sunny Corner, New South Wales. Wildlife Research 18: 49 61. Saunders, G., Kay, B., & Nicol, H. 1993. Factors affecting bait uptake and trapping success for feral pigs ( Sus scrofa ) in Kosciusko National Park. Wildlife Research 20 (5): 653 665. Schindler, J.R., Fulbright, T.E., & Forbes, T.D.A. 2004. Shrub regrowth, antiherbivore defenses, and nutritional value following fire. Rangeland Ecology & Management 57 (2): 178 186. Savage, M. & Swetnam, T.W. 1990. Early 19th century fire decline following sheep pasturing in a Navajo ponderosa pine forest. Ecology 71: 2374 2378. Schlesinger, W.H., Reynolds, J.F., Cunningham, G.L., Huenneke, L.F., Jarrell, W.M., Virginia, R.A., & Whitford, W.G. 1990. Biological feedbacks in global desertification. Science 247: 1043 1048. Schnell, D.E. 1976. Carnivorous plants of the United States and Canada. Timber Press, Winston Salem, NC, US. Schroeter, S.C., Dixon, J.D., Kastendiek, J., & Smith, R.O. 1993. Detecting the ecological effects of environmental impacts a case study of kelp forest invertebrates. Ecological Applications 3: 331 350. S hea, K., Roxburgh, S.H., & Rauschert, E.S.J. 2004. Moving from pattern to process: coexistence mechanisms under intermediate disturbance regimes. Ecology Letters 7 (6): 491 508. Shepherd, B.J., Miller, D.L., & Thetford, M. 2012. Fire season effects on flowering characteristics and g ermination of longleaf pine ( Pinus palustris ) savanna grasses. Restoration Ecology 20 (2): 268 276.
110 Siemann, E., Carrillo, J.A., Gabler, C.A., Zipp, R., & Rogers, W.E. 2009. Experimental test of the impacts of feral hogs on forest dynamics and processes in the southeastern US. Forest Ecology and Management 258: 546 553. Silvertown, J. & Bullock, J.M. 2003. Do seedlings in gaps interact? A field test of assumptions in ESS seed size models. Oikos 101(3): 499 504. Singer, F.J., Swank, W.T., & Clebsch, E.E.C. 1984. Effects of wild pig rooting in a deciduous forest. The Journal of Wildlife Management 48(2): 464 473. Slapcinsky, J.L., Gordon, D.R., & Menges, E.S. 2010. Responses of Rare Plant Species to Fire in Florida's Pyrogenic Communities. Na tural Areas Journal 30(1): 4 19. Slocum, M.G., Platt, W.J., & Cooley, H.C. 2003. Effects of differences in prescribed fire regimes on patchiness and intensity of fires in subtropical savannas of Everglades National Park, Florida. Restoration Ecology 11(1): 91 102. Sousa, W.P. 1984. The role of disturbance in natural communities. Annual Review of Ecology and Systematics 15(35): 3 391. Sparklin, B.D., Mitchell, M.S., Hanson, L.B., Jolley, D.B., & Ditchkoff, S.S. 2009. Territoriality of feral pigs in a highly persecuted population on Fort Benning, Georgia. Journal of Wildlife Management 73(4): 497 502. Spatz, G. & Mueller Dombois, D. 1975. Successional patterns after pig digging in grassland communities on Mauna Loa, Hawaii. Phytocoenologia 3: 346 373. Stout, I.J. & Marion, W.R. 1993. Pine flatwoods and xeric pine forests of the southern (lower) coastal plain. In: Martin, W.H., B oyce, S.G., & Echternacht, A.C. (eds ). Biodiversity of the southeastern United States: lowland terrestrial communities pp. 373 446. John Wiley & Sons, New York, US. Stritar, M.L., Schweitzer, J.A., Hart, S.C., & Bailey, J.K. 2010. Introduced ungulate herbivore alters soil processes after fire. Biological Invasions 12 (2): 313 324. Sweitzer, R.A. 1998. Conservation implications of feral pigs in island and mainland ecosystems and a case study of feral pig expansion in California. Proceedings 18 th Vertebrate Pest Conference 18: 26 34. Sweitzer, R.A. & Van Vuren, D.H. 2002. Rooting and foraging effects of wild pigs on tree regeneration and Proceedings of General Technical Report PSW GTR 184, pp. 219 231. Sweitzer, R.A., Van Vuren, D., Gardner, I.A., Boyce, W.M., & Waithman, J.D. 2000. Estimating sizes of wild pig populations in the north and central coast regions of California. Journal of Wildlife Management 64(2): 531 543.
111 Tardiff, S.E. & Stanford J.A. 19 98. Grizzly bear digging: effects on subalpine meadow plants in relation to mineral nitrogen availability. Ecology 7: 2219 2228. Taylor, R.B., Hellgren, E.C., Gabor, T.M., & Ilse, L.M. 1998. Reproduction of feral pigs in southern Texas. Journal of Mammalo gy 79: 1325 1331. Taylor, D.L., Leung, L.P., & Gordon, I.J. 2011. The impact of feral pigs ( Sus scrofa ) on an Australian lowland tropical rainforest. Wildlife Research 38 (5): 437 445. Tierney, T.A. & Cushman, J.H. 2006. Temporal changes in native and exoti c vegetation and soil characteristics following disturbances by feral pigs in a California grassland. Biological Invasions 8: 1073 1089. Troeng, S. & Rankin, E. 2005. Long term conservation efforts contribute to positive green turtle Chelonia mydas nesting trend at Tortuguero, Costa Rica. Biological Conservation 121: 111 116. Turner, M.G., Hargrove, W.W., Gardner, R.H., & Romme, W.H. 1994. Effects of fire on landscape heterogeneity in Yellowstone National Park, Wyoming. Journal of Vegetation Science 5(5): 7 31 742. United States Air Force. 2002. Integrated Natural Resources Management Plan. Eglin Air Force Base, FL. Van Auken, O.W. 2000. Shrub Invasions of North American Semiarid Grasslands. Annual Review of Ecology and Systematics. 31: 197 215. Van Langeveld e F., Van de Vijver, C.A.D., Kumar, L., Van de Koppel, J., De Ridder, N., Van Andel, J., Skidmore, A.K., Hearne, J.W., Stroosnijder, L., Bond, W.J., Prins, H.H.T., & Rietkerk, M. 2003. Effects of Fire and Herbivory on the Stability of Savanna Ecosystems. Ecology 84(2): 337 350. Van Vuren, D. & Coblentz, B.E. 1987. Some ecological effects of feral sheep on Santa Cruz Island, California, USA. Biological Conservation 41 (4): 253 268. Vermeire, L.T., Mitchell, R.B., Fuhlendorf, S.D., & Gillen, R.L. 2004. Patch burning effects on grazing distribution. Rangeland Ecology & Management 57 (3): 248 252. Vinton, M.A., Hartnett, D.C., Finck, E.J., & Briggs, J.M. 1993. Interactive effects of fire, bison ( Bison bison ) grazing and plant community c omposition in tallgrass prairie. American Midland Naturalist 129(1): 10 18. Vittoz, P. & Guisan, A. 2007. How reliable is the monitoring of permanent vegetation plots? A test with multiple observers. Journal of Vegetation Science 18(3): 413 422. Waldram, M .S., Bond, W.J., & Stock, W.D. 2008. Ecological engineering by a mega grazer: white rhino impacts on a South African savanna. Ecosystems 11: 101 112.
112 Waldrop, T.A., White, D.L., & Jones, S.M. 1992. Fire regimes for pine grassland communities in the southeastern United States. Forest Ecology and Management 47: 195 210. Walker, J. & Peet, R.K. 1984. Composition and species diversity of pine wiregrass savannas of the Green Swamp, Nort h Carolina. Vegetatio 55 (3): 163 179. Watts, W.A. 1971. Postglacial and interglacial vegetation history of southern Georgia and central Florida. Ecology 52(4): 676 690. Welander, J. 2001. Spatial and temporal dynamics of wild boar ( Sus scrofa ) rooting in a mosaic landscape. Journal of Zoology 252: 263 271. Wharton, C.H., Odum, H.T., Ewel, K., Duever, M., Lugo, A., Boyt, R., Bartholomew, J., DeBellevue, E., Brown, S., Brown, M., & Duever, L. 1976. Forested wetlands of Florida their management and use. Final report to Division of State Planning. 421pp. Center for Wetlands, Phelps Lab, University of Florida, Gainesville, FL, US. Wiggers, M.S. 2011. Some like it hot: fire and legume germination in the longleaf pine ecosystem. Ph.D. dissertation, Auburn Universit y, Auburn, AL, US. Williams, M. 1992. Americans and their forests: a historical geography Cambridge University Press, Cambridge, UK. Dumroese, D.S., Busse, M.D., Kirchner, J.W., & Risch, A.C. 2012. Do changes in soil proper ties after rooting by wild boars ( Sus scrofa ) affect understory vegetation in Swiss hardwood forests? Canadian Journal of Forest Research 42: 585 592. Wood, G.W. & Brenneman, R.E. 1980. Feral hog movements and habitat use in South Carolina. Journal of Wildlife Management 44: 420 7. Woodall, P.F. 1983. Distribution and population dynamics of dingoes ( Canis familiaris ) and feral pigs ( Sus scrofa ) in Queensland, 1945 1976. Journal of Applied Ecology 20: 85 95. Work, G. 1993. A rancher's view of the wild pi g as an economic and ecological asset to the ranching enterprise. Conference summaries of the wild pig in California oak woodland: ecology and economics San Luis Obispo, CA, US. Integrated Hardwood Range Management Program; 5, Berkeley, CA, US. Wunderlin R.P. & Hansen, B.F. 2008. Atlas of Florida Vascular Plants [S. M. Landry and K. N. Campbell (application development), Florida Center for Community Design and Research.] Institute for Systematic Botany, University of South Florida, Tampa, FL, US. Young, T.P., Chase, J.M., & Huddleston, R.T. 2001. Community succession and assembly comparing, contrasting and combining paradigms in the context of ecological restoration. Ecological restoration 19 (1) : 5 18. Zar, J.H. 2010. Biostatistical Analysis 5th ed. Prentice Hall, New Jersey, US.
113 BIOGRAPHICAL SKETCH Megan was born and raised in a small town south of Richmond, Virginia. She graduated from the University of Virginia in 2002 wit h a Bachelor of Arts degree in b iology. Me gan went on to earn a Master of Science degree in geological s ciences from Virginia Polytechnic Institute and State University in 2004. An internship working with the Bureau of Land Management and Chicago Botanic Garden took her to the tiny town of Vale in eastern Oregon. This experience fostered her interest in wetlands and restoration. After leaving Oregon, Megan went back east to teach biology laboratories at Virginia State University in 2005. Restless for a job with more time outside, she saw an opportu nity to go back to school and conduct wetland research at the University of Florida. In 2008, she became a fellow in the Adaptive Management of Water, Wetlands, and Watershed Integrative Graduate Education and Research Traineeship (IGERT) program that was funded by the National Science Foundation. Megan earned her Ph.D. in i nterdisciplinary e cology from the University of Florida in May of 2014.
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID E9URIMKLY_DYVVTN INGEST_TIME 2014-10-03T21:54:55Z PACKAGE UFE0046581_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC